]> code.citadel.org Git - citadel.git/blobdiff - libcitadel/lib/stringbuf.c
* reference free handler (Don't free payload ;-)
[citadel.git] / libcitadel / lib / stringbuf.c
index e2dae77306fc523734fb38282cc505de2b34f8aa..ab7dfb88661cbd34a709e4f329c2b57bfa4b11d7 100644 (file)
@@ -1,3 +1,4 @@
+#include "../sysdep.h"
 #include <ctype.h>
 #include <errno.h>
 #include <string.h>
 #include <stdarg.h>
 #include "libcitadel.h"
 
+#ifdef HAVE_ICONV
+#include <iconv.h>
+#endif
+
+#ifdef HAVE_ZLIB
+#include <zlib.h>
+#endif
+
+
+#ifdef HAVE_ZLIB
+#include <zlib.h>
+int ZEXPORT compress_gzip(Bytef * dest, size_t * destLen,
+                          const Bytef * source, uLong sourceLen, int level);
+#endif
 
 /**
  * Private Structure for the Stringbuffer
@@ -1125,6 +1140,29 @@ void StrBufCutRight(StrBuf *Buf, int nChars)
 }
 
 
+/*
+ * Strip leading and trailing spaces from a string; with premeasured and adjusted length.
+ * buf - the string to modify
+ * 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 (delta > 0) StrBufCutLeft(Buf, delta);
+
+       if (Buf->BufUsed == 0) return;
+       while (isspace(Buf->buf[Buf->BufUsed - 1])){
+               Buf->BufUsed --;
+       }
+       Buf->buf[Buf->BufUsed] = '\0';
+}
+
+
 void StrBufUpCase(StrBuf *Buf) 
 {
        char *pch, *pche;
@@ -1297,7 +1335,7 @@ int CompressBuffer(StrBuf *Buf)
                          &compressed_len,
                          (Bytef *) Buf->buf,
                          (uLongf) Buf->BufUsed, Z_BEST_SPEED) == Z_OK) {
-               if (!ConstBuf)
+               if (!Buf->ConstBuf)
                        free(Buf->buf);
                Buf->buf = compressed_data;
                Buf->BufUsed = compressed_len;
@@ -1462,3 +1500,363 @@ void StrBufReplaceChars(StrBuf *buf, char search, char replace)
                        buf->buf[i] = replace;
 
 }
+
+
+
+/*
+ * Wrapper around iconv_open()
+ * Our version adds aliases for non-standard Microsoft charsets
+ * such as 'MS950', aliasing them to names like 'CP950'
+ *
+ * tocode      Target encoding
+ * 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);
+               }
+       }
+       *(iconv_t *)pic = ic;
+#endif
+}
+
+
+
+static inline char *FindNextEnd (StrBuf *Buf, char *bptr)
+{
+       char * end;
+       /* Find the next ?Q? */
+       if (Buf->BufUsed - (bptr - Buf->buf)  < 6)
+               return NULL;
+
+       end = strchr(bptr + 2, '?');
+
+       if (end == NULL)
+               return NULL;
+
+       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, "?=");
+       }
+       else
+               /* sort of half valid encoding, try to find an end. */
+               end = strstr(bptr, "?=");
+       return end;
+}
+
+
+void StrBufConvert(StrBuf *ConvertBuf, StrBuf *TmpBuf, void *pic)
+{
+#ifdef HAVE_ICONV
+       int BufSize;
+       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 */
+
+
+       if (ConvertBuf->BufUsed > TmpBuf->BufSize)
+               IncreaseBuf(TmpBuf, 0, ConvertBuf->BufUsed);
+
+       ic = *(iconv_t*)pic;
+       ibuf = ConvertBuf->buf;
+       ibuflen = ConvertBuf->BufUsed;
+       obuf = TmpBuf->buf;
+       obuflen = TmpBuf->BufSize;
+       
+       iconv(ic, &ibuf, &ibuflen, &obuf, &obuflen);
+
+       /* little card game: wheres the red lady? */
+       ibuf = ConvertBuf->buf;
+       BufSize = ConvertBuf->BufSize;
+
+       ConvertBuf->buf = TmpBuf->buf;
+       ConvertBuf->BufSize = TmpBuf->BufSize;
+       ConvertBuf->BufUsed = TmpBuf->BufSize - obuflen;
+       ConvertBuf->buf[ConvertBuf->BufUsed] = '\0';
+       
+       TmpBuf->buf = ibuf;
+       TmpBuf->BufSize = BufSize;
+       TmpBuf->BufUsed = 0;
+       TmpBuf->buf[0] = '\0';
+#endif
+}
+
+
+
+
+inline static void DecodeSegment(StrBuf *Target, 
+                                StrBuf *DecodeMe, 
+                                char *SegmentStart, 
+                                char *SegmentEnd, 
+                                StrBuf *ConvertBuf,
+                                StrBuf *ConvertBuf2)
+{
+       StrBuf StaticBuf;
+       char charset[128];
+       char encoding[16];
+       iconv_t ic = (iconv_t)(-1);
+
+       /* 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);
+       extract_token(encoding, SegmentStart, 2, '?', sizeof encoding);
+       StrBufExtract_token(ConvertBuf, &StaticBuf, 3, '?');
+       
+       if (*encoding == 'B') { /**< base64 */
+               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++;
+               }
+               
+               ConvertBuf2->BufUsed = CtdlDecodeQuotedPrintable(
+                       ConvertBuf2->buf, 
+                       ConvertBuf->buf,
+                       ConvertBuf->BufUsed);
+       }
+       else {
+               StrBufAppendBuf(ConvertBuf2, ConvertBuf, 0);
+       }
+
+       ctdl_iconv_open("UTF-8", charset, &ic);
+       if (ic != (iconv_t)(-1) ) {             
+               StrBufConvert(ConvertBuf2, ConvertBuf, &ic);
+               StrBufAppendBuf(Target, ConvertBuf2, 0);
+               iconv_close(ic);
+       }
+       else {
+               StrBufAppendBufPlain(Target, HKEY("(unreadable)"), 0);
+       }
+}
+/*
+ * Handle subjects with RFC2047 encoding such as:
+ * =?koi8-r?B?78bP0s3Mxc7JxSDXz9rE1dvO2c3JINvB0sHNySDP?=
+ */
+void StrBuf_RFC822_to_Utf8(StrBuf *Target, StrBuf *DecodeMe, const StrBuf* DefaultCharset)
+{
+       StrBuf *ConvertBuf, *ConvertBuf2;
+       char *start, *end, *next, *nextend, *ptr;
+       iconv_t ic = (iconv_t)(-1) ;
+       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 */
+       
+       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; i<DecodeMe->BufUsed; ++i) {
+               if ((DecodeMe->buf[i] < 32) || (DecodeMe->buf[i] > 126)) {
+                       illegal_non_rfc2047_encoding = 1;
+                       break;
+               }
+       }
+
+       ConvertBuf = NewStrBufPlain(NULL, StrLength(DecodeMe));
+       if (illegal_non_rfc2047_encoding) {
+               if ( (strcasecmp(ChrPtr(DefaultCharset), "UTF-8")) && 
+                    (strcasecmp(ChrPtr(DefaultCharset), "us-ascii")) ) {
+                       ctdl_iconv_open("UTF-8", ChrPtr(DefaultCharset), &ic);
+                       if (ic != (iconv_t)(-1) ) {
+                               long BufSize;
+                               ibuf = DecodeMe->buf;
+                               obuf = ConvertBuf->buf;
+                               ibuflen = DecodeMe->BufUsed;
+                               obuflen = ConvertBuf->BufSize;
+
+                               iconv(ic, &ibuf, &ibuflen, &obuf, &obuflen);
+                               /* little card game: wheres the red lady? */
+                               ibuf = DecodeMe->buf;
+                               BufSize = DecodeMe->BufSize;
+                               DecodeMe->buf = ConvertBuf->buf;
+                               DecodeMe->BufSize = ConvertBuf->BufSize;
+                               DecodeMe->BufUsed = ConvertBuf->BufSize - obuflen;
+                               DecodeMe->buf[DecodeMe->BufUsed] = '\0';
+
+                               ConvertBuf->buf = ibuf;
+                               ConvertBuf->BufSize = BufSize;
+                               ConvertBuf->BufUsed = 0;
+                               ConvertBuf->buf[0] = '\0';
+
+                               iconv_close(ic);
+                       }
+               }
+       }
+
+       /* pre evaluate the first pair */
+       nextend = end = NULL;
+       len = StrLength(DecodeMe);
+       start = strstr(DecodeMe->buf, "=?");
+       eptr = DecodeMe->buf + DecodeMe->BufUsed;
+       if (start != NULL) 
+               end = FindNextEnd (DecodeMe, start);
+       else {
+               StrBufAppendBuf(Target, DecodeMe, 0);
+               FreeStrBuf(&ConvertBuf);
+               return;
+       }
+
+       ConvertBuf2 = NewStrBufPlain(NULL, StrLength(DecodeMe));
+
+       /*
+        * 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, 
+                             DecodeMe, 
+                             start, 
+                             end, 
+                             ConvertBuf,
+                             ConvertBuf2);
+               
+               next = strstr(end, "=?");
+               nextend = NULL;
+               if ((next != NULL) && 
+                   (next < eptr))
+                       nextend = FindNextEnd(DecodeMe, next);
+               if (nextend == NULL)
+                       next = NULL;
+
+               /* 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 (ptr == next)
+                       {
+                               memmove (end + 2,
+                                        next,
+                                        len - (next - start));
+                               
+                               /* now terminate the gab at the end */
+                               delta = (next - end) - 2;
+                               DecodeMe->BufUsed -= delta;
+                               DecodeMe->buf[DecodeMe->BufUsed] = '\0';
+
+                               /* move next to its new location. */
+                               next -= delta;
+                               nextend -= delta;
+                       }
+               }
+               /* our next-pair is our new first pair now. */
+               start = next;
+               end = nextend;
+       }
+       FreeStrBuf(&ConvertBuf);
+       FreeStrBuf(&ConvertBuf2);
+}
+
+
+
+long StrBuf_Utf8StrLen(StrBuf *Buf)
+{
+       return Ctdl_Utf8StrLen(Buf->buf);
+}
+
+long StrBuf_Utf8StrCut(StrBuf *Buf, int maxlen)
+{
+       char *CutAt;
+
+       CutAt = Ctdl_Utf8StrCut(Buf->buf, maxlen);
+       if (CutAt != NULL) {
+               Buf->BufUsed = CutAt - Buf->buf;
+               Buf->buf[Buf->BufUsed] = '\0';
+       }
+       return Buf->BufUsed;    
+}
+
+
+
+int StrBufSipLine(StrBuf *LineBuf, StrBuf *Buf, const char **Ptr)
+{
+       const char *aptr, *ptr, *eptr;
+       char *optr, *xptr;
+
+       if (Buf == NULL)
+               return 0;
+
+       if (*Ptr==NULL)
+               ptr = aptr = Buf->buf;
+       else
+               ptr = aptr = *Ptr;
+
+       optr = LineBuf->buf;
+       eptr = Buf->buf + Buf->BufUsed;
+       xptr = LineBuf->buf + LineBuf->BufSize;
+
+       while ((*ptr != '\n') &&
+              (*ptr != '\r') &&
+              (ptr < eptr))
+       {
+               *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;
+               }
+       }
+       LineBuf->BufUsed = optr - LineBuf->buf;
+       *optr = '\0';       
+       if (*ptr == '\r')
+               ptr ++;
+       if (*ptr == '\n')
+               ptr ++;
+
+       *Ptr = ptr;
+
+       return Buf->BufUsed - (ptr - Buf->buf);
+}