[Python-checkins] python/dist/src/Objects unicodeobject.c, 2.200, 2.201

[doerwalter at users.sourceforge.net]

Update of /cvsroot/python/python/dist/src/Objects
In directory sc8-pr-cvs1:/tmp/cvs-serv21163/Objects

Modified Files:
	unicodeobject.c 
Log Message:
Fix a bug in the memory reallocation code of PyUnicode_TranslateCharmap().
charmaptranslate_makespace() allocated more memory than required for the
next replacement but didn't remember that fact, so memory size was growing
exponentially every time a replacement string is longer that one character.
This fixes SF bug #828737.


Index: unicodeobject.c
===================================================================
RCS file: /cvsroot/python/python/dist/src/Objects/unicodeobject.c,v
retrieving revision 2.200
retrieving revision 2.201
diff -C2 -d -r2.200 -r2.201
*** unicodeobject.c	18 Oct 2003 09:55:08 -0000	2.200
--- unicodeobject.c	24 Oct 2003 14:25:28 -0000	2.201
***************
*** 3223,3239 ****
  Return 0 on success, -1 on error */
  static
! int charmaptranslate_makespace(PyObject **outobj, Py_UNICODE **outp, int *outsize,
      int requiredsize)
  {
!     if (requiredsize > *outsize) {
  	/* remember old output position */
  	int outpos = *outp-PyUnicode_AS_UNICODE(*outobj);
  	/* exponentially overallocate to minimize reallocations */
! 	if (requiredsize < 2 * *outsize)
! 	    requiredsize = 2 * *outsize;
  	if (_PyUnicode_Resize(outobj, requiredsize) < 0)
  	    return -1;
  	*outp = PyUnicode_AS_UNICODE(*outobj) + outpos;
- 	*outsize = requiredsize;
      }
      return 0;
--- 3223,3239 ----
  Return 0 on success, -1 on error */
  static
! int charmaptranslate_makespace(PyObject **outobj, Py_UNICODE **outp,
      int requiredsize)
  {
!     int oldsize = PyUnicode_GET_SIZE(*outobj);
!     if (requiredsize > oldsize) {
  	/* remember old output position */
  	int outpos = *outp-PyUnicode_AS_UNICODE(*outobj);
  	/* exponentially overallocate to minimize reallocations */
! 	if (requiredsize < 2 * oldsize)
! 	    requiredsize = 2 * oldsize;
  	if (_PyUnicode_Resize(outobj, requiredsize) < 0)
  	    return -1;
  	*outp = PyUnicode_AS_UNICODE(*outobj) + outpos;
      }
      return 0;
***************
*** 3246,3257 ****
     Return 0 on success, -1 on error. */
  static
! int charmaptranslate_output(Py_UNICODE c, PyObject *mapping,
!     PyObject **outobj, int *outsize, Py_UNICODE **outp, PyObject **res)
  {
!     if (charmaptranslate_lookup(c, mapping, res))
  	return -1;
      if (*res==NULL) {
  	/* not found => default to 1:1 mapping */
! 	*(*outp)++ = (Py_UNICODE)c;
      }
      else if (*res==Py_None)
--- 3246,3258 ----
     Return 0 on success, -1 on error. */
  static
! int charmaptranslate_output(const Py_UNICODE *startinp, const Py_UNICODE *curinp,
!     int insize, PyObject *mapping, PyObject **outobj, Py_UNICODE **outp,
!     PyObject **res)
  {
!     if (charmaptranslate_lookup(*curinp, mapping, res))
  	return -1;
      if (*res==NULL) {
  	/* not found => default to 1:1 mapping */
! 	*(*outp)++ = *curinp;
      }
      else if (*res==Py_None)
***************
*** 3269,3274 ****
  	else if (repsize!=0) {
  	    /* more than one character */
! 	    int requiredsize = *outsize + repsize - 1;
! 	    if (charmaptranslate_makespace(outobj, outp, outsize, requiredsize))
  		return -1;
  	    memcpy(*outp, PyUnicode_AS_UNICODE(*res), sizeof(Py_UNICODE)*repsize);
--- 3270,3277 ----
  	else if (repsize!=0) {
  	    /* more than one character */
! 	    int requiredsize = (*outp-PyUnicode_AS_UNICODE(*outobj)) +
! 		(insize - (*curinp-*startinp)) +
! 		repsize - 1;
! 	    if (charmaptranslate_makespace(outobj, outp, requiredsize))
  		return -1;
  	    memcpy(*outp, PyUnicode_AS_UNICODE(*res), sizeof(Py_UNICODE)*repsize);
***************
*** 3295,3299 ****
      /* current output position */
      int respos = 0;
-     int ressize;
      char *reason = "character maps to <undefined>";
      PyObject *errorHandler = NULL;
--- 3298,3301 ----
***************
*** 3313,3326 ****
      res = PyUnicode_FromUnicode(NULL, size);
      if (res == NULL)
!         goto onError;
      if (size == 0)
  	return res;
      str = PyUnicode_AS_UNICODE(res);
-     ressize = size;
  
      while (p<endp) {
  	/* try to encode it */
  	PyObject *x = NULL;
! 	if (charmaptranslate_output(*p, mapping, &res, &ressize, &str, &x)) {
  	    Py_XDECREF(x);
  	    goto onError;
--- 3315,3327 ----
      res = PyUnicode_FromUnicode(NULL, size);
      if (res == NULL)
! 	goto onError;
      if (size == 0)
  	return res;
      str = PyUnicode_AS_UNICODE(res);
  
      while (p<endp) {
  	/* try to encode it */
  	PyObject *x = NULL;
! 	if (charmaptranslate_output(startp, p, size, mapping, &res, &str, &x)) {
  	    Py_XDECREF(x);
  	    goto onError;
***************
*** 3341,3345 ****
  	    /* find all untranslatable characters */
  	    while (collend < endp) {
! 	    	if (charmaptranslate_lookup(*collend, mapping, &x))
  		    goto onError;
  		Py_XDECREF(x);
--- 3342,3346 ----
  	    /* find all untranslatable characters */
  	    while (collend < endp) {
! 		if (charmaptranslate_lookup(*collend, mapping, &x))
  		    goto onError;
  		Py_XDECREF(x);
***************
*** 3380,3384 ****
  			char *cp;
  			sprintf(buffer, "&#%d;", (int)*p);
! 			if (charmaptranslate_makespace(&res, &str, &ressize,
  			    (str-PyUnicode_AS_UNICODE(res))+strlen(buffer)+(endp-collend)))
  			    goto onError;
--- 3381,3385 ----
  			char *cp;
  			sprintf(buffer, "&#%d;", (int)*p);
! 			if (charmaptranslate_makespace(&res, &str,
  			    (str-PyUnicode_AS_UNICODE(res))+strlen(buffer)+(endp-collend)))
  			    goto onError;
***************
*** 3396,3400 ****
  		    /* generate replacement  */
  		    repsize = PyUnicode_GET_SIZE(repunicode);
! 		    if (charmaptranslate_makespace(&res, &str, &ressize,
  			(str-PyUnicode_AS_UNICODE(res))+repsize+(endp-collend))) {
  			Py_DECREF(repunicode);
--- 3397,3401 ----
  		    /* generate replacement  */
  		    repsize = PyUnicode_GET_SIZE(repunicode);
! 		    if (charmaptranslate_makespace(&res, &str,
  			(str-PyUnicode_AS_UNICODE(res))+repsize+(endp-collend))) {
  			Py_DECREF(repunicode);
***************
*** 3410,3414 ****
      /* Resize if we allocated to much */
      respos = str-PyUnicode_AS_UNICODE(res);
!     if (respos<ressize) {
  	if (_PyUnicode_Resize(&res, respos) < 0)
  	    goto onError;
--- 3411,3415 ----
      /* Resize if we allocated to much */
      respos = str-PyUnicode_AS_UNICODE(res);
!     if (respos<PyUnicode_GET_SIZE(res)) {
  	if (_PyUnicode_Resize(&res, respos) < 0)
  	    goto onError;