GNOME Bugzilla – Bug 410113
remove use of eel string truncation
Last modified: 2018-02-07 08:33:16 UTC
eel_str_middle_truncate and eel_truncate_text_for_menu_item produce invalid UTF-8 (see bug 409271). Even if they were fixed to produce valid UTF-8, they'd still have the problem of truncating composing characters incorrectly etc. I think it's better to remove them completely and just let pango ellipsise the strings when they're displayed.
Created attachment 92266 [details] [review] partial patch
* libnautilus-private/nautilus-bookmark.c: (nautilus_bookmark_menu_item_new): * libnautilus-private/nautilus-customization-data.c: (nautilus_customization_data_new), (nautilus_customization_data_get_next_element_for_display): * libnautilus-private/nautilus-customization-data.h: * src/nautilus-application.c: * src/nautilus-bookmark-list.c: (new_bookmark_from_uri): * src/nautilus-navigation-window-menus.c: (connect_proxy_cb), (refresh_go_menu): * src/nautilus-places-sidebar.c: (bookmarks_drop_uris), (drag_data_received_callback): * src/nautilus-property-browser.c: (make_properties_from_directories): * src/nautilus-window-bookmarks.c: (connect_proxy_cb), (update_bookmarks): * src/nautilus-window-menus.c: (nautilus_menus_append_bookmark_to_menu): * test/test-nautilus-wrap-table.c: (main): Use pango ellipsisation instead of truncating the strings. Bug #410113. Leaving open to eventually get rid of the rest of them too.
I count zero uses of eel_truncate_text* remaining in Nautilus, so I'm closing this. Reopen if someone spots one grep/me missed.
Not fixed. We still use eel_str_middle_truncate in some places.
The title is a bit confusing... As far as I understand, solving this bug should be done by fixing eel_str_middle_truncate and eel_truncate_text_for_menu_item, not by removing the use of it. Or did I miss something ?
*** Bug 466762 has been marked as a duplicate of this bug. ***
*** Bug 409271 has been marked as a duplicate of this bug. ***
Looked into this a bit. I think there are some uses of the function that we can just avoid. They don't add much value. However, there are some that are used indirectly that are very difficult to avoid without major restructuring. For instance: http://git.gnome.org/browse/nautilus/tree/libnautilus-private/nautilus-file-operations.c#n2930 (where %B is a middle truncated string) Asked Owen for some advice and he said suggested it would be easiest to just fix eel_str_middle_truncate to use pango_get_log_attrs() and correct utf-8 operations - is_char_break isn't exactly right for places where you can cut the string to ellipsize, but close enough for practical purposes.
Hmm, it seems that some of the issues were fixed during the gio port. http://git.gnome.org/browse/eel/commit/eel/eel-string.c?id=59511dfee330102ee42e09bef37cbc9cd12973c6 Can someone please attach a use case that is still broken?
Bug 410113 listed ääääääääää as a broken case, but that's fixed and we have a test for that. Since nobody mentioned a broken case for 4 years, I assume this is obsolete.