Index: blame.c =================================================================== --- blame.c (revision 1366813) +++ blame.c (working copy) @@ -82,6 +82,7 @@ svn_boolean_t ignore_mime_type; /* name of file containing the previous revision of the file */ const char *last_filename; + const char *last_filename_utf8; /* name of UTF-8 file ( created if format of file 'LAST_FILENAME' is UTF-16. ) */ struct rev *rev; /* the rev for which blame is being assigned during a diff */ struct blame_chain *chain; /* the original blame chain. */ @@ -96,6 +97,7 @@ struct blame_chain *merged_chain; /* the merged blame chain. */ /* name of file containing the previous merged revision of the file */ const char *last_original_filename; + const char *last_original_filename_utf8; /* name of UTF-8 file ( created if format of file 'LAST_ORIGINAL_FILENAME' is UTF-16. ) */ /* pools for files which may need to persist for more than one rev. */ apr_pool_t *filepool; apr_pool_t *prevfilepool; @@ -305,13 +307,58 @@ return SVN_NO_ERROR; } + +/* if format of file 'INPUT_FILENAME' is UTF-16, export it to UTF-8 file and return that file's name. + otherwise return NULL. */ +static const char* +create_utf8_file_if_utf16_file( const char *input_filename, apr_pool_t *pool ) +{ + svn_stringbuf_t *input_buf; + const char *buf_utf8; + apr_size_t len_utf8; + const char* utf8_filename = NULL; + svn_stream_t *stream; + apr_size_t bom_len = 2; + const char* bom; + svn_boolean_t is_utf16; + + if (input_filename == NULL) return NULL; + + /* check UTF 16 BOM */ + SVN_ERR(svn_stream_open_readonly(&stream, input_filename, pool, pool)); + bom = apr_palloc(pool, bom_len); + svn_stream_read(stream, bom, &bom_len); + svn_stream_close(stream); + is_utf16 = (bom_len >= 2 && bom[0] == (char)0xFF && bom[1] == (char)0xFE); + if (!is_utf16) return NULL; + + /* Read : input file => input buf */ + SVN_ERR(svn_stringbuf_from_file2(&input_buf, input_filename, pool)); + + /* Convert : input buf => buf_utf8 */ + len_utf8 = WideCharToMultiByte(CP_UTF8, 0, input_buf->data, input_buf->len/sizeof(WCHAR), 0, 0, 0, 0); + buf_utf8 = apr_palloc(pool, len_utf8); + len_utf8 = WideCharToMultiByte(CP_UTF8, 0, input_buf->data, input_buf->len/sizeof(WCHAR), buf_utf8, len_utf8, 0, 0); + + /* Write : buf_utf8 => utf8 file */ + SVN_ERR(svn_io_write_unique(&utf8_filename, NULL, buf_utf8, len_utf8, svn_io_file_del_on_pool_cleanup, pool)); + + return utf8_filename; +} + static svn_error_t * window_handler(svn_txdelta_window_t *window, void *baton) { struct delta_baton *dbaton = baton; struct file_rev_baton *frb = dbaton->file_rev_baton; struct blame_chain *chain; + + apr_pool_t *filepool; + const char *last_filename = NULL; + const char *curr_filename = NULL; + const char *curr_filename_utf8 = NULL; + const char *last_original_filename = NULL; /* Call the wrapped handler first. */ SVN_ERR(dbaton->wrapped_handler(window, dbaton->wrapped_baton)); @@ -334,9 +381,22 @@ else chain = frb->chain; + /* Set pool for UTF-8 file ( created if current file is UTF-16 ) */ + if (frb->include_merged_revisions && !frb->merged_revision) + filepool = frb->filepool; + else + filepool = frb->currpool; + + /* If last file's original format is UTF-16, use UTF-8 file created at last call. */ + last_filename = frb->last_filename_utf8 ? frb->last_filename_utf8 : frb->last_filename; + + /* If curr file's original format is UTF-16, export it to UTF-8 file and use that file. */ + curr_filename_utf8 = create_utf8_file_if_utf16_file(dbaton->filename, filepool); + curr_filename = curr_filename_utf8 ? curr_filename_utf8 : dbaton->filename; + /* Process this file. */ - SVN_ERR(add_file_blame(frb->last_filename, - dbaton->filename, chain, frb->rev, + SVN_ERR(add_file_blame(last_filename, + curr_filename, chain, frb->rev, frb->diff_options, frb->currpool)); /* If we are including merged revisions, and the current revision is not a @@ -346,8 +406,11 @@ { apr_pool_t *tmppool; - SVN_ERR(add_file_blame(frb->last_original_filename, - dbaton->filename, frb->chain, frb->rev, + /* If last file's original format is UTF-16, use UTF-8 file created at last call. */ + last_original_filename = frb->last_original_filename_utf8 ? frb->last_original_filename_utf8 : frb->last_original_filename; + + SVN_ERR(add_file_blame(last_original_filename, + curr_filename, frb->chain, frb->rev, frb->diff_options, frb->currpool)); /* This filename could be around for a while, potentially, so @@ -359,12 +422,14 @@ frb->last_original_filename = apr_pstrdup(frb->filepool, dbaton->filename); + frb->last_original_filename_utf8 = curr_filename_utf8; } /* Prepare for next revision. */ /* Remember the file name so we can diff it with the next revision. */ frb->last_filename = dbaton->filename; + frb->last_filename_utf8 = curr_filename_utf8; /* Switch pools. */ { @@ -596,6 +661,8 @@ svn_stream_t *stream; const char *target_abspath_or_url; + const char* last_filename; + if (start->kind == svn_opt_revision_unspecified || end->kind == svn_opt_revision_unspecified) return svn_error_create @@ -630,6 +697,8 @@ frb.include_merged_revisions = include_merged_revisions; frb.last_filename = NULL; frb.last_original_filename = NULL; + frb.last_filename_utf8 = NULL; + frb.last_original_filename_utf8 = NULL; frb.chain = apr_palloc(pool, sizeof(*frb.chain)); frb.chain->blame = NULL; frb.chain->avail = NULL; @@ -719,9 +788,12 @@ /* Create a pool for the iteration below. */ iterpool = svn_pool_create(pool); - + + /* If last file's original format is UTF-16, use UTF-8 file created at last call. */ + last_filename = frb.last_filename_utf8 ? frb.last_filename_utf8 : frb.last_filename; + /* Open the last file and get a stream. */ - SVN_ERR(svn_stream_open_readonly(&last_stream, frb.last_filename, + SVN_ERR(svn_stream_open_readonly(&last_stream, last_filename, pool, pool)); stream = svn_subst_stream_translated(last_stream, "\n", TRUE, NULL, FALSE, pool);