Author: julianfoad Date: Thu Jul 8 10:41:42 2010 New Revision: 961703 URL: http://svn.apache.org/viewvc?rev=961703&view=rev Log: * subversion/include/private/svn_eol_private.h (svn_eol__detect_eol, svn_eol__detect_file_eol): Mention two bugs and two suggestions in the doc strings.
Modified: subversion/trunk/subversion/include/private/svn_eol_private.h Modified: subversion/trunk/subversion/include/private/svn_eol_private.h URL: http://svn.apache.org/viewvc/subversion/trunk/subversion/include/private/svn_eol_private.h?rev=961703&r1=961702&r2=961703&view=diff ============================================================================== --- subversion/trunk/subversion/include/private/svn_eol_private.h (original) +++ subversion/trunk/subversion/include/private/svn_eol_private.h Thu Jul 8 10:41:42 2010 @@ -60,6 +60,9 @@ svn_eol__find_eol_start(char *buf, apr_s * * The returned string is statically allocated. * + * ### [JAF] Does it make sense for this to use "[buf, end)" while + * svn_eol__find_eol_start() uses "buf, len"? + * * @since New in 1.7 */ const char * @@ -78,6 +81,14 @@ svn_eol__detect_eol(char *buf, char *end * The string @a *eol is statically allocated. * Do temporary allocations in @a pool. * + * ### [JAF] Apparent bug: implementation suffers from the "one byte of a + * two-byte sequence" issue noted for svn_eol__detect_eol(). + * + * ### [JAF] Apparent bug: if file pointer is at EOF, returns "success" + * without initializing @a *eol. + * + * ### [JAF] Wouldn't it make more sense to accept a stream than a file? + * * @since New in 1.7 */ svn_error_t * svn_eol__detect_file_eol(const char **eol, apr_file_t *file, apr_pool_t *pool);