X-Git-Url: https://code.citadel.org/?a=blobdiff_plain;f=libcitadel%2Flib%2Fstringbuf.c;h=08e904b315fc36f896823e6a9f0b174abf556a72;hb=aa1bc013ff54eb118855365945aae4218d099003;hp=2a286636571dd7452098b3e8f80bc56df68f082f;hpb=a075d9548b6366bfd336ab77a37edad923d85daa;p=citadel.git diff --git a/libcitadel/lib/stringbuf.c b/libcitadel/lib/stringbuf.c index 2a2866365..08e904b31 100644 --- a/libcitadel/lib/stringbuf.c +++ b/libcitadel/lib/stringbuf.c @@ -1,4 +1,22 @@ -#include "../sysdep.h" +/* + * Copyright (c) 1987-2011 by the citadel.org team + * + * This program is open source software; you can redistribute it and/or modify + * it under the terms of the GNU General Public License as published by + * the Free Software Foundation; either version 3 of the License, or + * (at your option) any later version. + * + * This program is distributed in the hope that it will be useful, + * but WITHOUT ANY WARRANTY; without even the implied warranty of + * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the + * GNU General Public License for more details. + * + * You should have received a copy of the GNU General Public License + * along with this program; if not, write to the Free Software + * Foundation, Inc., 59 Temple Place, Suite 330, Boston, MA 02111-1307 USA + */ + +#include "sysdep.h" #include #include #include @@ -10,6 +28,9 @@ #include #define SHOW_ME_VAPPEND_PRINTF #include +#ifndef LINUX_SENDFILE +#include +#endif #include "libcitadel.h" #ifdef HAVE_ICONV @@ -20,6 +41,10 @@ #include #endif +#ifdef LINUX_SENDFILE +#include +#endif + #ifdef HAVE_ZLIB #include int ZEXPORT compress_gzip(Bytef * dest, size_t * destLen, @@ -29,18 +54,97 @@ int BaseStrBufSize = 64; const char *StrBufNOTNULL = ((char*) NULL) - 1; +const char HexList[256][3] = { + "00","01","02","03","04","05","06","07","08","09","0A","0B","0C","0D","0E","0F", + "10","11","12","13","14","15","16","17","18","19","1A","1B","1C","1D","1E","1F", + "20","21","22","23","24","25","26","27","28","29","2A","2B","2C","2D","2E","2F", + "30","31","32","33","34","35","36","37","38","39","3A","3B","3C","3D","3E","3F", + "40","41","42","43","44","45","46","47","48","49","4A","4B","4C","4D","4E","4F", + "50","51","52","53","54","55","56","57","58","59","5A","5B","5C","5D","5E","5F", + "60","61","62","63","64","65","66","67","68","69","6A","6B","6C","6D","6E","6F", + "70","71","72","73","74","75","76","77","78","79","7A","7B","7C","7D","7E","7F", + "80","81","82","83","84","85","86","87","88","89","8A","8B","8C","8D","8E","8F", + "90","91","92","93","94","95","96","97","98","99","9A","9B","9C","9D","9E","9F", + "A0","A1","A2","A3","A4","A5","A6","A7","A8","A9","AA","AB","AC","AD","AE","AF", + "B0","B1","B2","B3","B4","B5","B6","B7","B8","B9","BA","BB","BC","BD","BE","BF", + "C0","C1","C2","C3","C4","C5","C6","C7","C8","C9","CA","CB","CC","CD","CE","CF", + "D0","D1","D2","D3","D4","D5","D6","D7","D8","D9","DA","DB","DC","DD","DE","DF", + "E0","E1","E2","E3","E4","E5","E6","E7","E8","E9","EA","EB","EC","ED","EE","EF", + "F0","F1","F2","F3","F4","F5","F6","F7","F8","F9","FA","FB","FC","FD","FE","FF"}; + +/** + * @defgroup StrBuf Stringbuffer, A class for manipulating strings with dynamic buffers + * StrBuf is a versatile class, aiding the handling of dynamic strings + * * reduce de/reallocations + * * reduce the need to remeasure it + * * reduce scanning over the string (in @ref StrBuf_NextTokenizer "Tokenizers") + * * allow asyncroneous IO for line and Blob based operations + * * reduce the use of memove in those + * * Quick filling in several operations with append functions + */ + +/** + * @defgroup StrBuf_DeConstructors Create/Destroy StrBufs + * @ingroup StrBuf + */ + +/** + * @defgroup StrBuf_Cast Cast operators to interact with char* based code + * @ingroup StrBuf + * use these operators to interfere with code demanding char*; + * if you need to own the content, smash me. Avoid, since we loose the length information. + */ + +/** + * @defgroup StrBuf_Filler Create/Replace/Append Content into a StrBuf + * @ingroup StrBuf + * operations to get your Strings into a StrBuf, manipulating them, or appending + */ +/** + * @defgroup StrBuf_NextTokenizer Fast tokenizer to pull tokens in sequence + * @ingroup StrBuf + * Quick tokenizer; demands of the user to pull its tokens in sequence + */ + +/** + * @defgroup StrBuf_Tokenizer tokenizer Functions; Slow ones. + * @ingroup StrBuf + * versatile tokenizer; random access to tokens, but slower; Prefer the @ref StrBuf_NextTokenizer "Next Tokenizer" + */ + +/** + * @defgroup StrBuf_BufferedIO Buffered IO with Asynchroneous reads and no unneeded memmoves (the fast ones) + * @ingroup StrBuf + * File IO to fill StrBufs; Works with work-buffer shared across several calls; + * External Cursor to maintain the current read position inside of the buffer + * the non-fast ones will use memove to keep the start of the buffer the read buffer (which is slower) + */ + +/** + * @defgroup StrBuf_IO FileIO; Prefer @ref StrBuf_BufferedIO + * @ingroup StrBuf + * Slow I/O; avoid. + */ + +/** + * @defgroup StrBuf_DeEnCoder functions to translate the contents of a buffer + * @ingroup StrBuf + * these functions translate the content of a buffer into another representation; + * some are combined Fillers and encoders + */ + /** * Private Structure for the Stringbuffer */ struct StrBuf { char *buf; /**< the pointer to the dynamic buffer */ long BufSize; /**< how many spcae do we optain */ - long BufUsed; /**< StNumber of Chars used excluding the trailing \0 */ + long BufUsed; /**< StNumber of Chars used excluding the trailing \\0 */ int ConstBuf; /**< are we just a wrapper arround a static buffer and musn't we be changed? */ #ifdef SIZE_DEBUG - long nIncreases; - char bt [SIZ]; - char bt_lastinc [SIZ]; + long nIncreases; /**< for profiling; cound how many times we needed more */ + char bt [SIZ]; /**< Stacktrace of last increase */ + char bt_lastinc [SIZ]; /**< How much did we increase last time? */ #endif }; @@ -76,11 +180,87 @@ static void StrBufBacktrace(StrBuf *Buf, int which) } #endif + +void dbg_FreeStrBuf(StrBuf *FreeMe, char *FromWhere) +{ + if (hFreeDbglog == -1){ + pid_t pid = getpid(); + char path [SIZ]; + snprintf(path, SIZ, "/tmp/libcitadel_strbuf_realloc.log.%d", pid); + hFreeDbglog = open(path, O_APPEND|O_CREAT|O_WRONLY); + } + if ((*FreeMe)->nIncreases > 0) + { + char buf[SIZ * 3]; + long n; + n = snprintf(buf, SIZ * 3, "%c+|%ld|%ld|%ld|%s|%s|\n", + FromWhere, + (*FreeMe)->nIncreases, + (*FreeMe)->BufUsed, + (*FreeMe)->BufSize, + (*FreeMe)->bt, + (*FreeMe)->bt_lastinc); + n = write(hFreeDbglog, buf, n); + } + else + { + char buf[128]; + long n; + n = snprintf(buf, 128, "%c_|0|%ld%ld|\n", + FromWhere, + (*FreeMe)->BufUsed, + (*FreeMe)->BufSize); + n = write(hFreeDbglog, buf, n); + } +} + +void dbg_IncreaseBuf(StrBuf *IncMe) +{ + Buf->nIncreases++; +#ifdef HAVE_BACKTRACE + StrBufBacktrace(Buf, 1); +#endif +} + +void dbg_Init(StrBuf *Buf) +{ + Buf->nIncreases = 0; + Buf->bt[0] = '\0'; + Buf->bt_lastinc[0] = '\0'; +#ifdef HAVE_BACKTRACE + StrBufBacktrace(Buf, 0); +#endif +} + +#else +/* void it... */ +#define dbg_FreeStrBuf(a, b) +#define dbg_IncreaseBuf(a) +#define dbg_Init(a) + #endif +/** + * @ingroup StrBuf + * @brief swaps the contents of two StrBufs + * this is to be used to have cheap switched between a work-buffer and a target buffer + * @param A First one + * @param B second one + */ +static inline void SwapBuffers(StrBuf *A, StrBuf *B) +{ + StrBuf C; + + memcpy(&C, A, sizeof(*A)); + memcpy(A, B, sizeof(*B)); + memcpy(B, &C, sizeof(C)); + +} + /** + * @ingroup StrBuf_Cast * @brief Cast operator to Plain String - * Note: if the buffer is altered by StrBuf operations, this pointer may become + * @note if the buffer is altered by StrBuf operations, this pointer may become * invalid. So don't lean on it after altering the buffer! * Since this operation is considered cheap, rather call it often than risking * your pointer to become invalid! @@ -95,6 +275,7 @@ inline const char *ChrPtr(const StrBuf *Str) } /** + * @ingroup StrBuf_Cast * @brief since we know strlen()'s result, provide it here. * @param Str the string to return the length to * @returns contentlength of the buffer @@ -105,6 +286,7 @@ inline int StrLength(const StrBuf *Str) } /** + * @ingroup StrBuf_DeConstructors * @brief local utility function to resize the buffer * @param Buf the buffer whichs storage we should increase * @param KeepOriginal should we copy the original buffer or just start over with a new one @@ -119,9 +301,12 @@ static int IncreaseBuf(StrBuf *Buf, int KeepOriginal, int DestSize) return -1; if (DestSize > 0) - while (NewSize <= DestSize) + while ((NewSize <= DestSize) && (NewSize != 0)) NewSize *= 2; + if (NewSize == 0) + return -1; + NewBuf= (char*) malloc(NewSize); if (NewBuf == NULL) return -1; @@ -138,24 +323,24 @@ static int IncreaseBuf(StrBuf *Buf, int KeepOriginal, int DestSize) free (Buf->buf); Buf->buf = NewBuf; Buf->BufSize = NewSize; -#ifdef SIZE_DEBUG - Buf->nIncreases++; -#ifdef HAVE_BACKTRACE - StrBufBacktrace(Buf, 1); -#endif -#endif + + dbg_IncreaseBuf(Buf); + return Buf->BufSize; } /** - * @brief shrink an _EMPTY_ buffer if its Buffer superseeds threshhold to NewSize. Buffercontent is thoroughly ignored and flushed. + * @ingroup StrBuf_DeConstructors + * @brief shrink / increase an _EMPTY_ buffer to NewSize. Buffercontent is thoroughly ignored and flushed. * @param Buf Buffer to shrink (has to be empty) * @param ThreshHold if the buffer is bigger then this, its readjusted * @param NewSize if we Shrink it, how big are we going to be afterwards? */ void ReAdjustEmptyBuf(StrBuf *Buf, long ThreshHold, long NewSize) { - if (Buf->BufUsed > ThreshHold) { + if ((Buf != NULL) && + (Buf->BufUsed == 0) && + (Buf->BufSize < ThreshHold)) { free(Buf->buf); Buf->buf = (char*) malloc(NewSize); Buf->BufUsed = 0; @@ -164,6 +349,7 @@ void ReAdjustEmptyBuf(StrBuf *Buf, long ThreshHold, long NewSize) } /** + * @ingroup StrBuf_DeConstructors * @brief shrink long term buffers to their real size so they don't waste memory * @param Buf buffer to shrink * @param Force if not set, will just executed if the buffer is much to big; set for lifetime strings @@ -171,6 +357,8 @@ void ReAdjustEmptyBuf(StrBuf *Buf, long ThreshHold, long NewSize) */ long StrBufShrinkToFit(StrBuf *Buf, int Force) { + if (Buf == NULL) + return -1; if (Force || (Buf->BufUsed + (Buf->BufUsed / 3) > Buf->BufSize)) { @@ -184,6 +372,7 @@ long StrBufShrinkToFit(StrBuf *Buf, int Force) } /** + * @ingroup StrBuf_DeConstructors * @brief Allocate a new buffer with default buffer size * @returns the new stringbuffer */ @@ -197,20 +386,16 @@ StrBuf* NewStrBuf(void) NewBuf->BufSize = BaseStrBufSize; NewBuf->BufUsed = 0; NewBuf->ConstBuf = 0; -#ifdef SIZE_DEBUG - NewBuf->nIncreases = 0; - NewBuf->bt[0] = '\0'; - NewBuf->bt_lastinc[0] = '\0'; -#ifdef HAVE_BACKTRACE - StrBufBacktrace(NewBuf, 0); -#endif -#endif + + dbg_Init (NewBuf); + return NewBuf; } /** + * @ingroup StrBuf_DeConstructors * @brief Copy Constructor; returns a duplicate of CopyMe - * @params CopyMe Buffer to faxmilate + * @param CopyMe Buffer to faxmilate * @returns the new stringbuffer */ StrBuf* NewStrBufDup(const StrBuf *CopyMe) @@ -226,18 +411,83 @@ StrBuf* NewStrBufDup(const StrBuf *CopyMe) NewBuf->BufUsed = CopyMe->BufUsed; NewBuf->BufSize = CopyMe->BufSize; NewBuf->ConstBuf = 0; -#ifdef SIZE_DEBUG - NewBuf->nIncreases = 0; - NewBuf->bt[0] = '\0'; - NewBuf->bt_lastinc[0] = '\0'; -#ifdef HAVE_BACKTRACE - StrBufBacktrace(NewBuf, 0); -#endif -#endif + + dbg_Init(NewBuf); + return NewBuf; } +/** + * @ingroup StrBuf_DeConstructors + * @brief Copy Constructor; CreateRelpaceMe will contain CopyFlushMe afterwards. + * @param NoMe if non-NULL, we will use that buffer as value; KeepOriginal will abused as len. + * @param CopyFlushMe Buffer to faxmilate if KeepOriginal, or to move into CreateRelpaceMe if !KeepOriginal. + * @param CreateRelpaceMe If NULL, will be created, else Flushed and filled CopyFlushMe + * @param KeepOriginal should CopyFlushMe remain intact? or may we Steal its buffer? + * @returns the new stringbuffer + */ +void NewStrBufDupAppendFlush(StrBuf **CreateRelpaceMe, StrBuf *CopyFlushMe, const char *NoMe, int KeepOriginal) +{ + StrBuf *NewBuf; + + if (CreateRelpaceMe == NULL) + return; + + if (NoMe != NULL) + { + if (*CreateRelpaceMe != NULL) + StrBufPlain(*CreateRelpaceMe, NoMe, KeepOriginal); + else + *CreateRelpaceMe = NewStrBufPlain(NoMe, KeepOriginal); + return; + } + + if (CopyFlushMe == NULL) + { + if (*CreateRelpaceMe != NULL) + FlushStrBuf(*CreateRelpaceMe); + else + *CreateRelpaceMe = NewStrBuf(); + return; + } + + /* + * Randomly Chosen: bigger than 64 chars is cheaper to swap the buffers instead of copying. + * else *CreateRelpaceMe may use more memory than needed in a longer term, CopyFlushMe might + * be a big IO-Buffer... + */ + if (KeepOriginal || (StrLength(CopyFlushMe) < 256)) + { + if (*CreateRelpaceMe == NULL) + { + *CreateRelpaceMe = NewBuf = NewStrBufPlain(NULL, CopyFlushMe->BufUsed); + dbg_Init(NewBuf); + } + else + { + NewBuf = *CreateRelpaceMe; + FlushStrBuf(NewBuf); + } + StrBufAppendBuf(NewBuf, CopyFlushMe, 0); + } + else + { + if (*CreateRelpaceMe == NULL) + { + *CreateRelpaceMe = NewBuf = NewStrBufPlain(NULL, CopyFlushMe->BufUsed); + dbg_Init(NewBuf); + } + else + NewBuf = *CreateRelpaceMe; + SwapBuffers (NewBuf, CopyFlushMe); + } + if (!KeepOriginal) + FlushStrBuf(CopyFlushMe); + return; +} + /** + * @ingroup StrBuf_DeConstructors * @brief create a new Buffer using an existing c-string * this function should also be used if you want to pre-suggest * the buffer size to allocate in conjunction with ptr == NULL @@ -257,10 +507,20 @@ StrBuf* NewStrBufPlain(const char* ptr, int nChars) else CopySize = nChars; - while (Siz <= CopySize) + while ((Siz <= CopySize) && (Siz != 0)) Siz *= 2; + if (Siz == 0) + { + return NULL; + } + NewBuf->buf = (char*) malloc(Siz); + if (NewBuf->buf == NULL) + { + free(NewBuf); + return NULL; + } NewBuf->BufSize = Siz; if (ptr != NULL) { memcpy(NewBuf->buf, ptr, CopySize); @@ -272,36 +532,47 @@ StrBuf* NewStrBufPlain(const char* ptr, int nChars) NewBuf->BufUsed = 0; } NewBuf->ConstBuf = 0; -#ifdef SIZE_DEBUG - NewBuf->nIncreases = 0; - NewBuf->bt[0] = '\0'; - NewBuf->bt_lastinc[0] = '\0'; -#ifdef HAVE_BACKTRACE - StrBufBacktrace(NewBuf, 0); -#endif -#endif + + dbg_Init(NewBuf); + return NewBuf; } /** + * @ingroup StrBuf_DeConstructors * @brief Set an existing buffer from a c-string + * @param Buf buffer to load * @param ptr c-string to put into * @param nChars set to -1 if we should work 0-terminated * @returns the new length of the string */ int StrBufPlain(StrBuf *Buf, const char* ptr, int nChars) { - size_t Siz = Buf->BufSize; + size_t Siz; size_t CopySize; + if (Buf == NULL) + return -1; + if (ptr == NULL) { + FlushStrBuf(Buf); + return -1; + } + + Siz = Buf->BufSize; + if (nChars < 0) CopySize = strlen(ptr); else CopySize = nChars; - while (Siz <= CopySize) + while ((Siz <= CopySize) && (Siz != 0)) Siz *= 2; + if (Siz == 0) { + FlushStrBuf(Buf); + return -1; + } + if (Siz != Buf->BufSize) IncreaseBuf(Buf, 0, Siz); memcpy(Buf->buf, ptr, CopySize); @@ -313,9 +584,10 @@ int StrBufPlain(StrBuf *Buf, const char* ptr, int nChars) /** + * @ingroup StrBuf_DeConstructors * @brief use strbuf as wrapper for a string constant for easy handling * @param StringConstant a string to wrap - * @param SizeOfConstant should be sizeof(StringConstant)-1 + * @param SizeOfStrConstant should be sizeof(StringConstant)-1 */ StrBuf* _NewConstStrBuf(const char* StringConstant, size_t SizeOfStrConstant) { @@ -326,16 +598,15 @@ StrBuf* _NewConstStrBuf(const char* StringConstant, size_t SizeOfStrConstant) NewBuf->BufSize = SizeOfStrConstant; NewBuf->BufUsed = SizeOfStrConstant; NewBuf->ConstBuf = 1; -#ifdef SIZE_DEBUG - NewBuf->nIncreases = 0; - NewBuf->bt[0] = '\0'; - NewBuf->bt_lastinc[0] = '\0'; -#endif + + dbg_Init(NewBuf); + return NewBuf; } /** + * @ingroup StrBuf_DeConstructors * @brief flush the content of a Buf; keep its struct * @param buf Buffer to flush */ @@ -351,6 +622,7 @@ int FlushStrBuf(StrBuf *buf) } /** + * @ingroup StrBuf_DeConstructors * @brief wipe the content of a Buf thoroughly (overwrite it -> expensive); keep its struct * @param buf Buffer to wipe */ @@ -371,6 +643,7 @@ int FLUSHStrBuf(StrBuf *buf) int hFreeDbglog = -1; #endif /** + * @ingroup StrBuf_DeConstructors * @brief Release a Buffer * Its a double pointer, so it can NULL your pointer * so fancy SIG11 appear instead of random results @@ -380,35 +653,9 @@ void FreeStrBuf (StrBuf **FreeMe) { if (*FreeMe == NULL) return; -#ifdef SIZE_DEBUG - if (hFreeDbglog == -1){ - pid_t pid = getpid(); - char path [SIZ]; - snprintf(path, SIZ, "/tmp/libcitadel_strbuf_realloc.log.%d", pid); - hFreeDbglog = open(path, O_APPEND|O_CREAT|O_WRONLY); - } - if ((*FreeMe)->nIncreases > 0) - { - char buf[SIZ * 3]; - long n; - n = snprintf(buf, SIZ * 3, "+|%ld|%ld|%ld|%s|%s|\n", - (*FreeMe)->nIncreases, - (*FreeMe)->BufUsed, - (*FreeMe)->BufSize, - (*FreeMe)->bt, - (*FreeMe)->bt_lastinc); - n = write(hFreeDbglog, buf, n); - } - else - { - char buf[128]; - long n; - n = snprintf(buf, 128, "_|0|%ld%ld|\n", - (*FreeMe)->BufUsed, - (*FreeMe)->BufSize); - n = write(hFreeDbglog, buf, n); - } -#endif + + dbg_FreeStrBuf(FreeMe, 'F'); + if (!(*FreeMe)->ConstBuf) free((*FreeMe)->buf); free(*FreeMe); @@ -416,6 +663,7 @@ void FreeStrBuf (StrBuf **FreeMe) } /** + * @ingroup StrBuf_DeConstructors * @brief flatten a Buffer to the Char * we return * Its a double pointer, so it can NULL your pointer * so fancy SIG11 appear instead of random results @@ -427,37 +675,11 @@ char *SmashStrBuf (StrBuf **SmashMe) { char *Ret; - if (*SmashMe == NULL) + if ((SmashMe == NULL) || (*SmashMe == NULL)) return NULL; -#ifdef SIZE_DEBUG - if (hFreeDbglog == -1){ - pid_t pid = getpid(); - char path [SIZ]; - snprintf(path, SIZ, "/tmp/libcitadel_strbuf_realloc.log.%d", pid); - hFreeDbglog = open(path, O_APPEND|O_CREAT|O_WRONLY); - } - if ((*SmashMe)->nIncreases > 0) - { - char buf[SIZ * 3]; - long n; - n = snprintf(buf, SIZ * 3, "S+|%ld|%ld|%ld|%s|%s|\n", - (*SmashMe)->nIncreases, - (*SmashMe)->BufUsed, - (*SmashMe)->BufSize, - (*SmashMe)->bt, - (*SmashMe)->bt_lastinc); - n = write(hFreeDbglog, buf, n); - } - else - { - char buf[128]; - long n; - n = snprintf(buf, 128, "S_|0|%ld%ld|\n", - (*SmashMe)->BufUsed, - (*SmashMe)->BufSize); - n = write(hFreeDbglog, buf, n); - } -#endif + + dbg_FreeStrBuf(SmashMe, 'S'); + Ret = (*SmashMe)->buf; free(*SmashMe); *SmashMe = NULL; @@ -465,6 +687,7 @@ char *SmashStrBuf (StrBuf **SmashMe) } /** + * @ingroup StrBuf_DeConstructors * @brief Release the buffer * If you want put your StrBuf into a Hash, use this as Destructor. * @param VFreeMe untyped pointer to a StrBuf. be shure to do the right thing [TM] @@ -474,41 +697,21 @@ void HFreeStrBuf (void *VFreeMe) StrBuf *FreeMe = (StrBuf*)VFreeMe; if (FreeMe == NULL) return; -#ifdef SIZE_DEBUG - if (hFreeDbglog == -1){ - pid_t pid = getpid(); - char path [SIZ]; - snprintf(path, SIZ, "/tmp/libcitadel_strbuf_realloc.log.%d", pid); - hFreeDbglog = open(path, O_APPEND|O_CREAT|O_WRONLY); - } - if (FreeMe->nIncreases > 0) - { - char buf[SIZ * 3]; - long n; - n = snprintf(buf, SIZ * 3, "+|%ld|%ld|%ld|%s|%s|\n", - FreeMe->nIncreases, - FreeMe->BufUsed, - FreeMe->BufSize, - FreeMe->bt, - FreeMe->bt_lastinc); - write(hFreeDbglog, buf, n); - } - else - { - char buf[128]; - long n; - n = snprintf(buf, 128, "_|%ld|%ld%ld|\n", - FreeMe->nIncreases, - FreeMe->BufUsed, - FreeMe->BufSize); - } -#endif + + dbg_FreeStrBuf(SmashMe, 'H'); + if (!FreeMe->ConstBuf) free(FreeMe->buf); free(FreeMe); } + +/******************************************************************************* + * Simple string transformations * + *******************************************************************************/ + /** + * @ingroup StrBuf * @brief Wrapper around atol */ long StrTol(const StrBuf *Buf) @@ -522,6 +725,7 @@ long StrTol(const StrBuf *Buf) } /** + * @ingroup StrBuf * @brief Wrapper around atoi */ int StrToi(const StrBuf *Buf) @@ -535,22 +739,31 @@ int StrToi(const StrBuf *Buf) } /** + * @ingroup StrBuf * @brief Checks to see if the string is a pure number + * @param Buf The buffer to inspect + * @returns 1 if its a pure number, 0, if not. */ int StrBufIsNumber(const StrBuf *Buf) { - char * pEnd; - if (Buf == NULL) { - return 0; - } - strtoll(Buf->buf, &pEnd, 10); - if (pEnd == NULL && ((Buf->buf)-pEnd) != 0) { - return 1; - } - return 0; -} -/** - * @brief modifies a Single char of the Buf + char * pEnd; + if ((Buf == NULL) || (Buf->BufUsed == 0)) { + return 0; + } + strtoll(Buf->buf, &pEnd, 10); + if (pEnd == Buf->buf) + return 0; + if ((pEnd != NULL) && (pEnd == Buf->buf + Buf->BufUsed)) + return 1; + if (Buf->buf == pEnd) + return 0; + return 0; +} + +/** + * @ingroup StrBuf_Filler + * @brief modifies a Single char of the Buf * You can point to it via char* or a zero-based integer + * @param Buf The buffer to manipulate * @param ptr char* to zero; use NULL if unused * @param nThChar zero based pointer into the string; use -1 if unused * @param PeekValue The Character to place into the position @@ -568,6 +781,34 @@ long StrBufPeek(StrBuf *Buf, const char* ptr, long nThChar, char PeekValue) } /** + * @ingroup StrBuf_Filler + * @brief modifies a range of chars of the Buf + * You can point to it via char* or a zero-based integer + * @param Buf The buffer to manipulate + * @param ptr char* to zero; use NULL if unused + * @param nThChar zero based pointer into the string; use -1 if unused + * @param nChars how many chars are to be flushed? + * @param PookValue The Character to place into that area + */ +long StrBufPook(StrBuf *Buf, const char* ptr, long nThChar, long nChars, char PookValue) +{ + if (Buf == NULL) + return -1; + if (ptr != NULL) + nThChar = ptr - Buf->buf; + if ((nThChar < 0) || (nThChar > Buf->BufUsed)) + return -1; + if (nThChar + nChars > Buf->BufUsed) + nChars = Buf->BufUsed - nThChar; + + memset(Buf->buf + nThChar, PookValue, nChars); + /* just to be shure... */ + Buf->buf[Buf->BufUsed] = 0; + return nChars; +} + +/** + * @ingroup StrBuf_Filler * @brief Append a StringBuffer to the buffer * @param Buf Buffer to modify * @param AppendBuf Buffer to copy at the end of our buffer @@ -592,6 +833,7 @@ void StrBufAppendBuf(StrBuf *Buf, const StrBuf *AppendBuf, unsigned long Offset) /** + * @ingroup StrBuf_Filler * @brief Append a C-String to the buffer * @param Buf Buffer to modify * @param AppendBuf Buffer to copy at the end of our buffer @@ -623,8 +865,127 @@ void StrBufAppendBufPlain(StrBuf *Buf, const char *AppendBuf, long AppendSize, u } /** + * @ingroup StrBuf_Filler + * @brief sprintf like function appending the formated string to the buffer + * vsnprintf version to wrap into own calls + * @param Buf Buffer to extend by format and Params + * @param format printf alike format to add + * @param ap va_list containing the items for format + */ +void StrBufVAppendPrintf(StrBuf *Buf, const char *format, va_list ap) +{ + va_list apl; + size_t BufSize; + size_t nWritten; + size_t Offset; + size_t newused; + + if ((Buf == NULL) || (format == NULL)) + return; + + BufSize = Buf->BufSize; + nWritten = Buf->BufSize + 1; + Offset = Buf->BufUsed; + newused = Offset + nWritten; + + while (newused >= BufSize) { + va_copy(apl, ap); + nWritten = vsnprintf(Buf->buf + Offset, + Buf->BufSize - Offset, + format, apl); + va_end(apl); + newused = Offset + nWritten; + if (newused >= Buf->BufSize) { + if (IncreaseBuf(Buf, 1, newused) == -1) + return; /* TODO: error handling? */ + newused = Buf->BufSize + 1; + } + else { + Buf->BufUsed = Offset + nWritten; + BufSize = Buf->BufSize; + } + + } +} + +/** + * @ingroup StrBuf_Filler + * @brief sprintf like function appending the formated string to the buffer + * @param Buf Buffer to extend by format and Params + * @param format printf alike format to add + */ +void StrBufAppendPrintf(StrBuf *Buf, const char *format, ...) +{ + size_t BufSize; + size_t nWritten; + size_t Offset; + size_t newused; + va_list arg_ptr; + + if ((Buf == NULL) || (format == NULL)) + return; + + BufSize = Buf->BufSize; + nWritten = Buf->BufSize + 1; + Offset = Buf->BufUsed; + newused = Offset + nWritten; + + while (newused >= BufSize) { + va_start(arg_ptr, format); + nWritten = vsnprintf(Buf->buf + Buf->BufUsed, + Buf->BufSize - Buf->BufUsed, + format, arg_ptr); + va_end(arg_ptr); + newused = Buf->BufUsed + nWritten; + if (newused >= Buf->BufSize) { + if (IncreaseBuf(Buf, 1, newused) == -1) + return; /* TODO: error handling? */ + newused = Buf->BufSize + 1; + } + else { + Buf->BufUsed += nWritten; + BufSize = Buf->BufSize; + } + + } +} + +/** + * @ingroup StrBuf_Filler + * @brief sprintf like function putting the formated string into the buffer + * @param Buf Buffer to extend by format and Parameters + * @param format printf alike format to add + */ +void StrBufPrintf(StrBuf *Buf, const char *format, ...) +{ + size_t nWritten; + va_list arg_ptr; + + if ((Buf == NULL) || (format == NULL)) + return; + + nWritten = Buf->BufSize + 1; + while (nWritten >= Buf->BufSize) { + va_start(arg_ptr, format); + nWritten = vsnprintf(Buf->buf, Buf->BufSize, format, arg_ptr); + va_end(arg_ptr); + if (nWritten >= Buf->BufSize) { + if (IncreaseBuf(Buf, 0, 0) == -1) + return; /* TODO: error handling? */ + nWritten = Buf->BufSize + 1; + continue; + } + Buf->BufUsed = nWritten ; + } +} + +/** + * @ingroup StrBuf_Filler * @brief Callback for cURL to append the webserver reply to a buffer - * @params pre-defined by the cURL API; see man 3 curl for mre info + * @param ptr pre-defined by the cURL API; see man 3 curl for mre info + * @param size pre-defined by the cURL API; see man 3 curl for mre info + * @param nmemb pre-defined by the cURL API; see man 3 curl for mre info + * @param stream pre-defined by the cURL API; see man 3 curl for mre info */ size_t CurlFillStrBuf_callback(void *ptr, size_t size, size_t nmemb, void *stream) { @@ -640,954 +1001,574 @@ size_t CurlFillStrBuf_callback(void *ptr, size_t size, size_t nmemb, void *strea } -/** - * @brief Escape a string for feeding out as a URL while appending it to a Buffer - * @param outbuf the output buffer - * @param oblen the size of outbuf to sanitize - * @param strbuf the input buffer +/** + * @ingroup StrBuf + * @brief extracts a substring from Source into dest + * @param dest buffer to place substring into + * @param Source string to copy substring from + * @param Offset chars to skip from start + * @param nChars number of chars to copy + * @returns the number of chars copied; may be different from nChars due to the size of Source */ -void StrBufUrlescAppend(StrBuf *OutBuf, const StrBuf *In, const char *PlainIn) +int StrBufSub(StrBuf *dest, const StrBuf *Source, unsigned long Offset, size_t nChars) { - const char *pch, *pche; - char *pt, *pte; - int b, c, len; - const char ec[] = " +#&;`'|*?-~<>^()[]{}/$\"\\"; - int eclen = sizeof(ec) -1; - - if (((In == NULL) && (PlainIn == NULL)) || (OutBuf == NULL) ) - return; - if (PlainIn != NULL) { - len = strlen(PlainIn); - pch = PlainIn; - pche = pch + len; + size_t NCharsRemain; + if (Offset > Source->BufUsed) + { + if (dest != NULL) + FlushStrBuf(dest); + return 0; } - else { - pch = In->buf; - pche = pch + In->BufUsed; - len = In->BufUsed; + if (Offset + nChars < Source->BufUsed) + { + if (nChars >= dest->BufSize) + IncreaseBuf(dest, 0, nChars + 1); + memcpy(dest->buf, Source->buf + Offset, nChars); + dest->BufUsed = nChars; + dest->buf[dest->BufUsed] = '\0'; + return nChars; } + NCharsRemain = Source->BufUsed - Offset; + if (NCharsRemain >= dest->BufSize) + IncreaseBuf(dest, 0, NCharsRemain + 1); + memcpy(dest->buf, Source->buf + Offset, NCharsRemain); + dest->BufUsed = NCharsRemain; + dest->buf[dest->BufUsed] = '\0'; + return NCharsRemain; +} - if (len == 0) +/** + * @ingroup StrBuf + * @brief Cut nChars from the start of the string + * @param Buf Buffer to modify + * @param nChars how many chars should be skipped? + */ +void StrBufCutLeft(StrBuf *Buf, int nChars) +{ + if ((Buf == NULL) || (Buf->BufUsed == 0)) return; + if (nChars >= Buf->BufUsed) { + FlushStrBuf(Buf); return; - - pt = OutBuf->buf + OutBuf->BufUsed; - pte = OutBuf->buf + OutBuf->BufSize - 4; /**< we max append 3 chars at once plus the \0 */ - - while (pch < pche) { - if (pt >= pte) { - IncreaseBuf(OutBuf, 1, -1); - pte = OutBuf->buf + OutBuf->BufSize - 4; /**< we max append 3 chars at once plus the \0 */ - pt = OutBuf->buf + OutBuf->BufUsed; - } - - c = 0; - for (b = 0; b < eclen; ++b) { - if (*pch == ec[b]) { - c = 1; - b += eclen; - } - } - if (c == 1) { - sprintf(pt,"%%%02X", *pch); - pt += 3; - OutBuf->BufUsed += 3; - pch ++; - } - else { - *(pt++) = *(pch++); - OutBuf->BufUsed++; - } } - *pt = '\0'; + memmove(Buf->buf, Buf->buf + nChars, Buf->BufUsed - nChars); + Buf->BufUsed -= nChars; + Buf->buf[Buf->BufUsed] = '\0'; } /** - * @brief Append a string, escaping characters which have meaning in HTML. - * - * @param Target target buffer - * @param Source source buffer; set to NULL if you just have a C-String - * @param PlainIn Plain-C string to append; set to NULL if unused - * @param nbsp If nonzero, spaces are converted to non-breaking spaces. - * @param nolinebreaks if set to 1, linebreaks are removed from the string. - * if set to 2, linebreaks are replaced by <br/> + * @ingroup StrBuf + * @brief Cut the trailing n Chars from the string + * @param Buf Buffer to modify + * @param nChars how many chars should be trunkated? */ -long StrEscAppend(StrBuf *Target, const StrBuf *Source, const char *PlainIn, int nbsp, int nolinebreaks) +void StrBufCutRight(StrBuf *Buf, int nChars) { - const char *aptr, *eiptr; - char *bptr, *eptr; - long len; - - if (((Source == NULL) && (PlainIn == NULL)) || (Target == NULL) ) - return -1; - - if (PlainIn != NULL) { - aptr = PlainIn; - len = strlen(PlainIn); - eiptr = aptr + len; + if ((Buf == NULL) || (Buf->BufUsed == 0)) return; + if (nChars >= Buf->BufUsed) { + FlushStrBuf(Buf); + return; } - else { - aptr = Source->buf; - eiptr = aptr + Source->BufUsed; - len = Source->BufUsed; + Buf->BufUsed -= nChars; + Buf->buf[Buf->BufUsed] = '\0'; +} + +/** + * @ingroup StrBuf + * @brief Cut the string after n Chars + * @param Buf Buffer to modify + * @param AfternChars after how many chars should we trunkate the string? + * @param At if non-null and points inside of our string, cut it there. + */ +void StrBufCutAt(StrBuf *Buf, int AfternChars, const char *At) +{ + if ((Buf == NULL) || (Buf->BufUsed == 0)) return; + if (At != NULL){ + AfternChars = At - Buf->buf; } - if (len == 0) - return -1; + if ((AfternChars < 0) || (AfternChars >= Buf->BufUsed)) + return; + Buf->BufUsed = AfternChars; + Buf->buf[Buf->BufUsed] = '\0'; +} - bptr = Target->buf + Target->BufUsed; - eptr = Target->buf + Target->BufSize - 11; /* our biggest unit to put in... */ - while (aptr < eiptr){ - if(bptr >= eptr) { - IncreaseBuf(Target, 1, -1); - eptr = Target->buf + Target->BufSize - 11; /* our biggest unit to put in... */ - bptr = Target->buf + Target->BufUsed; - } - if (*aptr == '<') { - memcpy(bptr, "<", 4); - bptr += 4; - Target->BufUsed += 4; - } - else if (*aptr == '>') { - memcpy(bptr, ">", 4); - bptr += 4; - Target->BufUsed += 4; - } - else if (*aptr == '&') { - memcpy(bptr, "&", 5); - bptr += 5; - Target->BufUsed += 5; - } - else if (*aptr == '"') { - memcpy(bptr, """, 6); - bptr += 6; - Target->BufUsed += 6; - } - else if (*aptr == '\'') { - memcpy(bptr, "'", 5); - bptr += 5; - Target->BufUsed += 5; - } - else if (*aptr == LB) { - *bptr = '<'; - bptr ++; - Target->BufUsed ++; - } - else if (*aptr == RB) { - *bptr = '>'; - bptr ++; - Target->BufUsed ++; - } - else if (*aptr == QU) { - *bptr ='"'; - bptr ++; - Target->BufUsed ++; - } - else if ((*aptr == 32) && (nbsp == 1)) { - memcpy(bptr, " ", 6); - bptr += 6; - Target->BufUsed += 6; - } - else if ((*aptr == '\n') && (nolinebreaks == 1)) { - *bptr='\0'; /* nothing */ - } - else if ((*aptr == '\n') && (nolinebreaks == 2)) { - memcpy(bptr, "<br/>", 11); - bptr += 11; - Target->BufUsed += 11; - } +/** + * @ingroup StrBuf + * @brief Strip leading and trailing spaces from a string; with premeasured and adjusted length. + * @param Buf the string to modify + */ +void StrBufTrim(StrBuf *Buf) +{ + int delta = 0; + if ((Buf == NULL) || (Buf->BufUsed == 0)) return; + while ((Buf->BufUsed > 0) && + isspace(Buf->buf[Buf->BufUsed - 1])) + { + Buf->BufUsed --; + } + Buf->buf[Buf->BufUsed] = '\0'; - else if ((*aptr == '\r') && (nolinebreaks != 0)) { - *bptr='\0'; /* nothing */ - } - else{ - *bptr = *aptr; - bptr++; - Target->BufUsed ++; - } - aptr ++; + if (Buf->BufUsed == 0) return; + + while ((Buf->BufUsed > delta) && (isspace(Buf->buf[delta]))){ + delta ++; } - *bptr = '\0'; - if ((bptr = eptr - 1 ) && !IsEmptyStr(aptr) ) - return -1; - return Target->BufUsed; + if (delta > 0) StrBufCutLeft(Buf, delta); } - /** - * @brief Append a string, escaping characters which have meaning in HTML. - * Converts linebreaks into blanks; escapes single quotes - * @param Target target buffer - * @param Source source buffer; set to NULL if you just have a C-String - * @param PlainIn Plain-C string to append; set to NULL if unused + * @ingroup StrBuf + * @brief changes all spaces in the string (tab, linefeed...) to Blank (0x20) + * @param Buf the string to modify */ -void StrMsgEscAppend(StrBuf *Target, const StrBuf *Source, const char *PlainIn) +void StrBufSpaceToBlank(StrBuf *Buf) { - const char *aptr, *eiptr; - char *tptr, *eptr; - long len; + char *pche, *pch; - if (((Source == NULL) && (PlainIn == NULL)) || (Target == NULL) ) - return ; + if ((Buf == NULL) || (Buf->BufUsed == 0)) return; - if (PlainIn != NULL) { - aptr = PlainIn; - len = strlen(PlainIn); - eiptr = aptr + len; - } - else { - aptr = Source->buf; - eiptr = aptr + Source->BufUsed; - len = Source->BufUsed; + pch = Buf->buf; + pche = pch + Buf->BufUsed; + while (pch < pche) + { + if (isspace(*pch)) + *pch = ' '; + pch ++; } +} - if (len == 0) - return; +void StrBufStripAllBut(StrBuf *Buf, char leftboundary, char rightboundary) +{ + const char *pBuff; + const char *pLeft; + const char *pRight; - eptr = Target->buf + Target->BufSize - 8; - tptr = Target->buf + Target->BufUsed; - - while (aptr < eiptr){ - if(tptr >= eptr) { - IncreaseBuf(Target, 1, -1); - eptr = Target->buf + Target->BufSize - 8; - tptr = Target->buf + Target->BufUsed; - } - - if (*aptr == '\n') { - *tptr = ' '; - Target->BufUsed++; - } - else if (*aptr == '\r') { - *tptr = ' '; - Target->BufUsed++; - } - else if (*aptr == '\'') { - *(tptr++) = '&'; - *(tptr++) = '#'; - *(tptr++) = '3'; - *(tptr++) = '9'; - *tptr = ';'; - Target->BufUsed += 5; - } else { - *tptr = *aptr; - Target->BufUsed++; - } - tptr++; aptr++; + if (Buf == NULL) + return; + pLeft = pBuff = Buf->buf; + while (pBuff != NULL) { + pLeft = pBuff; + pBuff = strchr(pBuff, leftboundary); + if (pBuff != NULL) + pBuff++; } - *tptr = '\0'; + + if (pLeft != NULL) + pBuff = pLeft; + else + pBuff = Buf->buf; + pRight = strchr(pBuff, rightboundary); + if (pRight != NULL) + StrBufCutAt(Buf, 0, pRight); + if (pLeft != NULL) + StrBufCutLeft(Buf, pLeft - Buf->buf); } - /** - * @brief Append a string, escaping characters which have meaning in ICAL. - * [\n,] - * @param Target target buffer - * @param Source source buffer; set to NULL if you just have a C-String - * @param PlainIn Plain-C string to append; set to NULL if unused + * @ingroup StrBuf_Filler + * @brief uppercase the contents of a buffer + * @param Buf the buffer to translate */ -void StrIcalEscAppend(StrBuf *Target, const StrBuf *Source, const char *PlainIn) +void StrBufUpCase(StrBuf *Buf) { - const char *aptr, *eiptr; - char *tptr, *eptr; - long len; + char *pch, *pche; - if (((Source == NULL) && (PlainIn == NULL)) || (Target == NULL) ) - return ; + if ((Buf == NULL) || (Buf->BufUsed == 0)) return; - if (PlainIn != NULL) { - aptr = PlainIn; - len = strlen(PlainIn); - eiptr = aptr + len; - } - else { - aptr = Source->buf; - eiptr = aptr + Source->BufUsed; - len = Source->BufUsed; + pch = Buf->buf; + pche = pch + Buf->BufUsed; + while (pch < pche) { + *pch = toupper(*pch); + pch ++; } +} - if (len == 0) - return; - eptr = Target->buf + Target->BufSize - 8; - tptr = Target->buf + Target->BufUsed; - - while (aptr < eiptr){ - if(tptr + 3 >= eptr) { - IncreaseBuf(Target, 1, -1); - eptr = Target->buf + Target->BufSize - 8; - tptr = Target->buf + Target->BufUsed; - } - - if (*aptr == '\n') { - *tptr = '\\'; - Target->BufUsed++; - tptr++; - *tptr = 'n'; - Target->BufUsed++; - } - else if (*aptr == '\r') { - *tptr = '\\'; - Target->BufUsed++; - tptr++; - *tptr = 'r'; - Target->BufUsed++; - } - else if (*aptr == ',') { - *tptr = '\\'; - Target->BufUsed++; - tptr++; - *tptr = ','; - Target->BufUsed++; - } else { - *tptr = *aptr; - Target->BufUsed++; - } - tptr++; aptr++; +/** + * @ingroup StrBuf_Filler + * @brief lowercase the contents of a buffer + * @param Buf the buffer to translate + */ +void StrBufLowerCase(StrBuf *Buf) +{ + char *pch, *pche; + + if ((Buf == NULL) || (Buf->BufUsed == 0)) return; + + pch = Buf->buf; + pche = pch + Buf->BufUsed; + while (pch < pche) { + *pch = tolower(*pch); + pch ++; } - *tptr = '\0'; } + +/******************************************************************************* + * a tokenizer that kills, maims, and destroys * + *******************************************************************************/ + /** - * @brief Append a string, escaping characters which have meaning in JavaScript strings . - * - * @param Target target buffer - * @param Source source buffer; set to NULL if you just have a C-String - * @param PlainIn Plain-C string to append; set to NULL if unused - * @returns size of result or -1 + * @ingroup StrBuf_Tokenizer + * @brief Replace a token at a given place with a given length by another token with given length + * @param Buf String where to work on + * @param where where inside of the Buf is the search-token + * @param HowLong How long is the token to be replaced + * @param Repl Token to insert at 'where' + * @param ReplLen Length of repl + * @returns -1 if fail else length of resulting Buf */ -long StrECMAEscAppend(StrBuf *Target, const StrBuf *Source, const char *PlainIn) +int StrBufReplaceToken(StrBuf *Buf, long where, long HowLong, + const char *Repl, long ReplLen) { - const char *aptr, *eiptr; - char *bptr, *eptr; - long len; - if (((Source == NULL) && (PlainIn == NULL)) || (Target == NULL) ) + if ((Buf == NULL) || + (where > Buf->BufUsed) || + (where + HowLong > Buf->BufUsed)) return -1; - if (PlainIn != NULL) { - aptr = PlainIn; - len = strlen(PlainIn); - eiptr = aptr + len; - } - else { - aptr = Source->buf; - eiptr = aptr + Source->BufUsed; - len = Source->BufUsed; - } + if (where + ReplLen - HowLong > Buf->BufSize) + if (IncreaseBuf(Buf, 1, Buf->BufUsed + ReplLen) < 0) + return -1; - if (len == 0) - return -1; + memmove(Buf->buf + where + ReplLen, + Buf->buf + where + HowLong, + Buf->BufUsed - where - HowLong); + + memcpy(Buf->buf + where, + Repl, ReplLen); - bptr = Target->buf + Target->BufUsed; - eptr = Target->buf + Target->BufSize - 3; /* our biggest unit to put in... */ + Buf->BufUsed += ReplLen - HowLong; - while (aptr < eiptr){ - if(bptr >= eptr) { - IncreaseBuf(Target, 1, -1); - eptr = Target->buf + Target->BufSize - 3; - bptr = Target->buf + Target->BufUsed; - } - if (*aptr == '"') { - *bptr = '\\'; - bptr ++; - *bptr = '"'; - bptr ++; - Target->BufUsed += 2; - } else if (*aptr == '\\') { - *bptr = '\\'; - bptr ++; - *bptr = '\\'; - bptr ++; - Target->BufUsed += 2; - } - else{ - *bptr = *aptr; - bptr++; - Target->BufUsed ++; - } - aptr ++; + return Buf->BufUsed; +} + +/** + * @ingroup StrBuf_Tokenizer + * @brief Counts the numbmer of tokens in a buffer + * @param source String to count tokens in + * @param tok Tokenizer char to count + * @returns numbers of tokenizer chars found + */ +int StrBufNum_tokens(const StrBuf *source, char tok) +{ + char *pch, *pche; + long NTokens; + if ((source == NULL) || (source->BufUsed == 0)) + return 0; + if ((source->BufUsed == 1) && (*source->buf == tok)) + return 2; + NTokens = 1; + pch = source->buf; + pche = pch + source->BufUsed; + while (pch < pche) + { + if (*pch == tok) + NTokens ++; + pch ++; } - *bptr = '\0'; - if ((bptr == eptr - 1 ) && !IsEmptyStr(aptr) ) - return -1; - return Target->BufUsed; + return NTokens; } /** - * @brief Append a string, escaping characters which have meaning in HTML + json. - * - * @param Target target buffer - * @param Source source buffer; set to NULL if you just have a C-String - * @param PlainIn Plain-C string to append; set to NULL if unused - * @param nbsp If nonzero, spaces are converted to non-breaking spaces. - * @param nolinebreaks if set to 1, linebreaks are removed from the string. - * if set to 2, linebreaks are replaced by <br/> + * @ingroup StrBuf_Tokenizer + * @brief a string tokenizer + * @param Source StringBuffer to read into + * @param parmnum n'th Parameter to remove + * @param separator tokenizer character + * @returns -1 if not found, else length of token. */ -long StrHtmlEcmaEscAppend(StrBuf *Target, const StrBuf *Source, const char *PlainIn, int nbsp, int nolinebreaks) +int StrBufRemove_token(StrBuf *Source, int parmnum, char separator) { - const char *aptr, *eiptr; - char *bptr, *eptr; - long len; - int IsUtf8Sequence = 0; + int ReducedBy; + char *d, *s, *end; /* dest, source */ + int count = 0; - if (((Source == NULL) && (PlainIn == NULL)) || (Target == NULL) ) - return -1; + /* Find desired @parameter */ + end = Source->buf + Source->BufUsed; + d = Source->buf; + while ((d <= end) && + (count < parmnum)) + { + /* End of string, bail! */ + if (!*d) { + d = NULL; + break; + } + if (*d == separator) { + count++; + } + d++; + } + if ((d == NULL) || (d >= end)) + return 0; /* @Parameter not found */ - if (PlainIn != NULL) { - aptr = PlainIn; - len = strlen(PlainIn); - eiptr = aptr + len; + /* Find next @parameter */ + s = d; + while ((s <= end) && + (*s && *s != separator)) + { + s++; + } + if (*s == separator) + s++; + ReducedBy = d - s; + + /* Hack and slash */ + if (s >= end) { + return 0; + } + else if (*s) { + memmove(d, s, Source->BufUsed - (s - Source->buf)); + Source->BufUsed += ReducedBy; + Source->buf[Source->BufUsed] = '\0'; + } + else if (d == Source->buf) { + *d = 0; + Source->BufUsed = 0; } else { - aptr = Source->buf; - eiptr = aptr + Source->BufUsed; - len = Source->BufUsed; + *--d = '\0'; + Source->BufUsed += ReducedBy; + } + /* + while (*s) { + *d++ = *s++; } + *d = 0; + */ + return ReducedBy; +} - if (len == 0) - return -1; - bptr = Target->buf + Target->BufUsed; - eptr = Target->buf + Target->BufSize - 11; /* our biggest unit to put in... */ +/** + * @ingroup StrBuf_Tokenizer + * @brief a string tokenizer + * @param dest Destination StringBuffer + * @param Source StringBuffer to read into + * @param parmnum n'th Parameter to extract + * @param separator tokenizer character + * @returns -1 if not found, else length of token. + */ +int StrBufExtract_token(StrBuf *dest, const StrBuf *Source, int parmnum, char separator) +{ + const char *s, *e; //* source * / + int len = 0; //* running total length of extracted string * / + int current_token = 0; //* token currently being processed * / + + if (dest != NULL) { + dest->buf[0] = '\0'; + dest->BufUsed = 0; + } + else + return(-1); - while (aptr < eiptr){ - if(bptr >= eptr) { - IncreaseBuf(Target, 1, -1); - eptr = Target->buf + Target->BufSize - 11; /* our biggest unit to put in... */ - bptr = Target->buf + Target->BufUsed; - } - if (*aptr == '<') { - memcpy(bptr, "<", 4); - bptr += 4; - Target->BufUsed += 4; - } - else if (*aptr == '>') { - memcpy(bptr, ">", 4); - bptr += 4; - Target->BufUsed += 4; - } - else if (*aptr == '&') { - memcpy(bptr, "&", 5); - bptr += 5; - Target->BufUsed += 5; - } - else if (*aptr == LB) { - *bptr = '<'; - bptr ++; - Target->BufUsed ++; - } - else if (*aptr == RB) { - *bptr = '>'; - bptr ++; - Target->BufUsed ++; - } - else if ((*aptr == 32) && (nbsp == 1)) { - memcpy(bptr, " ", 6); - bptr += 6; - Target->BufUsed += 6; - } - else if ((*aptr == '\n') && (nolinebreaks == 1)) { - *bptr='\0'; /* nothing */ - } - else if ((*aptr == '\n') && (nolinebreaks == 2)) { - memcpy(bptr, "<br/>", 11); - bptr += 11; - Target->BufUsed += 11; - } + if ((Source == NULL) || (Source->BufUsed ==0)) { + return(-1); + } + s = Source->buf; + e = s + Source->BufUsed; - else if ((*aptr == '\r') && (nolinebreaks != 0)) { - *bptr='\0'; /* nothing */ - } + //cit_backtrace(); + //lprintf (CTDL_DEBUG, "test >: n: %d sep: %c source: %s \n willi \n", parmnum, separator, source); - else if ((*aptr == '"') || (*aptr == QU)) { - *bptr = '\\'; - bptr ++; - *bptr = '"'; - bptr ++; - Target->BufUsed += 2; - } else if (*aptr == '\\') { - *bptr = '\\'; - bptr ++; - *bptr = '\\'; - bptr ++; - Target->BufUsed += 2; + while ((s < e) && !IsEmptyStr(s)) { + if (*s == separator) { + ++current_token; } - else { - if (IsUtf8Sequence != 0) { - IsUtf8Sequence --; - *bptr = *aptr; - bptr++; - Target->BufUsed ++; - } - else { - if (*aptr >= 0x20) - { - IsUtf8Sequence = Ctdl_GetUtf8SequenceLength(aptr, eiptr); - - *bptr = *aptr; - bptr++; - Target->BufUsed ++; - } + if (len >= dest->BufSize) { + dest->BufUsed = len; + if (IncreaseBuf(dest, 1, -1) < 0) { + dest->BufUsed --; + break; } - } - aptr ++; + if ( (current_token == parmnum) && + (*s != separator)) { + dest->buf[len] = *s; + ++len; + } + else if (current_token > parmnum) { + break; + } + ++s; } - *bptr = '\0'; - if ((bptr = eptr - 1 ) && !IsEmptyStr(aptr) ) - return -1; - return Target->BufUsed; + + dest->buf[len] = '\0'; + dest->BufUsed = len; + + if (current_token < parmnum) { + //lprintf (CTDL_DEBUG,"test Source->BufUsed) - { - FlushStrBuf(dest); - return 0; - } - if (Offset + nChars < Source->BufUsed) - { - if (nChars >= dest->BufSize) - IncreaseBuf(dest, 0, nChars + 1); - memcpy(dest->buf, Source->buf + Offset, nChars); - dest->BufUsed = nChars; - dest->buf[dest->BufUsed] = '\0'; - return nChars; - } - NCharsRemain = Source->BufUsed - Offset; - if (NCharsRemain >= dest->BufSize) - IncreaseBuf(dest, 0, NCharsRemain + 1); - memcpy(dest->buf, Source->buf + Offset, NCharsRemain); - dest->BufUsed = NCharsRemain; - dest->buf[dest->BufUsed] = '\0'; - return NCharsRemain; -} + + /** - * @brief sprintf like function appending the formated string to the buffer - * vsnprintf version to wrap into own calls - * @param Buf Buffer to extend by format and @params - * @param format printf alike format to add - * @param ap va_list containing the items for format + * @ingroup StrBuf_Tokenizer + * @brief a string tokenizer to fetch an integer + * @param Source String containing tokens + * @param parmnum n'th Parameter to extract + * @param separator tokenizer character + * @returns 0 if not found, else integer representation of the token */ -void StrBufVAppendPrintf(StrBuf *Buf, const char *format, va_list ap) +int StrBufExtract_int(const StrBuf* Source, int parmnum, char separator) { - va_list apl; - size_t BufSize; - size_t nWritten; - size_t Offset; - size_t newused; - - if ((Buf == NULL) || (format == NULL)) - return; - - BufSize = Buf->BufSize; - nWritten = Buf->BufSize + 1; - Offset = Buf->BufUsed; - newused = Offset + nWritten; + StrBuf tmp; + char buf[64]; - while (newused >= BufSize) { - va_copy(apl, ap); - nWritten = vsnprintf(Buf->buf + Offset, - Buf->BufSize - Offset, - format, apl); - va_end(apl); - newused = Offset + nWritten; - if (newused >= Buf->BufSize) { - IncreaseBuf(Buf, 1, newused); - newused = Buf->BufSize + 1; - } - else { - Buf->BufUsed = Offset + nWritten; - BufSize = Buf->BufSize; - } - - } + tmp.buf = buf; + buf[0] = '\0'; + tmp.BufSize = 64; + tmp.BufUsed = 0; + tmp.ConstBuf = 1; + if (StrBufExtract_token(&tmp, Source, parmnum, separator) > 0) + return(atoi(buf)); + else + return 0; } /** - * @brief sprintf like function appending the formated string to the buffer - * @param Buf Buffer to extend by format and @params - * @param format printf alike format to add - * @param ap va_list containing the items for format + * @ingroup StrBuf_Tokenizer + * @brief a string tokenizer to fetch a long integer + * @param Source String containing tokens + * @param parmnum n'th Parameter to extract + * @param separator tokenizer character + * @returns 0 if not found, else long integer representation of the token */ -void StrBufAppendPrintf(StrBuf *Buf, const char *format, ...) +long StrBufExtract_long(const StrBuf* Source, int parmnum, char separator) { - size_t BufSize; - size_t nWritten; - size_t Offset; - size_t newused; - va_list arg_ptr; + StrBuf tmp; + char buf[64]; - if ((Buf == NULL) || (format == NULL)) - return; - - BufSize = Buf->BufSize; - nWritten = Buf->BufSize + 1; - Offset = Buf->BufUsed; - newused = Offset + nWritten; - - while (newused >= BufSize) { - va_start(arg_ptr, format); - nWritten = vsnprintf(Buf->buf + Buf->BufUsed, - Buf->BufSize - Buf->BufUsed, - format, arg_ptr); - va_end(arg_ptr); - newused = Buf->BufUsed + nWritten; - if (newused >= Buf->BufSize) { - IncreaseBuf(Buf, 1, newused); - newused = Buf->BufSize + 1; - } - else { - Buf->BufUsed += nWritten; - BufSize = Buf->BufSize; - } - - } + tmp.buf = buf; + buf[0] = '\0'; + tmp.BufSize = 64; + tmp.BufUsed = 0; + tmp.ConstBuf = 1; + if (StrBufExtract_token(&tmp, Source, parmnum, separator) > 0) + return(atoi(buf)); + else + return 0; } + /** - * @brief sprintf like function putting the formated string into the buffer - * @param Buf Buffer to extend by format and @params - * @param format printf alike format to add - * @param ap va_list containing the items for format + * @ingroup StrBuf_Tokenizer + * @brief a string tokenizer to fetch an unsigned long + * @param Source String containing tokens + * @param parmnum n'th Parameter to extract + * @param separator tokenizer character + * @returns 0 if not found, else unsigned long representation of the token */ -void StrBufPrintf(StrBuf *Buf, const char *format, ...) +unsigned long StrBufExtract_unsigned_long(const StrBuf* Source, int parmnum, char separator) { - size_t nWritten; - va_list arg_ptr; + StrBuf tmp; + char buf[64]; + char *pnum; - if ((Buf == NULL) || (format == NULL)) - return; - - nWritten = Buf->BufSize + 1; - while (nWritten >= Buf->BufSize) { - va_start(arg_ptr, format); - nWritten = vsnprintf(Buf->buf, Buf->BufSize, format, arg_ptr); - va_end(arg_ptr); - if (nWritten >= Buf->BufSize) { - IncreaseBuf(Buf, 0, 0); - nWritten = Buf->BufSize + 1; - continue; - } - Buf->BufUsed = nWritten ; + tmp.buf = buf; + buf[0] = '\0'; + tmp.BufSize = 64; + tmp.BufUsed = 0; + tmp.ConstBuf = 1; + if (StrBufExtract_token(&tmp, Source, parmnum, separator) > 0) { + pnum = &buf[0]; + if (*pnum == '-') + pnum ++; + return (unsigned long) atol(pnum); } + else + return 0; } + /** - * @brief Counts the numbmer of tokens in a buffer - * @param Source String to count tokens in - * @param tok Tokenizer char to count - * @returns numbers of tokenizer chars found + * @ingroup StrBuf_NextTokenizer + * @brief a string tokenizer; Bounds checker + * function to make shure whether StrBufExtract_NextToken and friends have reached the end of the string. + * @param Source our tokenbuffer + * @param pStart the token iterator pointer to inspect + * @returns whether the revolving pointer is inside of the search range */ -int StrBufNum_tokens(const StrBuf *source, char tok) +int StrBufHaveNextToken(const StrBuf *Source, const char **pStart) { - if (source == NULL) + if ((Source == NULL) || + (*pStart == StrBufNOTNULL) || + (Source->BufUsed == 0)) + { + return 0; + } + if (*pStart == NULL) + { + return 1; + } + else if (*pStart > Source->buf + Source->BufUsed) + { + return 0; + } + else if (*pStart <= Source->buf) + { return 0; - return num_tokens(source->buf, tok); + } + + return 1; } -/* - * remove_token() - a tokenizer that kills, maims, and destroys - */ /** + * @ingroup StrBuf_NextTokenizer * @brief a string tokenizer + * @param dest Destination StringBuffer * @param Source StringBuffer to read into - * @param parmnum n'th @parameter to remove - * @param separator tokenizer @param + * @param pStart pointer to the end of the last token. Feed with NULL on start. + * @param separator tokenizer * @returns -1 if not found, else length of token. */ -int StrBufRemove_token(StrBuf *Source, int parmnum, char separator) +int StrBufExtract_NextToken(StrBuf *dest, const StrBuf *Source, const char **pStart, char separator) { - int ReducedBy; - char *d, *s, *end; /* dest, source */ - int count = 0; + const char *s; /* source */ + const char *EndBuffer; /* end stop of source buffer */ + int current_token = 0; /* token currently being processed */ + int len = 0; /* running total length of extracted string */ - /* Find desired @parameter */ - end = Source->buf + Source->BufUsed; - d = Source->buf; - while ((count < parmnum) && - (d <= end)) + if ((Source == NULL) || + (Source->BufUsed == 0) ) { - /* End of string, bail! */ - if (!*d) { - d = NULL; - break; - } - if (*d == separator) { - count++; - } - d++; + *pStart = StrBufNOTNULL; + if (dest != NULL) + FlushStrBuf(dest); + return -1; } - if ((d == NULL) || (d >= end)) - return 0; /* @Parameter not found */ + + EndBuffer = Source->buf + Source->BufUsed; - /* Find next @parameter */ - s = d; - while ((*s && *s != separator) && - (s <= end)) + if (dest != NULL) { - s++; + dest->buf[0] = '\0'; + dest->BufUsed = 0; + } + else + { + *pStart = EndBuffer + 1; + return -1; } - if (*s == separator) - s++; - ReducedBy = d - s; - /* Hack and slash */ - if (*s) { - memmove(d, s, Source->BufUsed - (s - Source->buf)); - Source->BufUsed += ReducedBy; + if (*pStart == NULL) + { + *pStart = Source->buf; /* we're starting to examine this buffer. */ } - else if (d == Source->buf) { - *d = 0; - Source->BufUsed = 0; - } - else { - *--d = 0; - Source->BufUsed += ReducedBy; - } - /* - while (*s) { - *d++ = *s++; - } - *d = 0; - */ - return ReducedBy; -} - - -/** - * @brief a string tokenizer - * @param dest Destination StringBuffer - * @param Source StringBuffer to read into - * @param parmnum n'th @parameter to extract - * @param separator tokenizer @param - * @returns -1 if not found, else length of token. - */ -int StrBufExtract_token(StrBuf *dest, const StrBuf *Source, int parmnum, char separator) -{ - const char *s, *e; //* source * / - int len = 0; //* running total length of extracted string * / - int current_token = 0; //* token currently being processed * / - - if (dest != NULL) { - dest->buf[0] = '\0'; - dest->BufUsed = 0; - } - else - return(-1); - - if ((Source == NULL) || (Source->BufUsed ==0)) { - return(-1); - } - s = Source->buf; - e = s + Source->BufUsed; - - //cit_backtrace(); - //lprintf (CTDL_DEBUG, "test >: n: %d sep: %c source: %s \n willi \n", parmnum, separator, source); - - while ((s= dest->BufSize) { - dest->BufUsed = len; - if (IncreaseBuf(dest, 1, -1) < 0) { - dest->BufUsed --; - break; - } - } - if ( (current_token == parmnum) && - (*s != separator)) { - dest->buf[len] = *s; - ++len; - } - else if (current_token > parmnum) { - break; - } - ++s; - } - - dest->buf[len] = '\0'; - dest->BufUsed = len; - - if (current_token < parmnum) { - //lprintf (CTDL_DEBUG,"test 0) - return(atoi(buf)); - else - return 0; -} - -/** - * @brief a string tokenizer to fetch a long integer - * @param dest Destination StringBuffer - * @param parmnum n'th @parameter to extract - * @param separator tokenizer @param - * @returns 0 if not found, else long integer representation of the token - */ -long StrBufExtract_long(const StrBuf* Source, int parmnum, char separator) -{ - StrBuf tmp; - char buf[64]; - - tmp.buf = buf; - buf[0] = '\0'; - tmp.BufSize = 64; - tmp.BufUsed = 0; - tmp.ConstBuf = 1; - if (StrBufExtract_token(&tmp, Source, parmnum, separator) > 0) - return(atoi(buf)); - else - return 0; -} - - -/** - * @brief a string tokenizer to fetch an unsigned long - * @param dest Destination StringBuffer - * @param parmnum n'th @parameter to extract - * @param separator tokenizer @param - * @returns 0 if not found, else unsigned long representation of the token - */ -unsigned long StrBufExtract_unsigned_long(const StrBuf* Source, int parmnum, char separator) -{ - StrBuf tmp; - char buf[64]; - char *pnum; - - tmp.buf = buf; - buf[0] = '\0'; - tmp.BufSize = 64; - tmp.BufUsed = 0; - tmp.ConstBuf = 1; - if (StrBufExtract_token(&tmp, Source, parmnum, separator) > 0) { - pnum = &buf[0]; - if (*pnum == '-') - pnum ++; - return (unsigned long) atol(pnum); - } - else - return 0; -} - - - -/** - * @briefa string tokenizer; Bounds checker - * function to make shure whether StrBufExtract_NextToken and friends have reached the end of the string. - * @param Source our tokenbuffer - * @param pStart the token iterator pointer to inspect - * @returns whether the revolving pointer is inside of the search range - */ -int StrBufHaveNextToken(const StrBuf *Source, const char **pStart) -{ - if ((Source == NULL) || - (*pStart == StrBufNOTNULL) || - (Source->BufUsed == 0)) - { - return 0; - } - if (*pStart == NULL) - { - return 1; - } - else if (*pStart > Source->buf + Source->BufUsed) - { - return 0; - } - else if (*pStart <= Source->buf) - { - return 0; - } - - return 1; -} - -/** - * @brief a string tokenizer - * @param dest Destination StringBuffer - * @param Source StringBuffer to read into - * @param pStart pointer to the end of the last token. Feed with NULL on start. - * @param separator tokenizer @param - * @returns -1 if not found, else length of token. - */ -int StrBufExtract_NextToken(StrBuf *dest, const StrBuf *Source, const char **pStart, char separator) -{ - const char *s; /* source */ - const char *EndBuffer; /* end stop of source buffer */ - int current_token = 0; /* token currently being processed */ - int len = 0; /* running total length of extracted string */ - - if ((Source == NULL) || - (Source->BufUsed == 0) ) - { - *pStart = StrBufNOTNULL; - return -1; - } - - EndBuffer = Source->buf + Source->BufUsed; - - if (dest != NULL) - { - dest->buf[0] = '\0'; - dest->BufUsed = 0; - } - else - { - *pStart = EndBuffer + 1; - return -1; - } - - if (*pStart == NULL) - { - *pStart = Source->buf; /* we're starting to examine this buffer. */ - } - else if ((*pStart < Source->buf) || - (*pStart > EndBuffer ) ) - { - return -1; /* no more tokens to find. */ + else if ((*pStart < Source->buf) || + (*pStart > EndBuffer ) ) + { + return -1; /* no more tokens to find. */ } s = *pStart; @@ -1643,11 +1624,12 @@ int StrBufExtract_NextToken(StrBuf *dest, const StrBuf *Source, const char **pSt /** + * @ingroup StrBuf_NextTokenizer * @brief a string tokenizer - * @param dest Destination StringBuffer - * @param Source StringBuffer to read into + * @param Source StringBuffer to read from * @param pStart pointer to the end of the last token. Feed with NULL. - * @param separator tokenizer @param + * @param separator tokenizer character + * @param nTokens number of tokens to fastforward over * @returns -1 if not found, else length of token. */ int StrBufSkip_NTokenS(const StrBuf *Source, const char **pStart, char separator, int nTokens) @@ -1679,7 +1661,7 @@ int StrBufSkip_NTokenS(const StrBuf *Source, const char **pStart, char separator //cit_backtrace(); //lprintf (CTDL_DEBUG, "test >: n: %d sep: %c source: %s \n willi \n", parmnum, separator, source); - while ((sBufUsed; - while (1) { - rlen = read(*fd, &buf->buf[len], 1); - if (rlen < 1) { - *Error = strerror(errno); - - close(*fd); - *fd = -1; - - return -1; - } - if (buf->buf[len] == '\n') - break; - if (buf->buf[len] != '\r') - len ++; - if (len + 2 >= buf->BufSize) { - buf->BufUsed = len; - buf->buf[len+1] = '\0'; - IncreaseBuf(buf, 1, -1); - } - } - buf->BufUsed = len; - buf->buf[len] = '\0'; - return len - slen; -} +/******************************************************************************* + * Escape Appending * + *******************************************************************************/ -/** - * @brief Read a line from socket - * flushes and closes the FD on error - * @param buf the buffer to get the input to - * @param fd pointer to the filedescriptor to read - * @param append Append to an existing string or replace? - * @param Error strerror() on error - * @returns numbers of chars read +/** + * @ingroup StrBuf_DeEnCoder + * @brief Escape a string for feeding out as a URL while appending it to a Buffer + * @param OutBuf the output buffer + * @param In Buffer to encode + * @param PlainIn way in from plain old c strings */ -int StrBufTCP_read_buffered_line(StrBuf *Line, - StrBuf *buf, - int *fd, - int timeout, - int selectresolution, - const char **Error) +void StrBufUrlescAppend(StrBuf *OutBuf, const StrBuf *In, const char *PlainIn) { - int len, rlen; - int nSuccessLess = 0; - fd_set rfds; - char *pch = NULL; - int fdflags; - int IsNonBlock; - struct timeval tv; - - if (buf->BufUsed > 0) { - pch = strchr(buf->buf, '\n'); - if (pch != NULL) { - rlen = 0; - len = pch - buf->buf; - if (len > 0 && (*(pch - 1) == '\r') ) - rlen ++; - StrBufSub(Line, buf, 0, len - rlen); - StrBufCutLeft(buf, len + 1); - return len - rlen; - } - } + const char *pch, *pche; + char *pt, *pte; + int len; - if (buf->BufSize - buf->BufUsed < 10) - IncreaseBuf(buf, 1, -1); + if (((In == NULL) && (PlainIn == NULL)) || (OutBuf == NULL) ) + return; + if (PlainIn != NULL) { + len = strlen(PlainIn); + pch = PlainIn; + pche = pch + len; + } + else { + pch = In->buf; + pche = pch + In->BufUsed; + len = In->BufUsed; + } - fdflags = fcntl(*fd, F_GETFL); - IsNonBlock = (fdflags & O_NONBLOCK) == O_NONBLOCK; + if (len == 0) + return; - while ((nSuccessLess < timeout) && (pch == NULL)) { - if (IsNonBlock){ - tv.tv_sec = selectresolution; - tv.tv_usec = 0; - - FD_ZERO(&rfds); - FD_SET(*fd, &rfds); - if (select(*fd + 1, NULL, &rfds, NULL, &tv) == -1) { - *Error = strerror(errno); - close (*fd); - *fd = -1; - return -1; - } - } - if (IsNonBlock && ! FD_ISSET(*fd, &rfds)) { - nSuccessLess ++; - continue; - } - rlen = read(*fd, - &buf->buf[buf->BufUsed], - buf->BufSize - buf->BufUsed - 1); - if (rlen < 1) { - *Error = strerror(errno); - close(*fd); - *fd = -1; - return -1; + pt = OutBuf->buf + OutBuf->BufUsed; + pte = OutBuf->buf + OutBuf->BufSize - 4; /**< we max append 3 chars at once plus the \0 */ + + while (pch < pche) { + if (pt >= pte) { + IncreaseBuf(OutBuf, 1, -1); + pte = OutBuf->buf + OutBuf->BufSize - 4; /**< we max append 3 chars at once plus the \0 */ + pt = OutBuf->buf + OutBuf->BufUsed; } - else if (rlen > 0) { - nSuccessLess = 0; - buf->BufUsed += rlen; - buf->buf[buf->BufUsed] = '\0'; - if (buf->BufUsed + 10 > buf->BufSize) { - IncreaseBuf(buf, 1, -1); - } - pch = strchr(buf->buf, '\n'); - continue; + + if((*pch >= 'a' && *pch <= 'z') || + (*pch >= '@' && *pch <= 'Z') || /* @ A-Z */ + (*pch >= '0' && *pch <= ':') || /* 0-9 : */ + (*pch == '!') || (*pch == '_') || + (*pch == ',') || (*pch == '.')) + { + *(pt++) = *(pch++); + OutBuf->BufUsed++; + } + else { + *pt = '%'; + *(pt + 1) = HexList[(unsigned char)*pch][0]; + *(pt + 2) = HexList[(unsigned char)*pch][1]; + pt += 3; + OutBuf->BufUsed += 3; + pch ++; } - - } - if (pch != NULL) { - rlen = 0; - len = pch - buf->buf; - if (len > 0 && (*(pch - 1) == '\r') ) - rlen ++; - StrBufSub(Line, buf, 0, len - rlen); - StrBufCutLeft(buf, len + 1); - return len - rlen; } - return -1; - + *pt = '\0'; } -static const char *ErrRBLF_SelectFailed="StrBufTCP_read_buffered_line_fast: Select failed without reason"; -static const char *ErrRBLF_NotEnoughSentFromServer="StrBufTCP_read_buffered_line_fast: No complete line was sent from peer"; -/** - * @brief Read a line from socket - * flushes and closes the FD on error - * @param buf the buffer to get the input to - * @param Pos pointer to the current read position, should be NULL initialized! - * @param fd pointer to the filedescriptor to read - * @param append Append to an existing string or replace? - * @param Error strerror() on error - * @returns numbers of chars read +/** + * @ingroup StrBuf_DeEnCoder + * @brief append a string in hex encoding to the buffer + * @param OutBuf the output buffer + * @param In Buffer to encode + * @param PlainIn way in from plain old c strings + * @param PlainInLen way in from plain old c strings; maybe you've got binary data or know the length? */ -int StrBufTCP_read_buffered_line_fast(StrBuf *Line, - StrBuf *IOBuf, - const char **Pos, - int *fd, - int timeout, - int selectresolution, - const char **Error) +void StrBufHexEscAppend(StrBuf *OutBuf, const StrBuf *In, const unsigned char *PlainIn, long PlainInLen) { - const char *pche = NULL; - const char *pos = NULL; - int len, rlen; - int nSuccessLess = 0; - fd_set rfds; - const char *pch = NULL; - int fdflags; - int IsNonBlock; - struct timeval tv; - - pos = *Pos; - if ((IOBuf->BufUsed > 0) && - (pos != NULL) && - (pos < IOBuf->buf + IOBuf->BufUsed)) - { - pche = IOBuf->buf + IOBuf->BufUsed; - pch = pos; - while ((pch < pche) && (*pch != '\n')) - pch ++; - if ((pch >= pche) || (*pch == '\0')) - pch = NULL; - if ((pch != NULL) && - (pch <= pche)) - { - rlen = 0; - len = pch - pos; - if (len > 0 && (*(pch - 1) == '\r') ) - rlen ++; - StrBufSub(Line, IOBuf, (pos - IOBuf->buf), len - rlen); - *Pos = pch + 1; - return len - rlen; - } - } + const unsigned char *pch, *pche; + char *pt, *pte; + int len; - if (pos != NULL) { - if (pos > pche) - FlushStrBuf(IOBuf); - else - StrBufCutLeft(IOBuf, (pos - IOBuf->buf)); - *Pos = NULL; + if (((In == NULL) && (PlainIn == NULL)) || (OutBuf == NULL) ) + return; + if (PlainIn != NULL) { + if (PlainInLen < 0) + len = strlen((const char*)PlainIn); + else + len = PlainInLen; + pch = PlainIn; + pche = pch + len; } - - if (IOBuf->BufSize - IOBuf->BufUsed < 10) { - IncreaseBuf(IOBuf, 1, -1); + else { + pch = (const unsigned char*)In->buf; + pche = pch + In->BufUsed; + len = In->BufUsed; } - fdflags = fcntl(*fd, F_GETFL); - IsNonBlock = (fdflags & O_NONBLOCK) == O_NONBLOCK; + if (len == 0) + return; - pch = NULL; - while ((nSuccessLess < timeout) && - (pch == NULL) && - (*fd != -1)) { - if (IsNonBlock) - { - tv.tv_sec = 1; - tv.tv_usec = 0; - - FD_ZERO(&rfds); - FD_SET(*fd, &rfds); - if (select((*fd) + 1, &rfds, NULL, NULL, &tv) == -1) { - *Error = strerror(errno); - close (*fd); - *fd = -1; - if (*Error == NULL) - *Error = ErrRBLF_SelectFailed; - return -1; - } - if (! FD_ISSET(*fd, &rfds) != 0) { - nSuccessLess ++; - continue; - } - } - rlen = read(*fd, - &IOBuf->buf[IOBuf->BufUsed], - IOBuf->BufSize - IOBuf->BufUsed - 1); - if (rlen < 1) { - *Error = strerror(errno); - close(*fd); - *fd = -1; - return -1; - } - else if (rlen > 0) { - nSuccessLess = 0; - IOBuf->BufUsed += rlen; - IOBuf->buf[IOBuf->BufUsed] = '\0'; - if (IOBuf->BufUsed + 10 > IOBuf->BufSize) { - IncreaseBuf(IOBuf, 1, -1); - } - - pche = IOBuf->buf + IOBuf->BufUsed; - pch = IOBuf->buf; - while ((pch < pche) && (*pch != '\n')) - pch ++; - if ((pch >= pche) || (*pch == '\0')) - pch = NULL; - continue; + pt = OutBuf->buf + OutBuf->BufUsed; + pte = OutBuf->buf + OutBuf->BufSize - 3; /**< we max append 3 chars at once plus the \0 */ + + while (pch < pche) { + if (pt >= pte) { + IncreaseBuf(OutBuf, 1, -1); + pte = OutBuf->buf + OutBuf->BufSize - 3; /**< we max append 3 chars at once plus the \0 */ + pt = OutBuf->buf + OutBuf->BufUsed; } - } - if (pch != NULL) { - pos = IOBuf->buf; - rlen = 0; - len = pch - pos; - if (len > 0 && (*(pch - 1) == '\r') ) - rlen ++; - StrBufSub(Line, IOBuf, 0, len - rlen); - *Pos = pos + len + 1; - return len - rlen; - } - *Error = ErrRBLF_NotEnoughSentFromServer; - return -1; + *pt = HexList[*pch][0]; + pt ++; + *pt = HexList[*pch][1]; + pt ++; pch ++; OutBuf->BufUsed += 2; + } + *pt = '\0'; +} + +/** + * @ingroup StrBuf_DeEnCoder + * @brief append a string in hex encoding to the buffer + * @param OutBuf the output buffer + * @param In Buffer to encode + * @param PlainIn way in from plain old c strings + */ +void StrBufHexescAppend(StrBuf *OutBuf, const StrBuf *In, const char *PlainIn) +{ + StrBufHexEscAppend(OutBuf, In, (const unsigned char*) PlainIn, -1); +} + +/** + * @ingroup StrBuf_DeEnCoder + * @brief Append a string, escaping characters which have meaning in HTML. + * + * @param Target target buffer + * @param Source source buffer; set to NULL if you just have a C-String + * @param PlainIn Plain-C string to append; set to NULL if unused + * @param nbsp If nonzero, spaces are converted to non-breaking spaces. + * @param nolinebreaks if set to 1, linebreaks are removed from the string. + * if set to 2, linebreaks are replaced by <br/> + */ +long StrEscAppend(StrBuf *Target, const StrBuf *Source, const char *PlainIn, int nbsp, int nolinebreaks) +{ + const char *aptr, *eiptr; + char *bptr, *eptr; + long len; + + if (((Source == NULL) && (PlainIn == NULL)) || (Target == NULL) ) + return -1; + + if (PlainIn != NULL) { + aptr = PlainIn; + len = strlen(PlainIn); + eiptr = aptr + len; + } + else { + aptr = Source->buf; + eiptr = aptr + Source->BufUsed; + len = Source->BufUsed; + } + + if (len == 0) + return -1; + + bptr = Target->buf + Target->BufUsed; + eptr = Target->buf + Target->BufSize - 11; /* our biggest unit to put in... */ + + while (aptr < eiptr){ + if(bptr >= eptr) { + IncreaseBuf(Target, 1, -1); + eptr = Target->buf + Target->BufSize - 11; /* our biggest unit to put in... */ + bptr = Target->buf + Target->BufUsed; + } + if (*aptr == '<') { + memcpy(bptr, "<", 4); + bptr += 4; + Target->BufUsed += 4; + } + else if (*aptr == '>') { + memcpy(bptr, ">", 4); + bptr += 4; + Target->BufUsed += 4; + } + else if (*aptr == '&') { + memcpy(bptr, "&", 5); + bptr += 5; + Target->BufUsed += 5; + } + else if (*aptr == '"') { + memcpy(bptr, """, 6); + bptr += 6; + Target->BufUsed += 6; + } + else if (*aptr == '\'') { + memcpy(bptr, "'", 5); + bptr += 5; + Target->BufUsed += 5; + } + else if (*aptr == LB) { + *bptr = '<'; + bptr ++; + Target->BufUsed ++; + } + else if (*aptr == RB) { + *bptr = '>'; + bptr ++; + Target->BufUsed ++; + } + else if (*aptr == QU) { + *bptr ='"'; + bptr ++; + Target->BufUsed ++; + } + else if ((*aptr == 32) && (nbsp == 1)) { + memcpy(bptr, " ", 6); + bptr += 6; + Target->BufUsed += 6; + } + else if ((*aptr == '\n') && (nolinebreaks == 1)) { + *bptr='\0'; /* nothing */ + } + else if ((*aptr == '\n') && (nolinebreaks == 2)) { + memcpy(bptr, "<br/>", 11); + bptr += 11; + Target->BufUsed += 11; + } + + + else if ((*aptr == '\r') && (nolinebreaks != 0)) { + *bptr='\0'; /* nothing */ + } + else{ + *bptr = *aptr; + bptr++; + Target->BufUsed ++; + } + aptr ++; + } + *bptr = '\0'; + if ((bptr = eptr - 1 ) && !IsEmptyStr(aptr) ) + return -1; + return Target->BufUsed; +} + +/** + * @ingroup StrBuf_DeEnCoder + * @brief Append a string, escaping characters which have meaning in HTML. + * Converts linebreaks into blanks; escapes single quotes + * @param Target target buffer + * @param Source source buffer; set to NULL if you just have a C-String + * @param PlainIn Plain-C string to append; set to NULL if unused + */ +void StrMsgEscAppend(StrBuf *Target, const StrBuf *Source, const char *PlainIn) +{ + const char *aptr, *eiptr; + char *tptr, *eptr; + long len; + + if (((Source == NULL) && (PlainIn == NULL)) || (Target == NULL) ) + return ; + + if (PlainIn != NULL) { + aptr = PlainIn; + len = strlen(PlainIn); + eiptr = aptr + len; + } + else { + aptr = Source->buf; + eiptr = aptr + Source->BufUsed; + len = Source->BufUsed; + } + + if (len == 0) + return; + + eptr = Target->buf + Target->BufSize - 8; + tptr = Target->buf + Target->BufUsed; + + while (aptr < eiptr){ + if(tptr >= eptr) { + IncreaseBuf(Target, 1, -1); + eptr = Target->buf + Target->BufSize - 8; + tptr = Target->buf + Target->BufUsed; + } + + if (*aptr == '\n') { + *tptr = ' '; + Target->BufUsed++; + } + else if (*aptr == '\r') { + *tptr = ' '; + Target->BufUsed++; + } + else if (*aptr == '\'') { + *(tptr++) = '&'; + *(tptr++) = '#'; + *(tptr++) = '3'; + *(tptr++) = '9'; + *tptr = ';'; + Target->BufUsed += 5; + } else { + *tptr = *aptr; + Target->BufUsed++; + } + tptr++; aptr++; + } + *tptr = '\0'; +} + + + +/** + * @ingroup StrBuf_DeEnCoder + * @brief Append a string, escaping characters which have meaning in ICAL. + * [\n,] + * @param Target target buffer + * @param Source source buffer; set to NULL if you just have a C-String + * @param PlainIn Plain-C string to append; set to NULL if unused + */ +void StrIcalEscAppend(StrBuf *Target, const StrBuf *Source, const char *PlainIn) +{ + const char *aptr, *eiptr; + char *tptr, *eptr; + long len; + + if (((Source == NULL) && (PlainIn == NULL)) || (Target == NULL) ) + return ; + + if (PlainIn != NULL) { + aptr = PlainIn; + len = strlen(PlainIn); + eiptr = aptr + len; + } + else { + aptr = Source->buf; + eiptr = aptr + Source->BufUsed; + len = Source->BufUsed; + } + + if (len == 0) + return; + + eptr = Target->buf + Target->BufSize - 8; + tptr = Target->buf + Target->BufUsed; + + while (aptr < eiptr){ + if(tptr + 3 >= eptr) { + IncreaseBuf(Target, 1, -1); + eptr = Target->buf + Target->BufSize - 8; + tptr = Target->buf + Target->BufUsed; + } + + if (*aptr == '\n') { + *tptr = '\\'; + Target->BufUsed++; + tptr++; + *tptr = 'n'; + Target->BufUsed++; + } + else if (*aptr == '\r') { + *tptr = '\\'; + Target->BufUsed++; + tptr++; + *tptr = 'r'; + Target->BufUsed++; + } + else if (*aptr == ',') { + *tptr = '\\'; + Target->BufUsed++; + tptr++; + *tptr = ','; + Target->BufUsed++; + } else { + *tptr = *aptr; + Target->BufUsed++; + } + tptr++; aptr++; + } + *tptr = '\0'; +} + +/** + * @ingroup StrBuf_DeEnCoder + * @brief Append a string, escaping characters which have meaning in JavaScript strings . + * + * @param Target target buffer + * @param Source source buffer; set to NULL if you just have a C-String + * @param PlainIn Plain-C string to append; set to NULL if unused + * @returns size of result or -1 + */ +long StrECMAEscAppend(StrBuf *Target, const StrBuf *Source, const char *PlainIn) +{ + const char *aptr, *eiptr; + char *bptr, *eptr; + long len; + int IsUtf8Sequence; + + if (((Source == NULL) && (PlainIn == NULL)) || (Target == NULL) ) + return -1; + + if (PlainIn != NULL) { + aptr = PlainIn; + len = strlen(PlainIn); + eiptr = aptr + len; + } + else { + aptr = Source->buf; + eiptr = aptr + Source->BufUsed; + len = Source->BufUsed; + } + + if (len == 0) + return -1; + + bptr = Target->buf + Target->BufUsed; + eptr = Target->buf + Target->BufSize - 7; /* our biggest unit to put in... */ + + while (aptr < eiptr){ + if(bptr >= eptr) { + IncreaseBuf(Target, 1, -1); + eptr = Target->buf + Target->BufSize - 7; /* our biggest unit to put in... */ + bptr = Target->buf + Target->BufUsed; + } + switch (*aptr) { + case '\n': + memcpy(bptr, HKEY("\\n")); + bptr += 2; + Target->BufUsed += 2; + break; + case '\r': + memcpy(bptr, HKEY("\\r")); + bptr += 2; + Target->BufUsed += 2; + break; + case '"': + *bptr = '\\'; + bptr ++; + *bptr = '"'; + bptr ++; + Target->BufUsed += 2; + break; + case '\\': + if ((*(aptr + 1) == 'u') && + isxdigit(*(aptr + 2)) && + isxdigit(*(aptr + 3)) && + isxdigit(*(aptr + 4)) && + isxdigit(*(aptr + 5))) + { /* oh, a unicode escaper. let it pass through. */ + memcpy(bptr, aptr, 6); + aptr += 5; + bptr +=6; + Target->BufUsed += 6; + } + else + { + *bptr = '\\'; + bptr ++; + *bptr = '\\'; + bptr ++; + Target->BufUsed += 2; + } + break; + case '\b': + *bptr = '\\'; + bptr ++; + *bptr = 'b'; + bptr ++; + Target->BufUsed += 2; + break; + case '\f': + *bptr = '\\'; + bptr ++; + *bptr = 'f'; + bptr ++; + Target->BufUsed += 2; + break; + case '\t': + *bptr = '\\'; + bptr ++; + *bptr = 't'; + bptr ++; + Target->BufUsed += 2; + break; + default: + IsUtf8Sequence = Ctdl_GetUtf8SequenceLength(aptr, eiptr); + while (IsUtf8Sequence > 0){ + *bptr = *aptr; + Target->BufUsed ++; + if (--IsUtf8Sequence) + aptr++; + bptr++; + } + } + aptr ++; + } + *bptr = '\0'; + if ((bptr == eptr - 1 ) && !IsEmptyStr(aptr) ) + return -1; + return Target->BufUsed; +} + +/** + * @ingroup StrBuf_DeEnCoder + * @brief Append a string, escaping characters which have meaning in HTML + json. + * + * @param Target target buffer + * @param Source source buffer; set to NULL if you just have a C-String + * @param PlainIn Plain-C string to append; set to NULL if unused + * @param nbsp If nonzero, spaces are converted to non-breaking spaces. + * @param nolinebreaks if set to 1, linebreaks are removed from the string. + * if set to 2, linebreaks are replaced by <br/> + */ +long StrHtmlEcmaEscAppend(StrBuf *Target, const StrBuf *Source, const char *PlainIn, int nbsp, int nolinebreaks) +{ + const char *aptr, *eiptr; + char *bptr, *eptr; + long len; + int IsUtf8Sequence = 0; + + if (((Source == NULL) && (PlainIn == NULL)) || (Target == NULL) ) + return -1; + + if (PlainIn != NULL) { + aptr = PlainIn; + len = strlen(PlainIn); + eiptr = aptr + len; + } + else { + aptr = Source->buf; + eiptr = aptr + Source->BufUsed; + len = Source->BufUsed; + } + + if (len == 0) + return -1; + + bptr = Target->buf + Target->BufUsed; + eptr = Target->buf + Target->BufSize - 11; /* our biggest unit to put in... */ + + while (aptr < eiptr){ + if(bptr >= eptr) { + IncreaseBuf(Target, 1, -1); + eptr = Target->buf + Target->BufSize - 11; /* our biggest unit to put in... */ + bptr = Target->buf + Target->BufUsed; + } + switch (*aptr) { + case '<': + memcpy(bptr, HKEY("<")); + bptr += 4; + Target->BufUsed += 4; + break; + case '>': + memcpy(bptr, HKEY(">")); + bptr += 4; + Target->BufUsed += 4; + break; + case '&': + memcpy(bptr, HKEY("&")); + bptr += 5; + Target->BufUsed += 5; + break; + case LB: + *bptr = '<'; + bptr ++; + Target->BufUsed ++; + break; + case RB: + *bptr = '>'; + bptr ++; + Target->BufUsed ++; + break; + case '\n': + switch (nolinebreaks) { + case 1: + *bptr='\0'; /* nothing */ + break; + case 2: + memcpy(bptr, HKEY("<br/>")); + bptr += 11; + Target->BufUsed += 11; + break; + default: + memcpy(bptr, HKEY("\\n")); + bptr += 2; + Target->BufUsed += 2; + } + break; + case '\r': + switch (nolinebreaks) { + case 1: + case 2: + *bptr='\0'; /* nothing */ + break; + default: + memcpy(bptr, HKEY("\\r")); + bptr += 2; + Target->BufUsed += 2; + break; + } + break; + case '"': + case QU: + *bptr = '\\'; + bptr ++; + *bptr = '"'; + bptr ++; + Target->BufUsed += 2; + break; + case '\\': + if ((*(aptr + 1) == 'u') && + isxdigit(*(aptr + 2)) && + isxdigit(*(aptr + 3)) && + isxdigit(*(aptr + 4)) && + isxdigit(*(aptr + 5))) + { /* oh, a unicode escaper. let it pass through. */ + memcpy(bptr, aptr, 6); + aptr += 5; + bptr +=6; + Target->BufUsed += 6; + } + else + { + *bptr = '\\'; + bptr ++; + *bptr = '\\'; + bptr ++; + Target->BufUsed += 2; + } + break; + case '\b': + *bptr = '\\'; + bptr ++; + *bptr = 'b'; + bptr ++; + Target->BufUsed += 2; + break; + case '\f': + *bptr = '\\'; + bptr ++; + *bptr = 'f'; + bptr ++; + Target->BufUsed += 2; + break; + case '\t': + *bptr = '\\'; + bptr ++; + *bptr = 't'; + bptr ++; + Target->BufUsed += 2; + break; + case 32: + if (nbsp == 1) { + memcpy(bptr, HKEY(" ")); + bptr += 6; + Target->BufUsed += 6; + break; + } + default: + IsUtf8Sequence = Ctdl_GetUtf8SequenceLength(aptr, eiptr); + while (IsUtf8Sequence > 0){ + *bptr = *aptr; + Target->BufUsed ++; + if (--IsUtf8Sequence) + aptr++; + bptr++; + } + } + aptr ++; + } + *bptr = '\0'; + if ((bptr = eptr - 1 ) && !IsEmptyStr(aptr) ) + return -1; + return Target->BufUsed; +} + +/** + * @ingroup StrBuf_DeEnCoder + * @brief unhide special chars hidden to the HTML escaper + * @param target buffer to put the unescaped string in + * @param source buffer to unescape + */ +void StrBufEUid_unescapize(StrBuf *target, const StrBuf *source) +{ + int a, b, len; + char hex[3]; + + if (target != NULL) + FlushStrBuf(target); + + if (source == NULL ||target == NULL) + { + return; + } + + len = source->BufUsed; + for (a = 0; a < len; ++a) { + if (target->BufUsed >= target->BufSize) + IncreaseBuf(target, 1, -1); + + if (source->buf[a] == '=') { + hex[0] = source->buf[a + 1]; + hex[1] = source->buf[a + 2]; + hex[2] = 0; + b = 0; + sscanf(hex, "%02x", &b); + target->buf[target->BufUsed] = b; + target->buf[++target->BufUsed] = 0; + a += 2; + } + else { + target->buf[target->BufUsed] = source->buf[a]; + target->buf[++target->BufUsed] = 0; + } + } +} + + +/** + * @ingroup StrBuf_DeEnCoder + * @brief hide special chars from the HTML escapers and friends + * @param target buffer to put the escaped string in + * @param source buffer to escape + */ +void StrBufEUid_escapize(StrBuf *target, const StrBuf *source) +{ + int i, len; + + if (target != NULL) + FlushStrBuf(target); + + if (source == NULL ||target == NULL) + { + return; + } + + len = source->BufUsed; + for (i=0; iBufUsed + 4 >= target->BufSize) + IncreaseBuf(target, 1, -1); + if ( (isalnum(source->buf[i])) || + (source->buf[i]=='-') || + (source->buf[i]=='_') ) { + target->buf[target->BufUsed++] = source->buf[i]; + } + else { + sprintf(&target->buf[target->BufUsed], + "=%02X", + (0xFF &source->buf[i])); + target->BufUsed += 3; + } + } + target->buf[target->BufUsed + 1] = '\0'; +} + + +/******************************************************************************* + * Quoted Printable de/encoding * + *******************************************************************************/ + +/** + * @ingroup StrBuf_DeEnCoder + * @brief decode a buffer from base 64 encoding; destroys original + * @param Buf Buffor to transform + */ +int StrBufDecodeBase64(StrBuf *Buf) +{ + char *xferbuf; + size_t siz; + if (Buf == NULL) return -1; + + xferbuf = (char*) malloc(Buf->BufSize); + *xferbuf = '\0'; + siz = CtdlDecodeBase64(xferbuf, + Buf->buf, + Buf->BufUsed); + free(Buf->buf); + Buf->buf = xferbuf; + Buf->BufUsed = siz; + return siz; } /** - * @brief Input binary data from socket - * flushes and closes the FD on error - * @param buf the buffer to get the input to - * @param fd pointer to the filedescriptor to read - * @param append Append to an existing string or replace? - * @param nBytes the maximal number of bytes to read - * @param Error strerror() on error - * @returns numbers of chars read + * @ingroup StrBuf_DeEnCoder + * @brief decode a buffer from base 64 encoding; destroys original + * @param Buf Buffor to transform */ -int StrBufReadBLOB(StrBuf *Buf, int *fd, int append, long nBytes, const char **Error) +int StrBufDecodeHex(StrBuf *Buf) { - int fdflags; - int len, rlen, slen; - int nSuccessLess; - int nRead = 0; - char *ptr; - int IsNonBlock; - struct timeval tv; - fd_set rfds; - if ((Buf == NULL) || (*fd == -1)) + unsigned int ch; + char *pch, *pche, *pchi; + + if (Buf == NULL) return -1; + + pch = pchi = Buf->buf; + pche = pch + Buf->BufUsed; + + while (pchi < pche){ + ch = decode_hex(pchi); + *pch = ch; + pch ++; + pchi += 2; + } + + *pch = '\0'; + Buf->BufUsed = pch - Buf->buf; + return Buf->BufUsed; +} + +/** + * @ingroup StrBuf_DeEnCoder + * @brief replace all chars >0x20 && < 0x7F with Mute + * @param Mute char to put over invalid chars + * @param Buf Buffor to transform + */ +int StrBufSanitizeAscii(StrBuf *Buf, const char Mute) +{ + unsigned char *pch; + + if (Buf == NULL) return -1; + pch = (unsigned char *)Buf->buf; + while (pch < (unsigned char *)Buf->buf + Buf->BufUsed) { + if ((*pch < 0x20) || (*pch > 0x7F)) + *pch = Mute; + pch ++; + } + return Buf->BufUsed; +} + + +/** + * @ingroup StrBuf_DeEnCoder + * @brief remove escaped strings from i.e. the url string (like %20 for blanks) + * @param Buf Buffer to translate + * @param StripBlanks Reduce several blanks to one? + */ +long StrBufUnescape(StrBuf *Buf, int StripBlanks) +{ + int a, b; + char hex[3]; + long len; + + if (Buf == NULL) return -1; - if (!append) - FlushStrBuf(Buf); - if (Buf->BufUsed + nBytes >= Buf->BufSize) - IncreaseBuf(Buf, 1, Buf->BufUsed + nBytes); - ptr = Buf->buf + Buf->BufUsed; + while ((Buf->BufUsed > 0) && (isspace(Buf->buf[Buf->BufUsed - 1]))){ + Buf->buf[Buf->BufUsed - 1] = '\0'; + Buf->BufUsed --; + } + + a = 0; + while (a < Buf->BufUsed) { + if (Buf->buf[a] == '+') + Buf->buf[a] = ' '; + else if (Buf->buf[a] == '%') { + /* don't let % chars through, rather truncate the input. */ + if (a + 2 > Buf->BufUsed) { + Buf->buf[a] = '\0'; + Buf->BufUsed = a; + } + else { + hex[0] = Buf->buf[a + 1]; + hex[1] = Buf->buf[a + 2]; + hex[2] = 0; + b = 0; + sscanf(hex, "%02x", &b); + Buf->buf[a] = (char) b; + len = Buf->BufUsed - a - 2; + if (len > 0) + memmove(&Buf->buf[a + 1], &Buf->buf[a + 3], len); + + Buf->BufUsed -=2; + } + } + a++; + } + return a; +} + + +/** + * @ingroup StrBuf_DeEnCoder + * @brief RFC2047-encode a header field if necessary. + * If no non-ASCII characters are found, the string + * will be copied verbatim without encoding. + * + * @param target Target buffer. + * @param source Source string to be encoded. + * @returns encoded length; -1 if non success. + */ +int StrBufRFC2047encode(StrBuf **target, const StrBuf *source) +{ + const char headerStr[] = "=?UTF-8?Q?"; + int need_to_encode = 0; + int i = 0; + unsigned char ch; + + if ((source == NULL) || + (target == NULL)) + return -1; + + while ((i < source->BufUsed) && + (!IsEmptyStr (&source->buf[i])) && + (need_to_encode == 0)) { + if (((unsigned char) source->buf[i] < 32) || + ((unsigned char) source->buf[i] > 126)) { + need_to_encode = 1; + } + i++; + } + + if (!need_to_encode) { + if (*target == NULL) { + *target = NewStrBufPlain(source->buf, source->BufUsed); + } + else { + FlushStrBuf(*target); + StrBufAppendBuf(*target, source, 0); + } + return (*target)->BufUsed; + } + if (*target == NULL) + *target = NewStrBufPlain(NULL, sizeof(headerStr) + source->BufUsed * 2); + else if (sizeof(headerStr) + source->BufUsed >= (*target)->BufSize) + IncreaseBuf(*target, sizeof(headerStr) + source->BufUsed, 0); + memcpy ((*target)->buf, headerStr, sizeof(headerStr) - 1); + (*target)->BufUsed = sizeof(headerStr) - 1; + for (i=0; (i < source->BufUsed); ++i) { + if ((*target)->BufUsed + 4 >= (*target)->BufSize) + IncreaseBuf(*target, 1, 0); + ch = (unsigned char) source->buf[i]; + if ((ch < 32) || + (ch > 126) || + (ch == 61) || + (ch == '=') || + (ch == '?') || + (ch == '_') || + (ch == '[') || + (ch == ']') ) + { + sprintf(&(*target)->buf[(*target)->BufUsed], "=%02X", ch); + (*target)->BufUsed += 3; + } + else { + if (ch == ' ') + (*target)->buf[(*target)->BufUsed] = '_'; + else + (*target)->buf[(*target)->BufUsed] = ch; + (*target)->BufUsed++; + } + } + + if ((*target)->BufUsed + 4 >= (*target)->BufSize) + IncreaseBuf(*target, 1, 0); + + (*target)->buf[(*target)->BufUsed++] = '?'; + (*target)->buf[(*target)->BufUsed++] = '='; + (*target)->buf[(*target)->BufUsed] = '\0'; + return (*target)->BufUsed;; +} + + + +static void AddRecipient(StrBuf *Target, + StrBuf *UserName, + StrBuf *EmailAddress, + StrBuf *EncBuf) +{ + int QuoteMe = 0; + + if (StrLength(Target) > 0) StrBufAppendBufPlain(Target, HKEY(", "), 0); + if (strchr(ChrPtr(UserName), ',') != NULL) QuoteMe = 1; + + if (QuoteMe) StrBufAppendBufPlain(Target, HKEY("\""), 0); + StrBufRFC2047encode(&EncBuf, UserName); + StrBufAppendBuf(Target, EncBuf, 0); + if (QuoteMe) StrBufAppendBufPlain(Target, HKEY("\" "), 0); + else StrBufAppendBufPlain(Target, HKEY(" "), 0); + + if (StrLength(EmailAddress) > 0){ + StrBufAppendBufPlain(Target, HKEY("<"), 0); + StrBufAppendBuf(Target, EmailAddress, 0); /* TODO: what about IDN???? */ + StrBufAppendBufPlain(Target, HKEY(">"), 0); + } +} + + +/** + * \brief QP encode parts of an email TO/CC/BCC vector, and strip/filter invalid parts + * \param Recp Source list of email recipients + * \param UserName Temporary buffer for internal use; Please provide valid buffer. + * \param EmailAddress Temporary buffer for internal use; Please provide valid buffer. + * \param EncBuf Temporary buffer for internal use; Please provide valid buffer. + * \returns encoded & sanitized buffer with the contents of Recp; Caller owns this memory. + */ +StrBuf *StrBufSanitizeEmailRecipientVector(const StrBuf *Recp, + StrBuf *UserName, + StrBuf *EmailAddress, + StrBuf *EncBuf) +{ + StrBuf *Target; + const char *pch, *pche; + const char *UserStart, *UserEnd, *EmailStart, *EmailEnd, *At; + + if ((Recp == NULL) || (StrLength(Recp) == 0)) + return NULL; + + pch = ChrPtr(Recp); + pche = pch + StrLength(Recp); + + if (!CheckEncode(pch, -1, pche)) + return NewStrBufDup(Recp); + + Target = NewStrBufPlain(NULL, StrLength(Recp)); + + while ((pch != NULL) && (pch < pche)) + { + while (isspace(*pch)) pch++; + UserStart = UserEnd = EmailStart = EmailEnd = NULL; + + if ((*pch == '"') || (*pch == '\'')) { + UserStart = pch + 1; + + UserEnd = strchr(UserStart, *pch); + if (UserEnd == NULL) + break; ///TODO: Userfeedback?? + EmailStart = UserEnd + 1; + while (isspace(*EmailStart)) + EmailStart++; + if (UserEnd == UserStart) { + UserStart = UserEnd = NULL; + } + + if (*EmailStart == '<') { + EmailStart++; + EmailEnd = strchr(EmailStart, '>'); + if (EmailEnd == NULL) + EmailEnd = strchr(EmailStart, ','); + + } + else { + EmailEnd = strchr(EmailStart, ','); + } + if (EmailEnd == NULL) + EmailEnd = pche; + pch = EmailEnd + 1; + } + else { + int gt = 0; + UserStart = pch; + EmailEnd = strchr(UserStart, ','); + if (EmailEnd == NULL) { + EmailEnd = strchr(pch, '>'); + pch = NULL; + if (EmailEnd != NULL) { + gt = 1; + } + else { + EmailEnd = pche; + } + } + else { + + pch = EmailEnd + 1; + while ((EmailEnd > UserStart) && !gt && + ((*EmailEnd == ',') || + (*EmailEnd == '>') || + (isspace(*EmailEnd)))) + { + if (*EmailEnd == '>') + gt = 1; + else + EmailEnd--; + } + if (EmailEnd == UserStart) + break; + } + if (gt) { + EmailStart = strchr(UserStart, '<'); + if ((EmailStart == NULL) || (EmailStart > EmailEnd)) + break; + UserEnd = EmailStart; + + while ((UserEnd > UserStart) && + isspace (*(UserEnd - 1))) + UserEnd --; + EmailStart ++; + if (UserStart >= UserEnd) + UserStart = UserEnd = NULL; + At = strchr(EmailStart, '@'); + } + else { /* this is a local recipient... no domain, just a realname */ + EmailStart = UserStart; + At = strchr(EmailStart, '@'); + if (At == NULL) { + UserEnd = EmailEnd; + EmailEnd = NULL; + } + else { + EmailStart = UserStart; + UserStart = NULL; + } + } + } + + if ((UserStart != NULL) && (UserEnd != NULL)) + StrBufPlain(UserName, UserStart, UserEnd - UserStart); + else if ((UserStart != NULL) && (UserEnd == NULL)) + StrBufPlain(UserName, UserStart, UserEnd - UserStart); + else + FlushStrBuf(UserName); + + if ((EmailStart != NULL) && (EmailEnd != NULL)) + StrBufPlain(EmailAddress, EmailStart, EmailEnd - EmailStart); + else if ((EmailStart != NULL) && (EmailEnd == NULL)) + StrBufPlain(EmailAddress, EmailStart, EmailEnd - pche); + else + FlushStrBuf(EmailAddress); + + AddRecipient(Target, UserName, EmailAddress, EncBuf); + + if (pch == NULL) + break; + + if ((pch != NULL) && (*pch == ',')) + pch ++; + if (pch != NULL) while (isspace(*pch)) + pch ++; + } + return Target; +} + + +/** + * @ingroup StrBuf + * @brief replaces all occurances of 'search' by 'replace' + * @param buf Buffer to modify + * @param search character to search + * @param replace character to replace search by + */ +void StrBufReplaceChars(StrBuf *buf, char search, char replace) +{ + long i; + if (buf == NULL) + return; + for (i=0; iBufUsed; i++) + if (buf->buf[i] == search) + buf->buf[i] = replace; + +} - slen = len = Buf->BufUsed; +/** + * @ingroup StrBuf + * @brief removes all \\r s from the string, or replaces them with \n if its not a combination of both. + * @param buf Buffer to modify + */ +void StrBufToUnixLF(StrBuf *buf) +{ + char *pche, *pchS, *pchT; + if (buf == NULL) + return; - fdflags = fcntl(*fd, F_GETFL); - IsNonBlock = (fdflags & O_NONBLOCK) == O_NONBLOCK; - nSuccessLess = 0; - while ((nRead < nBytes) && - (*fd != -1)) + pche = buf->buf + buf->BufUsed; + pchS = pchT = buf->buf; + while (pchS < pche) { - if (IsNonBlock) + if (*pchS == '\r') { - tv.tv_sec = 1; - tv.tv_usec = 0; - - FD_ZERO(&rfds); - FD_SET(*fd, &rfds); - if (select(*fd + 1, &rfds, NULL, NULL, &tv) == -1) { - *Error = strerror(errno); - close (*fd); - *fd = -1; - if (*Error == NULL) - *Error = ErrRBLF_SelectFailed; - return -1; - } - if (! FD_ISSET(*fd, &rfds) != 0) { - nSuccessLess ++; - continue; + pchS ++; + if (*pchS != '\n') { + *pchT = '\n'; + pchT++; } } - - if ((rlen = read(*fd, - ptr, - nBytes - nRead)) == -1) { - close(*fd); - *fd = -1; - *Error = strerror(errno); - return rlen; - } - nRead += rlen; - ptr += rlen; - Buf->BufUsed += rlen; + *pchT = *pchS; + pchT++; pchS++; } - Buf->buf[Buf->BufUsed] = '\0'; - return nRead; + *pchT = '\0'; + buf->BufUsed = pchT - buf->buf; } -const char *ErrRBB_too_many_selects = "StrBufReadBLOBBuffered: to many selects; aborting."; + +/******************************************************************************* + * Iconv Wrapper; RFC822 de/encoding * + *******************************************************************************/ + /** - * @brief Input binary data from socket - * flushes and closes the FD on error - * @param buf the buffer to get the input to - * @param fd pointer to the filedescriptor to read - * @param append Append to an existing string or replace? - * @param nBytes the maximal number of bytes to read - * @param Error strerror() on error - * @returns numbers of chars read + * @ingroup StrBuf_DeEnCoder + * @brief Wrapper around iconv_open() + * Our version adds aliases for non-standard Microsoft charsets + * such as 'MS950', aliasing them to names like 'CP950' + * + * @param tocode Target encoding + * @param fromcode Source encoding + * @param pic anonimized pointer to iconv struct */ -int StrBufReadBLOBBuffered(StrBuf *Blob, - StrBuf *IOBuf, - const char **Pos, - int *fd, - int append, - long nBytes, - int check, - const char **Error) +void ctdl_iconv_open(const char *tocode, const char *fromcode, void *pic) { - const char *pche; - const char *pos; - int nSelects = 0; - int SelRes; - int fdflags; - int len = 0; - int rlen, slen; - int nRead = 0; - int nAlreadyRead = 0; - int IsNonBlock; - char *ptr; - fd_set rfds; - const char *pch; - struct timeval tv; - int nSuccessLess; +#ifdef HAVE_ICONV + iconv_t ic = (iconv_t)(-1) ; + ic = iconv_open(tocode, fromcode); + if (ic == (iconv_t)(-1) ) { + char alias_fromcode[64]; + if ( (strlen(fromcode) == 5) && (!strncasecmp(fromcode, "MS", 2)) ) { + safestrncpy(alias_fromcode, fromcode, sizeof alias_fromcode); + alias_fromcode[0] = 'C'; + alias_fromcode[1] = 'P'; + ic = iconv_open(tocode, alias_fromcode); + } + } + *(iconv_t *)pic = ic; +#endif +} - if ((Blob == NULL) || (*fd == -1) || (IOBuf == NULL) || (Pos == NULL)) - return -1; - if (!append) - FlushStrBuf(Blob); - if (Blob->BufUsed + nBytes >= Blob->BufSize) - IncreaseBuf(Blob, append, Blob->BufUsed + nBytes); - - pos = *Pos; - if (pos > 0) - len = pos - IOBuf->buf; - rlen = IOBuf->BufUsed - len; +/** + * @ingroup StrBuf_DeEnCoder + * @brief find one chunk of a RFC822 encoded string + * @param Buffer where to search + * @param bptr where to start searching + * @returns found position, NULL if none. + */ +static inline const char *FindNextEnd (const StrBuf *Buf, const char *bptr) +{ + const char * end; + /* Find the next ?Q? */ + if (Buf->BufUsed - (bptr - Buf->buf) < 6) + return NULL; + end = strchr(bptr + 2, '?'); - if ((IOBuf->BufUsed > 0) && - (pos != NULL) && - (pos < IOBuf->buf + IOBuf->BufUsed)) - { - pche = IOBuf->buf + IOBuf->BufUsed; - pch = pos; + if (end == NULL) + return NULL; - if (rlen < nBytes) { - memcpy(Blob->buf + Blob->BufUsed, pos, rlen); - Blob->BufUsed += rlen; - Blob->buf[Blob->BufUsed] = '\0'; - nAlreadyRead = nRead = rlen; - *Pos = NULL; - } - if (rlen >= nBytes) { - memcpy(Blob->buf + Blob->BufUsed, pos, nBytes); - Blob->BufUsed += nBytes; - Blob->buf[Blob->BufUsed] = '\0'; - if (rlen == nBytes) { - *Pos = NULL; - FlushStrBuf(IOBuf); - } - else - *Pos += nBytes; - return nBytes; - } + if ((Buf->BufUsed - (end - Buf->buf) > 3) && + (((*(end + 1) == 'B') || (*(end + 1) == 'Q')) || + ((*(end + 1) == 'b') || (*(end + 1) == 'q'))) && + (*(end + 2) == '?')) { + /* skip on to the end of the cluster, the next ?= */ + end = strstr(end + 3, "?="); } + else + /* sort of half valid encoding, try to find an end. */ + end = strstr(bptr, "?="); + return end; +} - FlushStrBuf(IOBuf); - if (IOBuf->BufSize < nBytes - nRead) - IncreaseBuf(IOBuf, 0, nBytes - nRead); - ptr = IOBuf->buf; - slen = len = Blob->BufUsed; - fdflags = fcntl(*fd, F_GETFL); - IsNonBlock = (fdflags & O_NONBLOCK) == O_NONBLOCK; +/** + * @ingroup StrBuf_DeEnCoder + * @brief convert one buffer according to the preselected iconv pointer PIC + * @param ConvertBuf buffer we need to translate + * @param TmpBuf To share a workbuffer over several iterations. prepare to have it filled with useless stuff afterwards. + * @param pic Pointer to the iconv-session Object + */ +void StrBufConvert(StrBuf *ConvertBuf, StrBuf *TmpBuf, void *pic) +{ +#ifdef HAVE_ICONV + long trycount = 0; + size_t siz; + iconv_t ic; + char *ibuf; /**< Buffer of characters to be converted */ + char *obuf; /**< Buffer for converted characters */ + size_t ibuflen; /**< Length of input buffer */ + size_t obuflen; /**< Length of output buffer */ + + + /* since we're converting to utf-8, one glyph may take up to 6 bytes */ + if (ConvertBuf->BufUsed * 6 >= TmpBuf->BufSize) + IncreaseBuf(TmpBuf, 0, ConvertBuf->BufUsed * 6); +TRYAGAIN: + ic = *(iconv_t*)pic; + ibuf = ConvertBuf->buf; + ibuflen = ConvertBuf->BufUsed; + obuf = TmpBuf->buf; + obuflen = TmpBuf->BufSize; + + siz = iconv(ic, &ibuf, &ibuflen, &obuf, &obuflen); + + if (siz < 0) { + if (errno == E2BIG) { + trycount ++; + IncreaseBuf(TmpBuf, 0, 0); + if (trycount < 5) + goto TRYAGAIN; - SelRes = 1; - nBytes -= nRead; - nRead = 0; - while ((nRead < nBytes) && - (*fd != -1)) { - if (IsNonBlock) - { - tv.tv_sec = 1; - tv.tv_usec = 0; - - FD_ZERO(&rfds); - FD_SET(*fd, &rfds); - if (select(*fd + 1, &rfds, NULL, NULL, &tv) == -1) { - *Error = strerror(errno); - close (*fd); - *fd = -1; - if (*Error == NULL) - *Error = ErrRBLF_SelectFailed; - return -1; - } - if (! FD_ISSET(*fd, &rfds) != 0) { - nSuccessLess ++; - continue; - } - } - nSuccessLess = 0; - rlen = read(*fd, - ptr, - nBytes - nRead); - if (rlen == -1) { - close(*fd); - *fd = -1; - *Error = strerror(errno); - return rlen; } - else if (rlen == 0){ - nSuccessLess ++; - if ((check == NNN_TERM) && - (nRead > 5) && - (strncmp(IOBuf->buf + IOBuf->BufUsed - 5, "\n000\n", 5) == 0)) - { - StrBufPlain(Blob, HKEY("\n000\n")); - StrBufCutRight(Blob, 5); - return Blob->BufUsed; - } - if (nSelects > 10) { - FlushStrBuf(IOBuf); - *Error = ErrRBB_too_many_selects; - return -1; - } + else if (errno == EILSEQ){ + /* hm, invalid utf8 sequence... what to do now? */ + /* An invalid multibyte sequence has been encountered in the input */ } - else if (rlen > 0) { - nRead += rlen; - ptr += rlen; - IOBuf->BufUsed += rlen; + else if (errno == EINVAL) { + /* An incomplete multibyte sequence has been encountered in the input. */ } + + FlushStrBuf(TmpBuf); } - if (nRead > nBytes) { - *Pos = IOBuf->buf + nBytes; - } - Blob->buf[Blob->BufUsed] = '\0'; - StrBufAppendBufPlain(Blob, IOBuf->buf, nBytes, 0); - if (*Pos == NULL) { - FlushStrBuf(IOBuf); + else { + TmpBuf->BufUsed = TmpBuf->BufSize - obuflen; + TmpBuf->buf[TmpBuf->BufUsed] = '\0'; + + /* little card game: wheres the red lady? */ + SwapBuffers(ConvertBuf, TmpBuf); + FlushStrBuf(TmpBuf); } - return nRead + nAlreadyRead; +#endif } + /** - * @brief Cut nChars from the start of the string - * @param Buf Buffer to modify - * @param nChars how many chars should be skipped? + * @ingroup StrBuf_DeEnCoder + * @brief catches one RFC822 encoded segment, and decodes it. + * @param Target buffer to fill with result + * @param DecodeMe buffer with stuff to process + * @param SegmentStart points to our current segment in DecodeMe + * @param SegmentEnd Points to the end of our current segment in DecodeMe + * @param ConvertBuf Workbuffer shared between several iterations. Random content; needs to be valid + * @param ConvertBuf2 Workbuffer shared between several iterations. Random content; needs to be valid + * @param FoundCharset Characterset to default decoding to; if we find another we will overwrite it. */ -void StrBufCutLeft(StrBuf *Buf, int nChars) +inline static void DecodeSegment(StrBuf *Target, + const StrBuf *DecodeMe, + const char *SegmentStart, + const char *SegmentEnd, + StrBuf *ConvertBuf, + StrBuf *ConvertBuf2, + StrBuf *FoundCharset) { - if (nChars >= Buf->BufUsed) { - FlushStrBuf(Buf); - return; + StrBuf StaticBuf; + char charset[128]; + char encoding[16]; +#ifdef HAVE_ICONV + iconv_t ic = (iconv_t)(-1); +#else + void *ic = NULL; +#endif + /* Now we handle foreign character sets properly encoded + * in RFC2047 format. + */ + StaticBuf.buf = (char*) SegmentStart; /*< it will just be read there... */ + StaticBuf.BufUsed = SegmentEnd - SegmentStart; + StaticBuf.BufSize = DecodeMe->BufSize - (SegmentStart - DecodeMe->buf); + extract_token(charset, SegmentStart, 1, '?', sizeof charset); + if (FoundCharset != NULL) { + FlushStrBuf(FoundCharset); + StrBufAppendBufPlain(FoundCharset, charset, -1, 0); } - memmove(Buf->buf, Buf->buf + nChars, Buf->BufUsed - nChars); - Buf->BufUsed -= nChars; - Buf->buf[Buf->BufUsed] = '\0'; + extract_token(encoding, SegmentStart, 2, '?', sizeof encoding); + StrBufExtract_token(ConvertBuf, &StaticBuf, 3, '?'); + + *encoding = toupper(*encoding); + if (*encoding == 'B') { /**< base64 */ + if (ConvertBuf2->BufSize < ConvertBuf->BufUsed) + IncreaseBuf(ConvertBuf2, 0, ConvertBuf->BufUsed); + ConvertBuf2->BufUsed = CtdlDecodeBase64(ConvertBuf2->buf, + ConvertBuf->buf, + ConvertBuf->BufUsed); + } + else if (*encoding == 'Q') { /**< quoted-printable */ + long pos; + + pos = 0; + while (pos < ConvertBuf->BufUsed) + { + if (ConvertBuf->buf[pos] == '_') + ConvertBuf->buf[pos] = ' '; + pos++; + } + + if (ConvertBuf2->BufSize < ConvertBuf->BufUsed) + IncreaseBuf(ConvertBuf2, 0, ConvertBuf->BufUsed); + + ConvertBuf2->BufUsed = CtdlDecodeQuotedPrintable( + ConvertBuf2->buf, + ConvertBuf->buf, + ConvertBuf->BufUsed); + } + else { + StrBufAppendBuf(ConvertBuf2, ConvertBuf, 0); + } +#ifdef HAVE_ICONV + ctdl_iconv_open("UTF-8", charset, &ic); + if (ic != (iconv_t)(-1) ) { +#endif + StrBufConvert(ConvertBuf2, ConvertBuf, &ic); + StrBufAppendBuf(Target, ConvertBuf2, 0); +#ifdef HAVE_ICONV + iconv_close(ic); + } + else { + StrBufAppendBufPlain(Target, HKEY("(unreadable)"), 0); + } +#endif } /** - * @brief Cut the trailing n Chars from the string - * @param Buf Buffer to modify - * @param nChars how many chars should be trunkated? + * @ingroup StrBuf_DeEnCoder + * @brief Handle subjects with RFC2047 encoding such as: [deprecated old syntax!] + * =?koi8-r?B?78bP0s3Mxc7JxSDXz9rE1dvO2c3JINvB0sHNySDP?= + * @param Target where to put the decoded string to + * @param DecodeMe buffer with encoded string + * @param DefaultCharset if we don't find one, which should we use? + * @param FoundCharset overrides DefaultCharset if non-empty; If we find a charset inside of the string, + * put it here for later use where no string might be known. */ -void StrBufCutRight(StrBuf *Buf, int nChars) +void StrBuf_RFC822_to_Utf8(StrBuf *Target, const StrBuf *DecodeMe, const StrBuf* DefaultCharset, StrBuf *FoundCharset) { - if (nChars >= Buf->BufUsed) { - FlushStrBuf(Buf); - return; - } - Buf->BufUsed -= nChars; - Buf->buf[Buf->BufUsed] = '\0'; + StrBuf *ConvertBuf; + StrBuf *ConvertBuf2; + ConvertBuf = NewStrBufPlain(NULL, StrLength(DecodeMe)); + ConvertBuf2 = NewStrBufPlain(NULL, StrLength(DecodeMe)); + + StrBuf_RFC822_2_Utf8(Target, + DecodeMe, + DefaultCharset, + FoundCharset, + ConvertBuf, + ConvertBuf2); + FreeStrBuf(&ConvertBuf); + FreeStrBuf(&ConvertBuf2); } /** - * @brief Cut the string after n Chars - * @param Buf Buffer to modify - * @param AfternChars after how many chars should we trunkate the string? - * @param At if non-null and points inside of our string, cut it there. + * @ingroup StrBuf_DeEnCoder + * @brief Handle subjects with RFC2047 encoding such as: + * =?koi8-r?B?78bP0s3Mxc7JxSDXz9rE1dvO2c3JINvB0sHNySDP?= + * @param Target where to put the decoded string to + * @param DecodeMe buffer with encoded string + * @param DefaultCharset if we don't find one, which should we use? + * @param FoundCharset overrides DefaultCharset if non-empty; If we find a charset inside of the string, + * put it here for later use where no string might be known. + * @param ConvertBuf workbuffer. feed in, you shouldn't care about its content. + * @param ConvertBuf2 workbuffer. feed in, you shouldn't care about its content. */ -void StrBufCutAt(StrBuf *Buf, int AfternChars, const char *At) +void StrBuf_RFC822_2_Utf8(StrBuf *Target, + const StrBuf *DecodeMe, + const StrBuf* DefaultCharset, + StrBuf *FoundCharset, + StrBuf *ConvertBuf, + StrBuf *ConvertBuf2) { - if (At != NULL){ - AfternChars = At - Buf->buf; + StrBuf *DecodedInvalidBuf = NULL; + const StrBuf *DecodeMee = DecodeMe; + const char *start, *end, *next, *nextend, *ptr = NULL; +#ifdef HAVE_ICONV + iconv_t ic = (iconv_t)(-1) ; +#endif + const char *eptr; + int passes = 0; + int i, len; + int illegal_non_rfc2047_encoding = 0; + + /* Sometimes, badly formed messages contain strings which were simply + * written out directly in some foreign character set instead of + * using RFC2047 encoding. This is illegal but we will attempt to + * handle it anyway by converting from a user-specified default + * charset to UTF-8 if we see any nonprintable characters. + */ + + len = StrLength(DecodeMe); + for (i=0; iBufUsed; ++i) { + if ((DecodeMe->buf[i] < 32) || (DecodeMe->buf[i] > 126)) { + illegal_non_rfc2047_encoding = 1; + break; + } } - if ((AfternChars < 0) || (AfternChars >= Buf->BufUsed)) - return; - Buf->BufUsed = AfternChars; - Buf->buf[Buf->BufUsed] = '\0'; -} + if ((illegal_non_rfc2047_encoding) && + (strcasecmp(ChrPtr(DefaultCharset), "UTF-8")) && + (strcasecmp(ChrPtr(DefaultCharset), "us-ascii")) ) + { +#ifdef HAVE_ICONV + ctdl_iconv_open("UTF-8", ChrPtr(DefaultCharset), &ic); + if (ic != (iconv_t)(-1) ) { + DecodedInvalidBuf = NewStrBufDup(DecodeMe); + StrBufConvert(DecodedInvalidBuf, ConvertBuf, &ic);///TODO: don't void const? + DecodeMee = DecodedInvalidBuf; + iconv_close(ic); + } +#endif + } + /* pre evaluate the first pair */ + nextend = end = NULL; + len = StrLength(DecodeMee); + start = strstr(DecodeMee->buf, "=?"); + eptr = DecodeMee->buf + DecodeMee->BufUsed; + if (start != NULL) + end = FindNextEnd (DecodeMee, start + 2); + else { + StrBufAppendBuf(Target, DecodeMee, 0); + FreeStrBuf(&DecodedInvalidBuf); + return; + } -/** - * @brief Strip leading and trailing spaces from a string; with premeasured and adjusted length. - * @param buf the string to modify - * @param len length of the string. - */ -void StrBufTrim(StrBuf *Buf) -{ - int delta = 0; - if ((Buf == NULL) || (Buf->BufUsed == 0)) return; - while ((Buf->BufUsed > delta) && (isspace(Buf->buf[delta]))){ - delta ++; + if (start != DecodeMee->buf) { + long nFront; + + nFront = start - DecodeMee->buf; + StrBufAppendBufPlain(Target, DecodeMee->buf, nFront, 0); + len -= nFront; } - if (delta > 0) StrBufCutLeft(Buf, delta); + /* + * Since spammers will go to all sorts of absurd lengths to get their + * messages through, there are LOTS of corrupt headers out there. + * So, prevent a really badly formed RFC2047 header from throwing + * this function into an infinite loop. + */ + while ((start != NULL) && + (end != NULL) && + (start < eptr) && + (end < eptr) && + (passes < 20)) + { + passes++; + DecodeSegment(Target, + DecodeMee, + start, + end, + ConvertBuf, + ConvertBuf2, + FoundCharset); + + next = strstr(end, "=?"); + nextend = NULL; + if ((next != NULL) && + (next < eptr)) + nextend = FindNextEnd(DecodeMee, next); + if (nextend == NULL) + next = NULL; - if (Buf->BufUsed == 0) return; - while (isspace(Buf->buf[Buf->BufUsed - 1])){ - Buf->BufUsed --; + /* did we find two partitions */ + if ((next != NULL) && + ((next - end) > 2)) + { + ptr = end + 2; + while ((ptr < next) && + (isspace(*ptr) || + (*ptr == '\r') || + (*ptr == '\n') || + (*ptr == '\t'))) + ptr ++; + /* + * did we find a gab just filled with blanks? + * if not, copy its stuff over. + */ + if (ptr != next) + { + StrBufAppendBufPlain(Target, + end + 2, + next - end - 2, + 0); + } + } + /* our next-pair is our new first pair now. */ + ptr = end + 2; + start = next; + end = nextend; + } + end = ptr; + nextend = DecodeMee->buf + DecodeMee->BufUsed; + if ((end != NULL) && (end < nextend)) { + ptr = end; + while ( (ptr < nextend) && + (isspace(*ptr) || + (*ptr == '\r') || + (*ptr == '\n') || + (*ptr == '\t'))) + ptr ++; + if (ptr < nextend) + StrBufAppendBufPlain(Target, end, nextend - end, 0); } - Buf->buf[Buf->BufUsed] = '\0'; + FreeStrBuf(&DecodedInvalidBuf); } +/******************************************************************************* + * Manipulating UTF-8 Strings * + *******************************************************************************/ + /** - * @brief uppercase the contents of a buffer - * @param Buf the buffer to translate + * @ingroup StrBuf + * @brief evaluate the length of an utf8 special character sequence + * @param Char the character to examine + * @returns width of utf8 chars in bytes; if the sequence is broken 0 is returned; 1 if its simply ASCII. */ -void StrBufUpCase(StrBuf *Buf) +static inline int Ctdl_GetUtf8SequenceLength(const char *CharS, const char *CharE) { - char *pch, *pche; + int n = 0; + unsigned char test = (1<<7); - pch = Buf->buf; - pche = pch + Buf->BufUsed; - while (pch < pche) { - *pch = toupper(*pch); - pch ++; + if ((*CharS & 0xC0) != 0xC0) + return 1; + + while ((n < 8) && + ((test & ((unsigned char)*CharS)) != 0)) + { + test = test >> 1; + n ++; } + if ((n > 6) || ((CharE - CharS) < n)) + n = 0; + return n; } - /** - * @brief lowercase the contents of a buffer - * @param Buf the buffer to translate + * @ingroup StrBuf + * @brief detect whether this char starts an utf-8 encoded char + * @param Char character to inspect + * @returns yes or no */ -void StrBufLowerCase(StrBuf *Buf) +static inline int Ctdl_IsUtf8SequenceStart(const char Char) { - char *pch, *pche; - - pch = Buf->buf; - pche = pch + Buf->BufUsed; - while (pch < pche) { - *pch = tolower(*pch); - pch ++; - } +/** 11??.???? indicates an UTF8 Sequence. */ + return ((Char & 0xC0) == 0xC0); } /** - * @brief removes double slashes from pathnames - * @param Dir directory string to filter - * @param RemoveTrailingSlash allows / disallows trailing slashes + * @ingroup StrBuf + * @brief measure the number of glyphs in an UTF8 string... + * @param Buf string to measure + * @returns the number of glyphs in Buf */ -void StrBufStripSlashes(StrBuf *Dir, int RemoveTrailingSlash) +long StrBuf_Utf8StrLen(StrBuf *Buf) { - char *a, *b; - - a = b = Dir->buf; + int n = 0; + int m = 0; + char *aptr, *eptr; - while (!IsEmptyStr(a)) { - if (*a == '/') { - while (*a == '/') - a++; - *b = '/'; - b++; + if ((Buf == NULL) || (Buf->BufUsed == 0)) + return 0; + aptr = Buf->buf; + eptr = Buf->buf + Buf->BufUsed; + while ((aptr < eptr) && (*aptr != '\0')) { + if (Ctdl_IsUtf8SequenceStart(*aptr)){ + m = Ctdl_GetUtf8SequenceLength(aptr, eptr); + while ((aptr < eptr) && (*aptr++ != '\0')&& (m-- > 0) ); + n ++; } else { - *b = *a; - b++; a++; + n++; + aptr++; } } - if ((RemoveTrailingSlash) && (*(b - 1) != '/')){ - *b = '/'; - b++; - } - *b = '\0'; - Dir->BufUsed = b - Dir->buf; + return n; } /** - * @brief unhide special chars hidden to the HTML escaper - * @param target buffer to put the unescaped string in - * @param source buffer to unescape + * @ingroup StrBuf + * @brief cuts a string after maxlen glyphs + * @param Buf string to cut to maxlen glyphs + * @param maxlen how long may the string become? + * @returns current length of the string */ -void StrBufEUid_unescapize(StrBuf *target, const StrBuf *source) +long StrBuf_Utf8StrCut(StrBuf *Buf, int maxlen) { - int a, b, len; - char hex[3]; - - if (target != NULL) - FlushStrBuf(target); - - if (source == NULL ||target == NULL) - { - return; - } - - len = source->BufUsed; - for (a = 0; a < len; ++a) { - if (target->BufUsed >= target->BufSize) - IncreaseBuf(target, 1, -1); + char *aptr, *eptr; + int n = 0, m = 0; - if (source->buf[a] == '=') { - hex[0] = source->buf[a + 1]; - hex[1] = source->buf[a + 2]; - hex[2] = 0; - b = 0; - sscanf(hex, "%02x", &b); - target->buf[target->BufUsed] = b; - target->buf[++target->BufUsed] = 0; - a += 2; + aptr = Buf->buf; + eptr = Buf->buf + Buf->BufUsed; + while ((aptr < eptr) && (*aptr != '\0')) { + if (Ctdl_IsUtf8SequenceStart(*aptr)){ + m = Ctdl_GetUtf8SequenceLength(aptr, eptr); + while ((*aptr++ != '\0') && (m-- > 0)); + n ++; } else { - target->buf[target->BufUsed] = source->buf[a]; - target->buf[++target->BufUsed] = 0; + n++; + aptr++; } + if (n > maxlen) { + *aptr = '\0'; + Buf->BufUsed = aptr - Buf->buf; + return Buf->BufUsed; + } } + return Buf->BufUsed; + } -/** - * @brief hide special chars from the HTML escapers and friends - * @param target buffer to put the escaped string in - * @param source buffer to escape - */ -void StrBufEUid_escapize(StrBuf *target, const StrBuf *source) -{ - int i, len; - if (target != NULL) - FlushStrBuf(target); - if (source == NULL ||target == NULL) - { - return; - } - len = source->BufUsed; - for (i=0; iBufUsed + 4 >= target->BufSize) - IncreaseBuf(target, 1, -1); - if ( (isalnum(source->buf[i])) || - (source->buf[i]=='-') || - (source->buf[i]=='_') ) { - target->buf[target->BufUsed++] = source->buf[i]; - } - else { - sprintf(&target->buf[target->BufUsed], - "=%02X", - (0xFF &source->buf[i])); - target->BufUsed += 3; - } - } - target->buf[target->BufUsed + 1] = '\0'; -} +/******************************************************************************* + * wrapping ZLib * + *******************************************************************************/ + +#ifdef HAVE_ZLIB +#define DEF_MEM_LEVEL 8 /*< memlevel??? */ +#define OS_CODE 0x03 /*< unix */ /** + * @ingroup StrBuf_DeEnCoder * @brief uses the same calling syntax as compress2(), but it * creates a stream compatible with HTTP "Content-encoding: gzip" + * @param dest compressed buffer + * @param destLen length of the compresed data + * @param source source to encode + * @param sourceLen length of source to encode + * @param level compression level */ -#ifdef HAVE_ZLIB -#define DEF_MEM_LEVEL 8 /*< memlevel??? */ -#define OS_CODE 0x03 /*< unix */ -int ZEXPORT compress_gzip(Bytef * dest, /*< compressed buffer*/ - size_t * destLen, /*< length of the compresed data */ - const Bytef * source, /*< source to encode */ - uLong sourceLen, /*< length of source to encode */ - int level) /*< compression level */ +int ZEXPORT compress_gzip(Bytef * dest, + size_t * destLen, + const Bytef * source, + uLong sourceLen, + int level) { const int gz_magic[2] = { 0x1f, 0x8b }; /* gzip magic header */ @@ -2552,7 +3518,10 @@ int ZEXPORT compress_gzip(Bytef * dest, /*< compressed buffer*/ /** + * @ingroup StrBuf_DeEnCoder + * @brief compress the buffer with gzip * Attention! If you feed this a Const String, you must maintain the uncompressed buffer yourself! + * @param Buf buffer whose content is to be gzipped */ int CompressBuffer(StrBuf *Buf) { @@ -2586,685 +3555,950 @@ int CompressBuffer(StrBuf *Buf) } else { free(compressed_data); } -#endif /* HAVE_ZLIB */ - return 0; -} +#endif /* HAVE_ZLIB */ + return 0; +} + +/******************************************************************************* + * File I/O; Callbacks to libevent * + *******************************************************************************/ + +long StrBuf_read_one_chunk_callback (int fd, short event, IOBuffer *FB) +{ + long bufremain = 0; + int n; + + if ((FB == NULL) || (FB->Buf == NULL)) + return -1; + + /* + * check whether the read pointer is somewhere in a range + * where a cut left is inexpensive + */ + + if (FB->ReadWritePointer != NULL) + { + long already_read; + + already_read = FB->ReadWritePointer - FB->Buf->buf; + bufremain = FB->Buf->BufSize - FB->Buf->BufUsed - 1; + + if (already_read != 0) { + long unread; + + unread = FB->Buf->BufUsed - already_read; + + /* else nothing to compact... */ + if (unread == 0) { + FB->ReadWritePointer = FB->Buf->buf; + bufremain = FB->Buf->BufSize; + } + else if ((unread < 64) || + (bufremain < already_read)) + { + /* + * if its just a tiny bit remaining, or we run out of space... + * lets tidy up. + */ + FB->Buf->BufUsed = unread; + if (unread < already_read) + memcpy(FB->Buf->buf, FB->ReadWritePointer, unread); + else + memmove(FB->Buf->buf, FB->ReadWritePointer, unread); + FB->ReadWritePointer = FB->Buf->buf; + bufremain = FB->Buf->BufSize - unread - 1; + } + else if (bufremain < (FB->Buf->BufSize / 10)) + { + /* get a bigger buffer */ + + IncreaseBuf(FB->Buf, 0, FB->Buf->BufUsed + 1); + + FB->ReadWritePointer = FB->Buf->buf + unread; + + bufremain = FB->Buf->BufSize - unread - 1; +/*TODO: special increase function that won't copy the already read! */ + } + } + else if (bufremain < 10) { + IncreaseBuf(FB->Buf, 1, FB->Buf->BufUsed + 10); + + FB->ReadWritePointer = FB->Buf->buf; + + bufremain = FB->Buf->BufSize - FB->Buf->BufUsed - 1; + } + + } + else { + FB->ReadWritePointer = FB->Buf->buf; + bufremain = FB->Buf->BufSize - 1; + } + + n = read(fd, FB->Buf->buf + FB->Buf->BufUsed, bufremain); + + if (n > 0) { + FB->Buf->BufUsed += n; + FB->Buf->buf[FB->Buf->BufUsed] = '\0'; + } + return n; +} + +int StrBuf_write_one_chunk_callback(int fd, short event, IOBuffer *FB) +{ + long WriteRemain; + int n; + + if ((FB == NULL) || (FB->Buf == NULL)) + return -1; + + if (FB->ReadWritePointer != NULL) + { + WriteRemain = FB->Buf->BufUsed - + (FB->ReadWritePointer - + FB->Buf->buf); + } + else { + FB->ReadWritePointer = FB->Buf->buf; + WriteRemain = FB->Buf->BufUsed; + } + + n = write(fd, FB->ReadWritePointer, WriteRemain); + if (n > 0) { + FB->ReadWritePointer += n; + + if (FB->ReadWritePointer == + FB->Buf->buf + FB->Buf->BufUsed) + { + FlushStrBuf(FB->Buf); + FB->ReadWritePointer = NULL; + return 0; + } + // check whether we've got something to write + // get the maximum chunk plus the pointer we can send + // write whats there + // if not all was sent, remember the send pointer for the next time + return FB->ReadWritePointer - FB->Buf->buf + FB->Buf->BufUsed; + } + return n; +} + +/** + * @ingroup StrBuf_IO + * @brief extract a "next line" from Buf; Ptr to persist across several iterations + * @param LineBuf your line will be copied here. + * @param FB BLOB with lines of text... + * @param Ptr moved arround to keep the next-line across several iterations + * has to be &NULL on start; will be &NotNULL on end of buffer + * @returns size of copied buffer + */ +eReadState StrBufChunkSipLine(StrBuf *LineBuf, IOBuffer *FB) +{ + const char *aptr, *ptr, *eptr; + char *optr, *xptr; + + if ((FB->Buf == NULL) || (FB->ReadWritePointer == StrBufNOTNULL)) { + FB->ReadWritePointer = StrBufNOTNULL; + return eReadFail; + } + + FlushStrBuf(LineBuf); + if (FB->ReadWritePointer == NULL) + ptr = aptr = FB->Buf->buf; + else + ptr = aptr = FB->ReadWritePointer; + + optr = LineBuf->buf; + eptr = FB->Buf->buf + FB->Buf->BufUsed; + xptr = LineBuf->buf + LineBuf->BufSize - 1; + + while ((ptr <= eptr) && + (*ptr != '\n') && + (*ptr != '\r') ) + { + *optr = *ptr; + optr++; ptr++; + if (optr == xptr) { + LineBuf->BufUsed = optr - LineBuf->buf; + IncreaseBuf(LineBuf, 1, LineBuf->BufUsed + 1); + optr = LineBuf->buf + LineBuf->BufUsed; + xptr = LineBuf->buf + LineBuf->BufSize - 1; + } + } -/** - * @brief decode a buffer from base 64 encoding; destroys original - * @param Buf Buffor to transform - */ -int StrBufDecodeBase64(StrBuf *Buf) -{ - char *xferbuf; - size_t siz; - if (Buf == NULL) return -1; + if (ptr >= eptr) { + if (optr > LineBuf->buf) + optr --; + if ((*(ptr - 1) != '\r') && (*(ptr - 1) != '\n')) { + LineBuf->BufUsed = optr - LineBuf->buf; + *optr = '\0'; + if ((FB->ReadWritePointer != NULL) && + (FB->ReadWritePointer != FB->Buf->buf)) + { + /* Ok, the client application read all the data + it was interested in so far. Since there is more to read, + we now shrink the buffer, and move the rest over. + */ + StrBufCutLeft(FB->Buf, + FB->ReadWritePointer - FB->Buf->buf); + FB->ReadWritePointer = FB->Buf->buf; + } + return eMustReadMore; + } + } + LineBuf->BufUsed = optr - LineBuf->buf; + *optr = '\0'; + if ((ptr <= eptr) && (*ptr == '\r')) + ptr ++; + if ((ptr <= eptr) && (*ptr == '\n')) + ptr ++; + + if (ptr < eptr) { + FB->ReadWritePointer = ptr; + } + else { + FlushStrBuf(FB->Buf); + FB->ReadWritePointer = NULL; + } - xferbuf = (char*) malloc(Buf->BufSize); - siz = CtdlDecodeBase64(xferbuf, - Buf->buf, - Buf->BufUsed); - free(Buf->buf); - Buf->buf = xferbuf; - Buf->BufUsed = siz; - return siz; + return eReadSuccess; } /** - * @brief decode a buffer from base 64 encoding; destroys original - * @param Buf Buffor to transform + * @ingroup StrBuf_CHUNKED_IO + * @brief check whether the chunk-buffer has more data waiting or not. + * @param FB Chunk-Buffer to inspect */ -int StrBufDecodeHex(StrBuf *Buf) +eReadState StrBufCheckBuffer(IOBuffer *FB) { - unsigned int ch; - char *pch, *pche, *pchi; + if (FB == NULL) + return eReadFail; + if (FB->Buf->BufUsed == 0) + return eReadSuccess; + if (FB->ReadWritePointer == NULL) + return eBufferNotEmpty; + if (FB->Buf->buf + FB->Buf->BufUsed > FB->ReadWritePointer) + return eBufferNotEmpty; + return eReadSuccess; +} - if (Buf == NULL) return -1; +long IOBufferStrLength(IOBuffer *FB) +{ + if (FB->ReadWritePointer == NULL) + return StrLength(FB->Buf); + + return StrLength(FB->Buf) - (FB->ReadWritePointer - FB->Buf->buf); +} - pch = pchi = Buf->buf; - pche = pch + Buf->BufUsed; - while (pchi < pche){ - ch = decode_hex(pchi); - *pch = ch; - pch ++; - pchi += 2; - } - *pch = '\0'; - Buf->BufUsed = pch - Buf->buf; - return Buf->BufUsed; +void FDIOBufferInit(FDIOBuffer *FDB, IOBuffer *IO, int FD, long TotalSendSize) +{ + memset(FDB, 0, sizeof(FDIOBuffer)); + FDB->ChunkSize = + FDB->TotalSendSize = TotalSendSize; + FDB->IOB = IO; +#ifndef LINUX_SENDFILE + FDB->ChunkBuffer = NewStrBuf(); +#endif + FDB->OtherFD = FD; } -/** - * @brief replace all chars >0x20 && < 0x7F with Mute - * @param Mute char to put over invalid chars - * @param Buf Buffor to transform - */ -int StrBufSanitizeAscii(StrBuf *Buf, const char Mute) +int FileSendChunked(FDIOBuffer *FDB, const char **Err) { - unsigned char *pch; - if (Buf == NULL) return -1; - pch = (unsigned char *)Buf->buf; - while (pch < (unsigned char *)Buf->buf + Buf->BufUsed) { - if ((*pch < 0x20) || (*pch > 0x7F)) - *pch = Mute; - pch ++; +#ifdef LINUX_SENDFILE + ssize_t sent; + sent = sendfile(FDB->IOB->fd, FDB->OtherFD, &FDB->TotalSentAlready, FDB->ChunkSendRemain); + if (sent == -1) + { + *Err = strerror(errno); + return sent; } - return Buf->BufUsed; + FDB->ChunkSendRemain -= sent; + return FDB->ChunkSendRemain; +#else +#endif + return 0; } - -/** - * @brief remove escaped strings from i.e. the url string (like %20 for blanks) - * @param Buf Buffer to translate - * @param StripBlanks Reduce several blanks to one? - */ -long StrBufUnescape(StrBuf *Buf, int StripBlanks) +int FileRecvChunked(FDIOBuffer *FDB, const char **Err) { - int a, b; - char hex[3]; - long len; - - while ((Buf->BufUsed > 0) && (isspace(Buf->buf[Buf->BufUsed - 1]))){ - Buf->buf[Buf->BufUsed - 1] = '\0'; - Buf->BufUsed --; - } - a = 0; - while (a < Buf->BufUsed) { - if (Buf->buf[a] == '+') - Buf->buf[a] = ' '; - else if (Buf->buf[a] == '%') { - /* don't let % chars through, rather truncate the input. */ - if (a + 2 > Buf->BufUsed) { - Buf->buf[a] = '\0'; - Buf->BufUsed = a; - } - else { - hex[0] = Buf->buf[a + 1]; - hex[1] = Buf->buf[a + 2]; - hex[2] = 0; - b = 0; - sscanf(hex, "%02x", &b); - Buf->buf[a] = (char) b; - len = Buf->BufUsed - a - 2; - if (len > 0) - memmove(&Buf->buf[a + 1], &Buf->buf[a + 3], len); - - Buf->BufUsed -=2; - } - } - a++; +#ifdef LINUX_SENDFILE + ssize_t sent; + sent = sendfile(FDB->OtherFD, FDB->IOB->fd, &FDB->TotalSentAlready, FDB->ChunkSendRemain); + if (sent == -1) + { + *Err = strerror(errno); + return sent; } - return a; + FDB->ChunkSendRemain -= sent; + return FDB->ChunkSendRemain; +#else +#endif + return 0; } - -/** - * @brief RFC2047-encode a header field if necessary. - * If no non-ASCII characters are found, the string - * will be copied verbatim without encoding. - * - * @param target Target buffer. - * @param source Source string to be encoded. - * @returns encoded length; -1 if non success. - */ -int StrBufRFC2047encode(StrBuf **target, const StrBuf *source) +eReadState WriteIOBAlreadyRead(FDIOBuffer *FDB, const char **Error) { - const char headerStr[] = "=?UTF-8?Q?"; - int need_to_encode = 0; - int i = 0; - unsigned char ch; + int IsNonBlock; + int fdflags; + long rlen; + long should_write; + int nSuccessLess = 0; + struct timeval tv; + fd_set rfds; - if ((source == NULL) || - (target == NULL)) - return -1; + fdflags = fcntl(FDB->OtherFD, F_GETFL); + IsNonBlock = (fdflags & O_NONBLOCK) == O_NONBLOCK; - while ((i < source->BufUsed) && - (!IsEmptyStr (&source->buf[i])) && - (need_to_encode == 0)) { - if (((unsigned char) source->buf[i] < 32) || - ((unsigned char) source->buf[i] > 126)) { - need_to_encode = 1; + while ((FDB->IOB->ReadWritePointer - FDB->IOB->Buf->buf < FDB->IOB->Buf->BufUsed) && + (FDB->ChunkSendRemain > 0)) + { + if (IsNonBlock){ + tv.tv_sec = 1; /* selectresolution; */ + tv.tv_usec = 0; + + FD_ZERO(&rfds); + FD_SET(FDB->OtherFD, &rfds); + if (select(FDB->OtherFD + 1, NULL, &rfds, NULL, &tv) == -1) { + *Error = strerror(errno); + return eReadFail; + } } - i++; - } - - if (!need_to_encode) { - if (*target == NULL) { - *target = NewStrBufPlain(source->buf, source->BufUsed); + if (IsNonBlock && ! FD_ISSET(FDB->OtherFD, &rfds)) { + nSuccessLess ++; + continue; } - else { - FlushStrBuf(*target); - StrBufAppendBuf(*target, source, 0); + + should_write = FDB->IOB->Buf->BufUsed - + (FDB->IOB->ReadWritePointer - FDB->IOB->Buf->buf); + if (should_write > FDB->ChunkSendRemain) + should_write = FDB->ChunkSendRemain; + + rlen = write(FDB->OtherFD, + FDB->IOB->ReadWritePointer, + should_write); + if (rlen < 1) { + *Error = strerror(errno); + + return eReadFail; } - return (*target)->BufUsed; + FDB->TotalSentAlready += rlen; + FDB->IOB->ReadWritePointer += rlen; + FDB->ChunkSendRemain -= rlen; } - if (*target == NULL) - *target = NewStrBufPlain(NULL, sizeof(headerStr) + source->BufUsed * 2); - else if (sizeof(headerStr) + source->BufUsed >= (*target)->BufSize) - IncreaseBuf(*target, sizeof(headerStr) + source->BufUsed, 0); - memcpy ((*target)->buf, headerStr, sizeof(headerStr) - 1); - (*target)->BufUsed = sizeof(headerStr) - 1; - for (i=0; (i < source->BufUsed); ++i) { - if ((*target)->BufUsed + 4 >= (*target)->BufSize) - IncreaseBuf(*target, 1, 0); - ch = (unsigned char) source->buf[i]; - if ((ch < 32) || (ch > 126) || (ch == 61)) { - sprintf(&(*target)->buf[(*target)->BufUsed], "=%02X", ch); - (*target)->BufUsed += 3; - } - else { - (*target)->buf[(*target)->BufUsed] = ch; - (*target)->BufUsed++; - } + if (FDB->IOB->ReadWritePointer >= FDB->IOB->Buf->buf + FDB->IOB->Buf->BufUsed) + { + FlushStrBuf(FDB->IOB->Buf); + FDB->IOB->ReadWritePointer = NULL; } - - if ((*target)->BufUsed + 4 >= (*target)->BufSize) - IncreaseBuf(*target, 1, 0); - (*target)->buf[(*target)->BufUsed++] = '?'; - (*target)->buf[(*target)->BufUsed++] = '='; - (*target)->buf[(*target)->BufUsed] = '\0'; - return (*target)->BufUsed;; + if (FDB->ChunkSendRemain == 0) + return eReadSuccess; + else + return eMustReadMore; } +/******************************************************************************* + * File I/O; Prefer buffered read since its faster! * + *******************************************************************************/ + /** - * @brief replaces all occurances of 'search' by 'replace' - * @param buf Buffer to modify - * @param search character to search - * @param relpace character to replace search by + * @ingroup StrBuf_IO + * @brief Read a line from socket + * flushes and closes the FD on error + * @param buf the buffer to get the input to + * @param fd pointer to the filedescriptor to read + * @param append Append to an existing string or replace? + * @param Error strerror() on error + * @returns numbers of chars read */ -void StrBufReplaceChars(StrBuf *buf, char search, char replace) +int StrBufTCP_read_line(StrBuf *buf, int *fd, int append, const char **Error) { - long i; - if (buf == NULL) - return; - for (i=0; iBufUsed; i++) - if (buf->buf[i] == search) - buf->buf[i] = replace; - -} - + int len, rlen, slen; + if (!append) + FlushStrBuf(buf); -/** - * @brief Wrapper around iconv_open() - * Our version adds aliases for non-standard Microsoft charsets - * such as 'MS950', aliasing them to names like 'CP950' - * - * @param tocode Target encoding - * @param fromcode Source encoding - */ -void ctdl_iconv_open(const char *tocode, const char *fromcode, void *pic) -{ -#ifdef HAVE_ICONV - iconv_t ic = (iconv_t)(-1) ; - ic = iconv_open(tocode, fromcode); - if (ic == (iconv_t)(-1) ) { - char alias_fromcode[64]; - if ( (strlen(fromcode) == 5) && (!strncasecmp(fromcode, "MS", 2)) ) { - safestrncpy(alias_fromcode, fromcode, sizeof alias_fromcode); - alias_fromcode[0] = 'C'; - alias_fromcode[1] = 'P'; - ic = iconv_open(tocode, alias_fromcode); + slen = len = buf->BufUsed; + while (1) { + rlen = read(*fd, &buf->buf[len], 1); + if (rlen < 1) { + *Error = strerror(errno); + + close(*fd); + *fd = -1; + + return -1; + } + if (buf->buf[len] == '\n') + break; + if (buf->buf[len] != '\r') + len ++; + if (len + 2 >= buf->BufSize) { + buf->BufUsed = len; + buf->buf[len+1] = '\0'; + IncreaseBuf(buf, 1, -1); } } - *(iconv_t *)pic = ic; -#endif + buf->BufUsed = len; + buf->buf[len] = '\0'; + return len - slen; } - /** - * @brief find one chunk of a RFC822 encoded string - * @param Buffer where to search - * @param bptr where to start searching - * @returns found position, NULL if none. + * @ingroup StrBuf_BufferedIO + * @brief Read a line from socket + * flushes and closes the FD on error + * @param Line the line to read from the fd / I/O Buffer + * @param buf the buffer to get the input to + * @param fd pointer to the filedescriptor to read + * @param timeout number of successless selects until we bail out + * @param selectresolution how long to wait on each select + * @param Error strerror() on error + * @returns numbers of chars read */ -static inline char *FindNextEnd (const StrBuf *Buf, char *bptr) +int StrBufTCP_read_buffered_line(StrBuf *Line, + StrBuf *buf, + int *fd, + int timeout, + int selectresolution, + const char **Error) { - char * end; - /* Find the next ?Q? */ - if (Buf->BufUsed - (bptr - Buf->buf) < 6) - return NULL; + int len, rlen; + int nSuccessLess = 0; + fd_set rfds; + char *pch = NULL; + int fdflags; + int IsNonBlock; + struct timeval tv; - end = strchr(bptr + 2, '?'); + if (buf->BufUsed > 0) { + pch = strchr(buf->buf, '\n'); + if (pch != NULL) { + rlen = 0; + len = pch - buf->buf; + if (len > 0 && (*(pch - 1) == '\r') ) + rlen ++; + StrBufSub(Line, buf, 0, len - rlen); + StrBufCutLeft(buf, len + 1); + return len - rlen; + } + } + + if (buf->BufSize - buf->BufUsed < 10) + IncreaseBuf(buf, 1, -1); - if (end == NULL) - return NULL; + fdflags = fcntl(*fd, F_GETFL); + IsNonBlock = (fdflags & O_NONBLOCK) == O_NONBLOCK; - if ((Buf->BufUsed - (end - Buf->buf) > 3) && - ((*(end + 1) == 'B') || (*(end + 1) == 'Q')) && - (*(end + 2) == '?')) { - /* skip on to the end of the cluster, the next ?= */ - end = strstr(end + 3, "?="); + while ((nSuccessLess < timeout) && (pch == NULL)) { + if (IsNonBlock){ + tv.tv_sec = selectresolution; + tv.tv_usec = 0; + + FD_ZERO(&rfds); + FD_SET(*fd, &rfds); + if (select(*fd + 1, NULL, &rfds, NULL, &tv) == -1) { + *Error = strerror(errno); + close (*fd); + *fd = -1; + return -1; + } + } + if (IsNonBlock && ! FD_ISSET(*fd, &rfds)) { + nSuccessLess ++; + continue; + } + rlen = read(*fd, + &buf->buf[buf->BufUsed], + buf->BufSize - buf->BufUsed - 1); + if (rlen < 1) { + *Error = strerror(errno); + close(*fd); + *fd = -1; + return -1; + } + else if (rlen > 0) { + nSuccessLess = 0; + buf->BufUsed += rlen; + buf->buf[buf->BufUsed] = '\0'; + if (buf->BufUsed + 10 > buf->BufSize) { + IncreaseBuf(buf, 1, -1); + } + pch = strchr(buf->buf, '\n'); + continue; + } + } - else - /* sort of half valid encoding, try to find an end. */ - end = strstr(bptr, "?="); - return end; + if (pch != NULL) { + rlen = 0; + len = pch - buf->buf; + if (len > 0 && (*(pch - 1) == '\r') ) + rlen ++; + StrBufSub(Line, buf, 0, len - rlen); + StrBufCutLeft(buf, len + 1); + return len - rlen; + } + return -1; + } +static const char *ErrRBLF_PreConditionFailed="StrBufTCP_read_buffered_line_fast: Wrong arguments or invalid Filedescriptor"; +static const char *ErrRBLF_SelectFailed="StrBufTCP_read_buffered_line_fast: Select failed without reason"; +static const char *ErrRBLF_NotEnoughSentFromServer="StrBufTCP_read_buffered_line_fast: No complete line was sent from peer"; /** - * @brief swaps the contents of two StrBufs - * this is to be used to have cheap switched between a work-buffer and a target buffer - * @param A First one - * @param B second one + * @ingroup StrBuf_BufferedIO + * @brief Read a line from socket + * flushes and closes the FD on error + * @param Line where to append our Line read from the fd / I/O Buffer; + * @param IOBuf the buffer to get the input to; lifetime pair to FD + * @param Pos pointer to the current read position, should be NULL initialized on opening the FD it belongs to.! + * @param fd pointer to the filedescriptor to read + * @param timeout number of successless selects until we bail out + * @param selectresolution how long to wait on each select + * @param Error strerror() on error + * @returns numbers of chars read or -1 in case of error. "\n" will become 0 */ -static inline void SwapBuffers(StrBuf *A, StrBuf *B) +int StrBufTCP_read_buffered_line_fast(StrBuf *Line, + StrBuf *IOBuf, + const char **Pos, + int *fd, + int timeout, + int selectresolution, + const char **Error) { - StrBuf C; + const char *pche = NULL; + const char *pos = NULL; + const char *pLF; + int len, rlen, retlen; + int nSuccessLess = 0; + fd_set rfds; + const char *pch = NULL; + int fdflags; + int IsNonBlock; + struct timeval tv; + + retlen = 0; + if ((Line == NULL) || + (Pos == NULL) || + (IOBuf == NULL) || + (*fd == -1)) + { + if (Pos != NULL) + *Pos = NULL; + *Error = ErrRBLF_PreConditionFailed; + return -1; + } - memcpy(&C, A, sizeof(*A)); - memcpy(A, B, sizeof(*B)); - memcpy(B, &C, sizeof(C)); + pos = *Pos; + if ((IOBuf->BufUsed > 0) && + (pos != NULL) && + (pos < IOBuf->buf + IOBuf->BufUsed)) + { + char *pcht; -} + pche = IOBuf->buf + IOBuf->BufUsed; + pch = pos; + pcht = Line->buf; + while ((pch < pche) && (*pch != '\n')) + { + if (Line->BufUsed + 10 > Line->BufSize) + { + long apos; + apos = pcht - Line->buf; + *pcht = '\0'; + IncreaseBuf(Line, 1, -1); + pcht = Line->buf + apos; + } + *pcht++ = *pch++; + Line->BufUsed++; + retlen++; + } -/** - * @brief convert one buffer according to the preselected iconv pointer PIC - * @param ConvertBuf buffer we need to translate - * @param TmpBuf To share a workbuffer over several iterations. prepare to have it filled with useless stuff afterwards. - * @param pic Pointer to the iconv-session Object - */ -void StrBufConvert(StrBuf *ConvertBuf, StrBuf *TmpBuf, void *pic) -{ -#ifdef HAVE_ICONV - long trycount = 0; - size_t siz; - iconv_t ic; - char *ibuf; /**< Buffer of characters to be converted */ - char *obuf; /**< Buffer for converted characters */ - size_t ibuflen; /**< Length of input buffer */ - size_t obuflen; /**< Length of output buffer */ + len = pch - pos; + if (len > 0 && (*(pch - 1) == '\r') ) + { + retlen--; + len --; + pcht --; + Line->BufUsed --; + } + *pcht = '\0'; + + if ((pch >= pche) || (*pch == '\0')) + { + FlushStrBuf(IOBuf); + *Pos = NULL; + pch = NULL; + pos = 0; + } + if ((pch != NULL) && + (pch <= pche)) + { + if (pch + 1 >= pche) { + *Pos = NULL; + FlushStrBuf(IOBuf); + } + else + *Pos = pch + 1; + + return retlen; + } + else + FlushStrBuf(IOBuf); + } - /* since we're converting to utf-8, one glyph may take up to 6 bytes */ - if (ConvertBuf->BufUsed * 6 >= TmpBuf->BufSize) - IncreaseBuf(TmpBuf, 0, ConvertBuf->BufUsed * 6); -TRYAGAIN: - ic = *(iconv_t*)pic; - ibuf = ConvertBuf->buf; - ibuflen = ConvertBuf->BufUsed; - obuf = TmpBuf->buf; - obuflen = TmpBuf->BufSize; + /* If we come here, Pos is Unset since we read everything into Line, and now go for more. */ - siz = iconv(ic, &ibuf, &ibuflen, &obuf, &obuflen); + if (IOBuf->BufSize - IOBuf->BufUsed < 10) + IncreaseBuf(IOBuf, 1, -1); - if (siz < 0) { - if (errno == E2BIG) { - trycount ++; - IncreaseBuf(TmpBuf, 0, 0); - if (trycount < 5) - goto TRYAGAIN; + fdflags = fcntl(*fd, F_GETFL); + IsNonBlock = (fdflags & O_NONBLOCK) == O_NONBLOCK; + pLF = NULL; + while ((nSuccessLess < timeout) && + (pLF == NULL) && + (*fd != -1)) { + if (IsNonBlock) + { + tv.tv_sec = 1; + tv.tv_usec = 0; + + FD_ZERO(&rfds); + FD_SET(*fd, &rfds); + if (select((*fd) + 1, &rfds, NULL, NULL, &tv) == -1) { + *Error = strerror(errno); + close (*fd); + *fd = -1; + if (*Error == NULL) + *Error = ErrRBLF_SelectFailed; + return -1; + } + if (! FD_ISSET(*fd, &rfds) != 0) { + nSuccessLess ++; + continue; + } } - else if (errno == EILSEQ){ - /* hm, invalid utf8 sequence... what to do now? */ - /* An invalid multibyte sequence has been encountered in the input */ - } - else if (errno == EINVAL) { - /* An incomplete multibyte sequence has been encountered in the input. */ + rlen = read(*fd, + &IOBuf->buf[IOBuf->BufUsed], + IOBuf->BufSize - IOBuf->BufUsed - 1); + if (rlen < 1) { + *Error = strerror(errno); + close(*fd); + *fd = -1; + return -1; } + else if (rlen > 0) { + nSuccessLess = 0; + pLF = IOBuf->buf + IOBuf->BufUsed; + IOBuf->BufUsed += rlen; + IOBuf->buf[IOBuf->BufUsed] = '\0'; + + pche = IOBuf->buf + IOBuf->BufUsed; + + while ((pLF < pche) && (*pLF != '\n')) + pLF ++; + if ((pLF >= pche) || (*pLF == '\0')) + pLF = NULL; - FlushStrBuf(TmpBuf); - } - else { - TmpBuf->BufUsed = TmpBuf->BufSize - obuflen; - TmpBuf->buf[TmpBuf->BufUsed] = '\0'; - - /* little card game: wheres the red lady? */ - SwapBuffers(ConvertBuf, TmpBuf); - FlushStrBuf(TmpBuf); - } -#endif -} + if (IOBuf->BufUsed + 10 > IOBuf->BufSize) + { + long apos = 0; + if (pLF != NULL) apos = pLF - IOBuf->buf; + IncreaseBuf(IOBuf, 1, -1); + if (pLF != NULL) pLF = IOBuf->buf + apos; + } -/** - * @brief catches one RFC822 encoded segment, and decodes it. - * @param Target buffer to fill with result - * @param DecodeMe buffer with stuff to process - * @param SegmentStart points to our current segment in DecodeMe - * @param SegmentEnd Points to the end of our current segment in DecodeMe - * @param ConvertBuf Workbuffer shared between several iterations. Random content; needs to be valid - * @param ConvertBuf2 Workbuffer shared between several iterations. Random content; needs to be valid - * @param FoundCharset Characterset to default decoding to; if we find another we will overwrite it. - */ -inline static void DecodeSegment(StrBuf *Target, - const StrBuf *DecodeMe, - char *SegmentStart, - char *SegmentEnd, - StrBuf *ConvertBuf, - StrBuf *ConvertBuf2, - StrBuf *FoundCharset) -{ - StrBuf StaticBuf; - char charset[128]; - char encoding[16]; -#ifdef HAVE_ICONV - iconv_t ic = (iconv_t)(-1); -#else - void *ic = NULL; -#endif - /* Now we handle foreign character sets properly encoded - * in RFC2047 format. - */ - StaticBuf.buf = SegmentStart; - StaticBuf.BufUsed = SegmentEnd - SegmentStart; - StaticBuf.BufSize = DecodeMe->BufSize - (SegmentStart - DecodeMe->buf); - extract_token(charset, SegmentStart, 1, '?', sizeof charset); - if (FoundCharset != NULL) { - FlushStrBuf(FoundCharset); - StrBufAppendBufPlain(FoundCharset, charset, -1, 0); - } - extract_token(encoding, SegmentStart, 2, '?', sizeof encoding); - StrBufExtract_token(ConvertBuf, &StaticBuf, 3, '?'); - - *encoding = toupper(*encoding); - if (*encoding == 'B') { /**< base64 */ - ConvertBuf2->BufUsed = CtdlDecodeBase64(ConvertBuf2->buf, - ConvertBuf->buf, - ConvertBuf->BufUsed); + continue; + } } - else if (*encoding == 'Q') { /**< quoted-printable */ - long pos; - - pos = 0; - while (pos < ConvertBuf->BufUsed) + *Pos = NULL; + if (pLF != NULL) { + pos = IOBuf->buf; + len = pLF - pos; + if (len > 0 && (*(pLF - 1) == '\r') ) + len --; + StrBufAppendBufPlain(Line, ChrPtr(IOBuf), len, 0); + if (pLF + 1 >= IOBuf->buf + IOBuf->BufUsed) { - if (ConvertBuf->buf[pos] == '_') - ConvertBuf->buf[pos] = ' '; - pos++; + FlushStrBuf(IOBuf); } - - ConvertBuf2->BufUsed = CtdlDecodeQuotedPrintable( - ConvertBuf2->buf, - ConvertBuf->buf, - ConvertBuf->BufUsed); - } - else { - StrBufAppendBuf(ConvertBuf2, ConvertBuf, 0); - } -#ifdef HAVE_ICONV - ctdl_iconv_open("UTF-8", charset, &ic); - if (ic != (iconv_t)(-1) ) { -#endif - StrBufConvert(ConvertBuf2, ConvertBuf, &ic); - StrBufAppendBuf(Target, ConvertBuf2, 0); -#ifdef HAVE_ICONV - iconv_close(ic); - } - else { - StrBufAppendBufPlain(Target, HKEY("(unreadable)"), 0); + else + *Pos = pLF + 1; + return retlen + len; } -#endif + *Error = ErrRBLF_NotEnoughSentFromServer; + return -1; + } +static const char *ErrRBLF_BLOBPreConditionFailed="StrBufReadBLOB: Wrong arguments or invalid Filedescriptor"; /** - * @brief Handle subjects with RFC2047 encoding such as: - * =?koi8-r?B?78bP0s3Mxc7JxSDXz9rE1dvO2c3JINvB0sHNySDP?= - * @param Target where to put the decoded string to - * @param DecodeMe buffer with encoded string - * @param DefaultCharset if we don't find one, which should we use? - * @param FoundCharset overrides DefaultCharset if non-empty; If we find a charset inside of the string, - * put it here for later use where no string might be known. + * @ingroup StrBuf_IO + * @brief Input binary data from socket + * flushes and closes the FD on error + * @param Buf the buffer to get the input to + * @param fd pointer to the filedescriptor to read + * @param append Append to an existing string or replace? + * @param nBytes the maximal number of bytes to read + * @param Error strerror() on error + * @returns numbers of chars read */ -void StrBuf_RFC822_to_Utf8(StrBuf *Target, const StrBuf *DecodeMe, const StrBuf* DefaultCharset, StrBuf *FoundCharset) +int StrBufReadBLOB(StrBuf *Buf, int *fd, int append, long nBytes, const char **Error) { - StrBuf *DecodedInvalidBuf = NULL; - StrBuf *ConvertBuf, *ConvertBuf2; - const StrBuf *DecodeMee = DecodeMe; - char *start, *end, *next, *nextend, *ptr = NULL; -#ifdef HAVE_ICONV - iconv_t ic = (iconv_t)(-1) ; -#endif - const char *eptr; - int passes = 0; - int i, len, delta; - int illegal_non_rfc2047_encoding = 0; - - /* Sometimes, badly formed messages contain strings which were simply - * written out directly in some foreign character set instead of - * using RFC2047 encoding. This is illegal but we will attempt to - * handle it anyway by converting from a user-specified default - * charset to UTF-8 if we see any nonprintable characters. - */ - - len = StrLength(DecodeMe); - for (i=0; iBufUsed; ++i) { - if ((DecodeMe->buf[i] < 32) || (DecodeMe->buf[i] > 126)) { - illegal_non_rfc2047_encoding = 1; - break; - } - } + int fdflags; + int rlen; + int nSuccessLess; + int nRead = 0; + char *ptr; + int IsNonBlock; + struct timeval tv; + fd_set rfds; - ConvertBuf = NewStrBufPlain(NULL, StrLength(DecodeMe)); - if ((illegal_non_rfc2047_encoding) && - (strcasecmp(ChrPtr(DefaultCharset), "UTF-8")) && - (strcasecmp(ChrPtr(DefaultCharset), "us-ascii")) ) + if ((Buf == NULL) || (*fd == -1)) { -#ifdef HAVE_ICONV - ctdl_iconv_open("UTF-8", ChrPtr(DefaultCharset), &ic); - if (ic != (iconv_t)(-1) ) { - DecodedInvalidBuf = NewStrBufDup(DecodeMe); - StrBufConvert(DecodedInvalidBuf, ConvertBuf, &ic);///TODO: don't void const? - DecodeMee = DecodedInvalidBuf; - iconv_close(ic); - } -#endif - } - - /* pre evaluate the first pair */ - nextend = end = NULL; - len = StrLength(DecodeMee); - start = strstr(DecodeMee->buf, "=?"); - eptr = DecodeMee->buf + DecodeMee->BufUsed; - if (start != NULL) - end = FindNextEnd (DecodeMee, start); - else { - StrBufAppendBuf(Target, DecodeMee, 0); - FreeStrBuf(&ConvertBuf); - FreeStrBuf(&DecodedInvalidBuf); - return; + *Error = ErrRBLF_BLOBPreConditionFailed; + return -1; } + if (!append) + FlushStrBuf(Buf); + if (Buf->BufUsed + nBytes >= Buf->BufSize) + IncreaseBuf(Buf, 1, Buf->BufUsed + nBytes); - ConvertBuf2 = NewStrBufPlain(NULL, StrLength(DecodeMee)); + ptr = Buf->buf + Buf->BufUsed; - if (start != DecodeMee->buf) { - long nFront; - - nFront = start - DecodeMee->buf; - StrBufAppendBufPlain(Target, DecodeMee->buf, nFront, 0); - len -= nFront; - } - /* - * Since spammers will go to all sorts of absurd lengths to get their - * messages through, there are LOTS of corrupt headers out there. - * So, prevent a really badly formed RFC2047 header from throwing - * this function into an infinite loop. - */ - while ((start != NULL) && - (end != NULL) && - (start < eptr) && - (end < eptr) && - (passes < 20)) + fdflags = fcntl(*fd, F_GETFL); + IsNonBlock = (fdflags & O_NONBLOCK) == O_NONBLOCK; + nSuccessLess = 0; + while ((nRead < nBytes) && + (*fd != -1)) { - passes++; - DecodeSegment(Target, - DecodeMee, - start, - end, - ConvertBuf, - ConvertBuf2, - FoundCharset); - - next = strstr(end, "=?"); - nextend = NULL; - if ((next != NULL) && - (next < eptr)) - nextend = FindNextEnd(DecodeMee, next); - if (nextend == NULL) - next = NULL; - - /* did we find two partitions */ - if ((next != NULL) && - ((next - end) > 2)) + if (IsNonBlock) { - ptr = end + 2; - while ((ptr < next) && - (isspace(*ptr) || - (*ptr == '\r') || - (*ptr == '\n') || - (*ptr == '\t'))) - ptr ++; - /* did we find a gab just filled with blanks? */ - if (ptr == next) - { - long gap = next - start; - memmove (end + 2, - next, - len - (gap)); - len -= gap; - /* now terminate the gab at the end */ - delta = (next - end) - 2; ////TODO: const! - ((StrBuf*)DecodeMee)->BufUsed -= delta; - ((StrBuf*)DecodeMee)->buf[DecodeMee->BufUsed] = '\0'; - - /* move next to its new location. */ - next -= delta; - nextend -= delta; - } - } - /* our next-pair is our new first pair now. */ - ptr = end + 2; - start = next; - end = nextend; - } - end = ptr; - nextend = DecodeMee->buf + DecodeMee->BufUsed; - if ((end != NULL) && (end < nextend)) { - ptr = end; - while ( (ptr < nextend) && - (isspace(*ptr) || - (*ptr == '\r') || - (*ptr == '\n') || - (*ptr == '\t'))) - ptr ++; - if (ptr < nextend) - StrBufAppendBufPlain(Target, end, nextend - end, 0); + tv.tv_sec = 1; + tv.tv_usec = 0; + + FD_ZERO(&rfds); + FD_SET(*fd, &rfds); + if (select(*fd + 1, &rfds, NULL, NULL, &tv) == -1) { + *Error = strerror(errno); + close (*fd); + *fd = -1; + if (*Error == NULL) + *Error = ErrRBLF_SelectFailed; + return -1; + } + if (! FD_ISSET(*fd, &rfds) != 0) { + nSuccessLess ++; + continue; + } + } + + if ((rlen = read(*fd, + ptr, + nBytes - nRead)) == -1) { + close(*fd); + *fd = -1; + *Error = strerror(errno); + return rlen; + } + nRead += rlen; + ptr += rlen; + Buf->BufUsed += rlen; } - FreeStrBuf(&ConvertBuf); - FreeStrBuf(&ConvertBuf2); - FreeStrBuf(&DecodedInvalidBuf); + Buf->buf[Buf->BufUsed] = '\0'; + return nRead; } +const char *ErrRBB_BLOBFPreConditionFailed = "StrBufReadBLOBBuffered: to many selects; aborting."; +const char *ErrRBB_too_many_selects = "StrBufReadBLOBBuffered: to many selects; aborting."; /** - * @brief evaluate the length of an utf8 special character sequence - * @param Char the character to examine - * @returns width of utf8 chars in bytes + * @ingroup StrBuf_BufferedIO + * @brief Input binary data from socket + * flushes and closes the FD on error + * @param Blob put binary thing here + * @param IOBuf the buffer to get the input to + * @param Pos offset inside of IOBuf + * @param fd pointer to the filedescriptor to read + * @param append Append to an existing string or replace? + * @param nBytes the maximal number of bytes to read + * @param check whether we should search for '000\n' terminators in case of timeouts + * @param Error strerror() on error + * @returns numbers of chars read */ -static inline int Ctdl_GetUtf8SequenceLength(const char *CharS, const char *CharE) +int StrBufReadBLOBBuffered(StrBuf *Blob, + StrBuf *IOBuf, + const char **Pos, + int *fd, + int append, + long nBytes, + int check, + const char **Error) { - int n = 1; - char test = (1<<7); - - while ((n < 8) && ((test & *CharS) != 0)) { - test = test << 1; - n ++; + const char *pos; + int fdflags; + int len = 0; + int rlen; + int nRead = 0; + int nAlreadyRead = 0; + int IsNonBlock; + char *ptr; + fd_set rfds; + struct timeval tv; + int nSuccessLess = 0; + int MaxTries; + + if ((Blob == NULL) || (*fd == -1) || (IOBuf == NULL) || (Pos == NULL)) + { + if (*Pos != NULL) + *Pos = NULL; + *Error = ErrRBB_BLOBFPreConditionFailed; + return -1; } - if ((n > 6) || ((CharE - CharS) > n)) - n = 1; - return n; -} -/** - * @brief detect whether this char starts an utf-8 encoded char - * @param Char character to inspect - * @returns yes or no - */ -static inline int Ctdl_IsUtf8SequenceStart(const char Char) -{ -/** 11??.???? indicates an UTF8 Sequence. */ - return ((Char & 0xC0) != 0); -} + if (!append) + FlushStrBuf(Blob); + if (Blob->BufUsed + nBytes >= Blob->BufSize) + IncreaseBuf(Blob, append, Blob->BufUsed + nBytes); + + pos = *Pos; -/** - * @brief measure the number of glyphs in an UTF8 string... - * @param str string to measure - * @returns the length of str - */ -long StrBuf_Utf8StrLen(StrBuf *Buf) -{ - int n = 0; - int m = 0; - char *aptr, *eptr; + if (pos != NULL) + len = pos - IOBuf->buf; + rlen = IOBuf->BufUsed - len; - if ((Buf == NULL) || (Buf->BufUsed == 0)) - return 0; - aptr = Buf->buf; - eptr = Buf->buf + Buf->BufUsed; - while ((aptr < eptr) && (*aptr != '\0')) { - if (Ctdl_IsUtf8SequenceStart(*aptr)){ - m = Ctdl_GetUtf8SequenceLength(aptr, eptr); - while ((aptr < eptr) && (m-- > 0) && (*aptr++ != '\0')) - n ++; + + if ((IOBuf->BufUsed > 0) && + (pos != NULL) && + (pos < IOBuf->buf + IOBuf->BufUsed)) + { + if (rlen < nBytes) { + memcpy(Blob->buf + Blob->BufUsed, pos, rlen); + Blob->BufUsed += rlen; + Blob->buf[Blob->BufUsed] = '\0'; + nAlreadyRead = nRead = rlen; + *Pos = NULL; } - else { - n++; - aptr++; + if (rlen >= nBytes) { + memcpy(Blob->buf + Blob->BufUsed, pos, nBytes); + Blob->BufUsed += nBytes; + Blob->buf[Blob->BufUsed] = '\0'; + if (rlen == nBytes) { + *Pos = NULL; + FlushStrBuf(IOBuf); + } + else + *Pos += nBytes; + return nBytes; } - } - return n; -} -/** - * @brief cuts a string after maxlen glyphs - * @param str string to cut to maxlen glyphs - * @param maxlen how long may the string become? - * @returns pointer to maxlen or the end of the string - */ -long StrBuf_Utf8StrCut(StrBuf *Buf, int maxlen) -{ - char *aptr, *eptr; - int n = 0, m = 0; + FlushStrBuf(IOBuf); + *Pos = NULL; + if (IOBuf->BufSize < nBytes - nRead) + IncreaseBuf(IOBuf, 0, nBytes - nRead); + ptr = IOBuf->buf; - aptr = Buf->buf; - eptr = Buf->buf + Buf->BufUsed; - while ((aptr < eptr) && (*aptr != '\0')) { - if (Ctdl_IsUtf8SequenceStart(*aptr)){ - m = Ctdl_GetUtf8SequenceLength(aptr, eptr); - while ((m-- > 0) && (*aptr++ != '\0')) - n ++; + len = Blob->BufUsed; + + fdflags = fcntl(*fd, F_GETFL); + IsNonBlock = (fdflags & O_NONBLOCK) == O_NONBLOCK; + if (IsNonBlock) + MaxTries = 1000; + else + MaxTries = 100000; + + nBytes -= nRead; + nRead = 0; + while ((nSuccessLess < MaxTries) && + (nRead < nBytes) && + (*fd != -1)) { + if (IsNonBlock) + { + tv.tv_sec = 1; + tv.tv_usec = 0; + + FD_ZERO(&rfds); + FD_SET(*fd, &rfds); + if (select(*fd + 1, &rfds, NULL, NULL, &tv) == -1) { + *Error = strerror(errno); + close (*fd); + *fd = -1; + if (*Error == NULL) + *Error = ErrRBLF_SelectFailed; + return -1; + } + if (! FD_ISSET(*fd, &rfds) != 0) { + nSuccessLess ++; + continue; + } } - else { - n++; - aptr++; + rlen = read(*fd, + ptr, + IOBuf->BufSize - (ptr - IOBuf->buf)); + if (rlen == -1) { + close(*fd); + *fd = -1; + *Error = strerror(errno); + return rlen; + } + else if (rlen == 0){ + if ((check == NNN_TERM) && + (nRead > 5) && + (strncmp(IOBuf->buf + IOBuf->BufUsed - 5, "\n000\n", 5) == 0)) + { + StrBufPlain(Blob, HKEY("\n000\n")); + StrBufCutRight(Blob, 5); + return Blob->BufUsed; + } + else if (!IsNonBlock) + nSuccessLess ++; + else if (nSuccessLess > MaxTries) { + FlushStrBuf(IOBuf); + *Error = ErrRBB_too_many_selects; + return -1; + } + } + else if (rlen > 0) { + nSuccessLess = 0; + nRead += rlen; + ptr += rlen; + IOBuf->BufUsed += rlen; } - if (n > maxlen) { - *aptr = '\0'; - Buf->BufUsed = aptr - Buf->buf; - return Buf->BufUsed; - } } - return Buf->BufUsed; + if (nSuccessLess >= MaxTries) { + FlushStrBuf(IOBuf); + *Error = ErrRBB_too_many_selects; + return -1; + } + if (nRead > nBytes) { + *Pos = IOBuf->buf + nBytes; + } + Blob->buf[Blob->BufUsed] = '\0'; + StrBufAppendBufPlain(Blob, IOBuf->buf, nBytes, 0); + if (*Pos == NULL) { + FlushStrBuf(IOBuf); + } + return nRead + nAlreadyRead; } - /** + * @ingroup StrBuf_IO * @brief extract a "next line" from Buf; Ptr to persist across several iterations * @param LineBuf your line will be copied here. * @param Buf BLOB with lines of text... * @param Ptr moved arround to keep the next-line across several iterations * has to be &NULL on start; will be &NotNULL on end of buffer - * @returns size of copied buffer + * @returns size of remaining buffer */ -int StrBufSipLine(StrBuf *LineBuf, StrBuf *Buf, const char **Ptr) +int StrBufSipLine(StrBuf *LineBuf, const StrBuf *Buf, const char **Ptr) { const char *aptr, *ptr, *eptr; char *optr, *xptr; @@ -3316,3 +4550,38 @@ int StrBufSipLine(StrBuf *LineBuf, StrBuf *Buf, const char **Ptr) return Buf->BufUsed - (ptr - Buf->buf); } + + +/** + * @ingroup StrBuf_IO + * @brief removes double slashes from pathnames + * @param Dir directory string to filter + * @param RemoveTrailingSlash allows / disallows trailing slashes + */ +void StrBufStripSlashes(StrBuf *Dir, int RemoveTrailingSlash) +{ + char *a, *b; + + a = b = Dir->buf; + + while (!IsEmptyStr(a)) { + if (*a == '/') { + while (*a == '/') + a++; + *b = '/'; + b++; + } + else { + *b = *a; + b++; a++; + } + } + if ((RemoveTrailingSlash) && (*(b - 1) != '/')){ + *b = '/'; + b++; + } + *b = '\0'; + Dir->BufUsed = b - Dir->buf; +} + +