Control: tag -1 moreinfo Hi Pablo, Pablo Barciela <scow@riseup.net> (2017-04-27): > 1) Fix: open new window with tree view in side panel (Closes: #851523). > > In the side pane, with tree view, right click in a item, click in "open in > new window". > -without the patch, it shows in the same window > -with the patch, as expected, it open new window Yeah, this looks a bit silly for a file manager not to be abe to do something so basic. From reading the diff, it looks like one function was traded for another one, more generic, with different flags. It would be helpful to have cleaner diffs, e.g. removing lines instead of removing them to replace them with a commented out version, e.g. avoid this: - fm_tree_view_activate_file (view, view->details->popup_file, CAJA_WINDOW_OPEN_FLAG_NEW_WINDOW); + /* fm_tree_view_activate_file (view, view->details->popup_file, CAJA_WINDOW_OPEN_FLAG_NEW_WINDOW); */ > 2) Don't crash on toggling "Show hidden and backup files" (Closes: #797723). > > easy to reproduce with gdb > edit -> preferences -> show hidden and backup files > caja crashes randomly activating/deactivating the checkbox Yep, this is sad as well. So, a function renamed because it no longer handles a single event, but two instead (the original one for trash, and the new one for preferences)? Out of curiosity, are the disconnected signals reconnected properly afterwards? [Trying to find appropriate line-wrapping, long function names ahead…] The trash monitor is handled through: caja_window_initialize_trash_icon_monitor() called by: caja_window_initialize_menus() which also contains a g_signal_connect_swapped() call with: G_CALLBACK(show_hidden_files_preference_callback) as a parameter, so we have the required symmetry for both signals, being: - set up from caja_window_initialize_menus() - and unset from caja_window_finalize_menus()? > 3) Allow the user to drag'n'drop files into the bookmark section. (Closes: > #786395). > > We can dran'n'drop files to everywhere in the side pane except bookmarks, > this is the fix to work too with bookmarks As Adam mentioned, this is borderline feature addition, plus the diff contains a lot of whitespace noise: part of hunk 2, all of hunks 3 and 4, part of hunk 5. Even with a cleaner patch, I'm wondering whether regressions could appear with drag'n'dropping into places which haven't been considered in the implemented checks. > 4) Filename font color now gets picked up from theme correctly for all > themes. (Closes: #770760). The patch itself doesn't make it obvious what the changes do… This is slightly more informative: https://github.com/mate-desktop/caja/pull/526/commits/828aea9083e19cec1a712c349285553197bc1c6f so at least having “icon container: restore original font color select logic” in the patch description would have helped. Ditto for: https://github.com/mate-desktop/caja/pull/526/commits/c74212b4630767b3b11b41cb26a8df20090096f4 with “eel: never try to block background change signal” Ditto for: https://github.com/mate-desktop/caja/pull/526/commits/ce7cc9580809d4017c74b0128f7e82f94eb173d9 and “icon container: don't set label colors right after widget realize” → Please mention all three commit (short) messages after the “taken from” line in your fourth patch. > >The above bug is filed as minor severity. In fact, the highest severity > >of any of them is currently "normal". Is that correct? > > yes, it was reported as severity minor, but the font color black in dark > themes is erroneous, and the patch fixes it. I'm not going to debate each severity in turn, but I can understand how the combination of these 4 bugs makes you want to see an update in jessie. If you can confirm my understanding of patches 1 & 2, can lower the noise in patch 3, and can improve the documentation in patch 4, feel free to post an updated debdiff for a new review. Tagging this jessie-pu request with moreinfo in the meanwhile. KiBi.
Attachment:
signature.asc
Description: Digital signature