]> code.citadel.org Git - citadel.git/blobdiff - libcitadel/lib/stringbuf.c
* add cURL read-callbackhook, so we can read HTTP answers into StrBufs
[citadel.git] / libcitadel / lib / stringbuf.c
index 94728a9bb160417d8386f5dce137d220e18c3886..599b482c0d9b6ff0475fb4cd301429daeb1864e4 100644 (file)
@@ -25,7 +25,7 @@
 int ZEXPORT compress_gzip(Bytef * dest, size_t * destLen,
                           const Bytef * source, uLong sourceLen, int level);
 #endif
-int BaseStrBufSize = SIZ;
+int BaseStrBufSize = 64;
 
 /**
  * Private Structure for the Stringbuffer
@@ -138,6 +138,22 @@ static int IncreaseBuf(StrBuf *Buf, int KeepOriginal, int DestSize)
        return Buf->BufSize;
 }
 
+/**
+ * \brief shrink an _EMPTY_ buffer if its Buffer superseeds threshhold 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) {
+               free(Buf->buf);
+               Buf->buf = (char*) malloc(NewSize);
+               Buf->BufUsed = 0;
+               Buf->BufSize = NewSize;
+       }
+}
+
 /**
  * \brief shrink long term buffers to their real size so they don't waste memory
  * \param Buf buffer to shrink
@@ -548,6 +564,23 @@ void StrBufAppendBufPlain(StrBuf *Buf, const char *AppendBuf, long AppendSize, u
        Buf->buf[Buf->BufUsed] = '\0';
 }
 
+/**
+ * \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
+ */
+size_t CurlFillStrBuf_callback(void *ptr, size_t size, size_t nmemb, void *stream)
+{
+
+       StrBuf *Target;
+
+       Target = stream;
+       if (ptr == NULL)
+               return 0;
+
+       StrBufAppendBufPlain(Target, ptr, size * nmemb, 0);
+       return size * nmemb;
+}
+
 
 /** 
  * \brief Escape a string for feeding out as a URL while appending it to a Buffer
@@ -730,7 +763,7 @@ long StrEscAppend(StrBuf *Target, const StrBuf *Source, const char *PlainIn, int
  * \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, StrBuf *Source, const char *PlainIn)
+void StrMsgEscAppend(StrBuf *Target, const StrBuf *Source, const char *PlainIn)
 {
        const char *aptr, *eiptr;
        char *tptr, *eptr;
@@ -787,6 +820,77 @@ void StrMsgEscAppend(StrBuf *Target, StrBuf *Source, const char *PlainIn)
        *tptr = '\0';
 }
 
+
+
+/*
+ * \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';
+}
+
 /*
  * \brief Append a string, escaping characters which have meaning in JavaScript strings .  
  *
@@ -827,13 +931,17 @@ long StrECMAEscAppend(StrBuf *Target, const StrBuf *Source, const char *PlainIn)
                        bptr = Target->buf + Target->BufUsed;
                }
                else if (*aptr == '"') {
-                       memcpy(bptr, "\\\"", 2);
-                       bptr += 2;
+                       *bptr = '\\';
+                       bptr ++;
+                       *bptr = '"';
+                       bptr ++;
                        Target->BufUsed += 2;
                } else if (*aptr == '\\') {
-                 memcpy(bptr, "\\\\", 2);
-                 bptr += 2;
-                 Target->BufUsed += 2;
+                       *bptr = '\\';
+                       bptr ++;
+                       *bptr = '\\';
+                       bptr ++;
+                       Target->BufUsed += 2;
                }
                else{
                        *bptr = *aptr;
@@ -843,7 +951,7 @@ long StrECMAEscAppend(StrBuf *Target, const StrBuf *Source, const char *PlainIn)
                aptr ++;
        }
        *bptr = '\0';
-       if ((bptr = eptr - 1 ) && !IsEmptyStr(aptr) )
+       if ((bptr == eptr - 1 ) && !IsEmptyStr(aptr) )
                return -1;
        return Target->BufUsed;
 }
@@ -1084,9 +1192,11 @@ int StrBufExtract_token(StrBuf *dest, const StrBuf *Source, int parmnum, char se
                if (*s == separator) {
                        ++current_token;
                }
-               if (len >= dest->BufSize)
+               if (len >= dest->BufSize) {
+                       dest->BufUsed = len;
                        if (!IncreaseBuf(dest, 1, -1))
                                break;
+               }
                if ( (current_token == parmnum) && 
                     (*s != separator)) {
                        dest->buf[len] = *s;
@@ -1235,11 +1345,13 @@ int StrBufExtract_NextToken(StrBuf *dest, const StrBuf *Source, const char **pSt
                if (*s == separator) {
                        ++current_token;
                }
-               if (len >= dest->BufSize)
+               if (len >= dest->BufSize) {
+                       dest->BufUsed = len;
                        if (!IncreaseBuf(dest, 1, -1)) {
                                *pStart = EndBuffer + 1;
                                break;
                        }
+               }
                if ( (current_token == 0) && 
                     (*s != separator)) {
                        dest->buf[len] = *s;
@@ -1542,8 +1654,8 @@ int StrBufTCP_read_buffered_line_fast(StrBuf *Line,
                                      int selectresolution, 
                                      const char **Error)
 {
-       const char *pche;
-       const char *pos;
+       const char *pche = NULL;
+       const char *pos = NULL;
        int len, rlen;
        int nSuccessLess = 0;
        fd_set rfds;
@@ -1576,7 +1688,10 @@ int StrBufTCP_read_buffered_line_fast(StrBuf *Line,
        }
        
        if (pos != NULL) {
-               StrBufCutLeft(IOBuf, (pos - IOBuf->buf));
+               if (pos > pche)
+                       FlushStrBuf(IOBuf);
+               else 
+                       StrBufCutLeft(IOBuf, (pos - IOBuf->buf));
                *Pos = NULL;
        }
        
@@ -1588,6 +1703,7 @@ int StrBufTCP_read_buffered_line_fast(StrBuf *Line,
        if ((fdflags & O_NONBLOCK) == O_NONBLOCK)
                return -1;
 
+       pch = NULL;
        while ((nSuccessLess < timeout) && (pch == NULL)) {
                tv.tv_sec = selectresolution;
                tv.tv_usec = 0;
@@ -1617,7 +1733,13 @@ int StrBufTCP_read_buffered_line_fast(StrBuf *Line,
                                if (IOBuf->BufUsed + 10 > IOBuf->BufSize) {
                                        IncreaseBuf(IOBuf, 1, -1);
                                }
-                               pch = strchr(IOBuf->buf, '\n');
+
+                               pche = IOBuf->buf + IOBuf->BufUsed;
+                               pch = IOBuf->buf;
+                               while ((pch < pche) && (*pch != '\n'))
+                                       pch ++;
+                               if ((pch >= pche) || (*pch == '\0'))
+                                       pch = NULL;
                                continue;
                        }
                }
@@ -1928,6 +2050,36 @@ void StrBufLowerCase(StrBuf *Buf)
        }
 }
 
+/**
+ * \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;
+}
 
 /**
  * \brief unhide special chars hidden to the HTML escaper
@@ -2361,11 +2513,21 @@ static inline char *FindNextEnd (const StrBuf *Buf, char *bptr)
        return end;
 }
 
+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));
+
+}
 
 void StrBufConvert(StrBuf *ConvertBuf, StrBuf *TmpBuf, void *pic)
 {
 #ifdef HAVE_ICONV
-       int BufSize;
+       long trycount = 0;
+       size_t siz;
        iconv_t ic;
        char *ibuf;                     /**< Buffer of characters to be converted */
        char *obuf;                     /**< Buffer for converted characters */
@@ -2375,28 +2537,41 @@ void StrBufConvert(StrBuf *ConvertBuf, StrBuf *TmpBuf, void *pic)
 
        if (ConvertBuf->BufUsed >= TmpBuf->BufSize)
                IncreaseBuf(TmpBuf, 0, ConvertBuf->BufUsed);
-
+TRYAGAIN:
        ic = *(iconv_t*)pic;
        ibuf = ConvertBuf->buf;
        ibuflen = ConvertBuf->BufUsed;
        obuf = TmpBuf->buf;
        obuflen = TmpBuf->BufSize;
        
-       iconv(ic, &ibuf, &ibuflen, &obuf, &obuflen);
+       siz = iconv(ic, &ibuf, &ibuflen, &obuf, &obuflen);
 
-       /* little card game: wheres the red lady? */
-       ibuf = ConvertBuf->buf;
-       BufSize = ConvertBuf->BufSize;
+       if (siz < 0) {
+               if (errno == E2BIG) {
+                       trycount ++;                    
+                       IncreaseBuf(TmpBuf, 0, 0);
+                       if (trycount < 5) 
+                               goto TRYAGAIN;
 
-       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';
+               }
+               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. */
+               }
+
+               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
 }
 
@@ -2575,10 +2750,11 @@ void StrBuf_RFC822_to_Utf8(StrBuf *Target, const StrBuf *DecodeMe, const StrBuf*
                        /* did we find a gab just filled with blanks? */
                        if (ptr == next)
                        {
+                               long gap = next - start;
                                memmove (end + 2,
                                         next,
-                                        len - (next - start));
-                               
+                                        len - (gap));
+                               len -= gap;
                                /* now terminate the gab at the end */
                                delta = (next - end) - 2; ////TODO: const! 
                                ((StrBuf*)DecodeMe)->BufUsed -= delta;
@@ -2611,23 +2787,97 @@ void StrBuf_RFC822_to_Utf8(StrBuf *Target, const StrBuf *DecodeMe, const StrBuf*
        FreeStrBuf(&ConvertBuf2);
 }
 
+/**
+ * \brief evaluate the length of an utf8 special character sequence
+ * \param Char the character to examine
+ * \returns width of utf8 chars in bytes
+ */
+static inline int Ctdl_GetUtf8SequenceLength(char *CharS, char *CharE)
+{
+       int n = 1;
+        char test = (1<<7);
+       
+       while ((n < 8) && ((test & *CharS) != 0)) {
+               test = test << 1;
+               n ++;
+       }
+       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(char Char)
+{
+/** 11??.???? indicates an UTF8 Sequence. */
+       return ((Char & 0xC0) != 0);
+}
 
+/**
+ * \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)
 {
-       return Ctdl_Utf8StrLen(Buf->buf);
+       int n = 0;
+       int m = 0;
+       char *aptr, *eptr;
+
+       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 ++;
+               }
+               else {
+                       n++;
+                       aptr++;
+               }
+                       
+       }
+       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 *CutAt;
+       char *aptr, *eptr;
+       int n = 0, m = 0;
 
-       CutAt = Ctdl_Utf8StrCut(Buf->buf, maxlen);
-       if (CutAt != NULL) {
-               Buf->BufUsed = CutAt - Buf->buf;
-               Buf->buf[Buf->BufUsed] = '\0';
+       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 ++;
+               }
+               else {
+                       n++;
+                       aptr++;
+               }
+               if (n > maxlen) {
+                       *aptr = '\0';
+                       Buf->BufUsed = aptr - Buf->buf;
+                       return Buf->BufUsed;
+               }                       
        }
-       return Buf->BufUsed;    
+       return Buf->BufUsed;
+
 }