Branch: refs/heads/master Author: Colomban Wendling ban@herbesfolles.org Committer: Colomban Wendling ban@herbesfolles.org Date: Sat, 17 Sep 2016 20:46:04 UTC Commit: 9d98ddc2e90cff84b7fa790da8e3e9f34d3aa001 https://github.com/geany/geany-plugins/commit/9d98ddc2e90cff84b7fa790da8e3e9...

Log Message: ----------- git-changebar: Fix handling of documents with a Unicode BOM

Fixes #482.

Modified Paths: -------------- git-changebar/src/gcb-plugin.c

Modified: git-changebar/src/gcb-plugin.c 52 lines changed, 44 insertions(+), 8 deletions(-) =================================================================== @@ -598,14 +598,17 @@ encoding_needs_coversion (const gchar *encoding) * @brief Converts encoding * @param buffer Input and output buffer * @param length Input and output buffer length + * @param free_buffer whether @p buffer should be freed when replaced * @param to Target encoding * @param from Source encoding * @param error Return location for errors, or @c NULL - * @returns @c TRUE if the conversion succeeded and the buffer has been - * replaced and should be freed, or @c FALSE otherwise. + * @returns @c TRUE if @p buffer should be freed, or @c FALSE otherwise. * * @warning This function has a very weird API, but it is practical for * how it's used. + * @note the only way to tell if the conversion succeeded when @p free_buffer + * is @c TRUE is to compare the output value of @buffer against the + * one it had as an input value. * * Converts between encodings (using g_convert()) in-place. * @@ -615,12 +618,13 @@ encoding_needs_coversion (const gchar *encoding) * of memory that should be freed with @c g_free(). * If the conversion succeeds, it replaces @p buffer and @p length with * the converted values and returns @c TRUE. If it fails, it does not - * modify the values and returns @c FALSE so that the passed-in + * modify the values and returns @p free_buffer so that the passed-in * variables can be used as a fallback if the conversion was optional. */ static gboolean convert_encoding_inplace (gchar **buffer, gsize *length, + gboolean free_buffer, const gchar *to, const gchar *from, GError **error) @@ -630,11 +634,38 @@ convert_encoding_inplace (gchar **buffer, NULL, &tmp_len, error);

if (tmp_buf) { + if (free_buffer) { + g_free (*buffer); + } + *buffer = tmp_buf; *length = tmp_len; + free_buffer = TRUE; }

- return tmp_buf != NULL; + return free_buffer; +} + +static gboolean +add_utf8_bom (gchar **buffer, + gsize *length, + gboolean free_buffer) +{ + gchar *new_buf = g_malloc (*length + 3); + + new_buf[0] = (gchar) 0xef; + new_buf[1] = (gchar) 0xbb; + new_buf[2] = (gchar) 0xbf; + memcpy (&new_buf[3], *buffer, *length); + + if (free_buffer) { + g_free (*buffer); + } + + *buffer = new_buf; + *length += 3; + + return TRUE; }

static int @@ -653,10 +684,15 @@ diff_buf_to_doc (const git_buf *old_buf, buf = (gchar *) scintilla_send_message (sci, SCI_GETCHARACTERPOINTER, 0, 0); len = sci_get_length (sci);

+ /* add the BOM if needed */ + if (doc->has_bom) { + /* UTF-8 BOM, converted below */ + free_buf = add_utf8_bom (&buf, &len, free_buf); + } /* convert the buffer back to in-file encoding if necessary */ if (encoding_needs_coversion (doc->encoding)) { - free_buf = convert_encoding_inplace (&buf, &len, doc->encoding, "UTF-8", - NULL); + free_buf = convert_encoding_inplace (&buf, &len, free_buf, + doc->encoding, "UTF-8", NULL); }

/* no context lines, and no need to bother about binary checks */ @@ -734,8 +770,8 @@ get_widget_for_buf_range (GeanyDocument *doc,

/* convert the buffer to UTF-8 if necessary */ if (encoding_needs_coversion (doc->encoding)) { - free_buf = convert_encoding_inplace (&buf, &buf_len, "UTF-8", doc->encoding, - NULL); + free_buf = convert_encoding_inplace (&buf, &buf_len, free_buf, + "UTF-8", doc->encoding, NULL); }

scintilla_send_message (sci, SCI_ADDTEXT, buf_len, (glong) buf);

-------------- This E-Mail was brought to you by github_commit_mail.py (Source: https://github.com/geany/infrastructure).