summaryrefslogtreecommitdiff
Commit message (Collapse)AuthorAgeFilesLines
* nautilus-directory: handle change eventsBUG_usbNelson Benítez León2021-05-223-0/+53
| | | | | | | | | | | | | | | | | | | | | for files that were still not known to nautilus because they were in the midst of being added to Nautilus. This happens when eg. inotify fires two consecutive NEW + CHANGE events for a new file, and it's important to handle the CHANGE event as the file may have changed its properties with respect the ones reported at the time the NEW event fired. A case this happens is for mountpoint directories of removable devices, as seen in the referenced bug below. We now queue CHANGE events received for files unknown to Nautilus only when the parent folder is currently adding new files. When the folder finish adding the new files, we then process the list of queued files. If there are still files unknown to Nautilus in this list, they will be ignored as before. https://bugzilla.gnome.org/show_bug.cgi?id=703179 Issue #1576
* Update Norwegian Bokmål translationKjartan Maraas2021-05-111-64/+19
|
* Update Occitan translationQuentin PAGÈS2021-05-101-2188/+2304
|
* Updated Spanish translationDaniel Mustieles2021-05-101-207/+209
|
* Update Persian translationDanial Behzadi2021-05-061-651/+1270
|
* Update Persian translationDanial Behzadi2021-05-061-1363/+748
|
* files-view: Set PWD to current dir to run scriptsAntónio Fernandes2021-05-021-0/+5
| | | | | | | | | | | | | | When running an executable text file as a program, it's reasonable to expect that the directory currently displayed by the file browser becomes the current working directory for that program. This used to be handled correctly by the activation action. While taking "Run as a Program" out of activation into a standalone action, this behavior was left behind. Let's add it back to the standalone action to fix the regression. Fixes https://gitlab.gnome.org/GNOME/nautilus/-/issues/1842
* Post release version bumpOndrej Holy2021-04-301-1/+1
|
* Release version 40.140.1Ondrej Holy2021-04-302-1/+8
|
* eel-vfs-extensions: Document get_extension_offset1841-crashing-when-copying-folder-with-same-name-and-click-to-renameAntónio Fernandes2021-04-301-0/+9
| | | | | Mostly to remember callers must be ready to handle a NULL return value, trying to prevent more bugs like the one fixed by the previous commit.
* file-conflict-dialog: Fix crash on extensionless filenamesAntónio Fernandes2021-04-301-6/+19
| | | | | | | | | | | | | | The new filename entry has the automatically suggested part of the name selected for convenience since 9108b4028b7f476b598bc664079fd6d115801b66 However, the logic for this relies on extensions offsets, causing crashes when failing to handle the extensionless case, which is common for folders. So, improve the logic here to handle extensionless filenames, looking at their length instead. Fixes https://gitlab.gnome.org/GNOME/nautilus/-/issues/1841
* file-operations: Return NULL if password dialog was cancelledOndrej Holy2021-04-221-2/+7
| | | | | | | | | | Currently, an empty string is passed to gnome-autoar if the archive extraction password dialog is cancelled. This is not problem currently as immediately the `abort_job` function is called, however, it would be nice to return `NULL` to make obvious that the prompt dialog was cancelled. https://gitlab.gnome.org/GNOME/nautilus/-/merge_requests/657#note_1084750
* file-operations: Fix crashes when extracting encrypted archivesOndrej Holy2021-04-221-22/+52
| | | | | | | | | | | When extracting encrypted archives, Nautilus randomly crashes when opening the password dialog. This happens pretty often under X11, but I have already seen a few crashes on Wayland as well. I suppose that all those crashes are caused because the password dialog is constructed on the extraction thread and GTK is not thread-safe. Let's modify the code to call GTK on the main thread only to fix those crashes. Fixes: https://gitlab.gnome.org/GNOME/nautilus/-/issues/1825
* gtkplacesview: Update to the latest codeOndrej Holy2021-04-221-0/+9
| | | | Fixes: https://gitlab.gnome.org/GNOME/nautilus/-/issues/1811
* ci: Specify test dependencies to fix pipelineOndrej Holy2021-04-211-0/+9
| | | | | | | | | The `validate-desktop`, `validate-desktop-autorun-software` and `validate-appdata` meson tests fail currently because of the updated meson. This is because of the recent change, which requires explicit specification of dependencies. Let's specify them to fix the pipeline. https://gitlab.gnome.org/GNOME/gnome-build-meta/-/issues/372
* Update Slovak translationDušan Kazik2021-04-191-985/+1011
|
* uncrustify: Enforce single space after control flow keywordsAntónio Fernandes2021-04-195-6/+7
| | | | | This is the style we use, but we were missing the uncrustify configuration for it.
* Update Nepali translationPawan Chitrakar2021-04-171-2092/+2258
|
* flatpak: Drop network requirementOndrej Holy2021-04-113-3/+3
| | | | | | | | | | | The network access is no more needed for gvfs support as it currently uses named sockets instead of abstract ones. It is enough to allow access to `xdg-run/gvfsd` folder, where the named sockets reside. I'm not aware of other reasons for allowing network access. Let's drop the network access. This also workarounds "Too many open files" errors caused by leftover sockets (GNOME/gvfs#542). Relates: https://gitlab.gnome.org/GNOME/gvfs/-/issues/515
* Update Portuguese translationHugo Carvalho2021-04-081-217/+222
|
* .desktop: claim handling of rar mime typesSimental Magana, Marcos2021-04-031-1/+1
| | | | | | | | | This addition follows the logic of adding back RAR support in gnome-autoar[1] Closes #1813 [1] - https://gitlab.gnome.org/GNOME/gnome-autoar/-/commit/9356fae1a735b9f8e9daf5633b9709596543ff2e
* Update Chinese (Taiwan) translationChao-Hsiung Liao2021-04-031-975/+992
|
* Update Hebrew translationYosef Or Boczko2021-03-281-1003/+1014
|
* Update Swedish translationAnders Jonsson2021-03-231-203/+208
|
* Post release version bumpOndrej Holy2021-03-201-1/+1
|
* Release version 40.040.0Ondrej Holy2021-03-203-4/+4
| | | | | | | I made 40 release yesterday by mistake. Let's use 40.0 instead to ensure the correct sort order. Fixes: https://gitlab.gnome.org/GNOME/nautilus/-/issues/1807
* Post release version bumpOndrej Holy2021-03-191-1/+1
|
* Release version 4040Ondrej Holy2021-03-192-1/+5
|
* Update Vietnamese translationNgọc Quân Trần2021-03-191-728/+737
|
* Update Dutch translationHannie Dumoleyn2021-03-161-982/+1018
|
* Update Italian translationMilo Casagrande2021-03-151-947/+801
|
* Update Polish translationPiotr Drąg2021-03-131-983/+844
|
* Update Norwegian Bokmål translationKjartan Maraas2021-03-131-981/+839
|
* Post release version bumpOndrej Holy2021-03-131-1/+1
|
* Release version 40.rc40.rcOndrej Holy2021-03-132-1/+9
|
* mime-actions: Revert unintended string freeze breakAntónio Fernandes2021-03-121-2/+3
| | | Reuse existing strings with approximate meaning.
* Update Brazilian Portuguese translationEnrico Nicoletto2021-03-121-222/+211
|
* Update Ukrainian translationYuri Chornoivan2021-03-121-207/+217
|
* mime-action: Fix reinstated regressionAntónio Fernandes2021-03-121-1/+1
| | | | | | | Commit [1] has reinstated a regression from the earlier commit [2]. [1] 080f83385ff79a8be54ee31e7a45422138226f1f [2] 02ba130f3d61fae414459cb308adf270edb01e6e
* mime-actions: Don't leak errorwip/hadess/reinstate-multi-file-launchingAntónio Fernandes2021-03-121-1/+1
|
* mime-actions: Add all portal-opened files to recentsAntónio Fernandes2021-03-121-1/+4
| | | | | | | | | | | While sandboxed, we add the first file in the queue to recents. This has two problems: * The file is added to recents even if we fail to launch. * If opening multiple files, we forget to add the others to recents Instead, add each file to recents for each successful launch.
* mime-actions: Drop dead code pathAntónio Fernandes2021-03-121-11/+0
| | | | | Now that the async files opening path is only used while sandboxed, the conditions for these if blocks (not being sanboxed) is never met.
* mime-actions: Check flatpak-info only onceAntónio Fernandes2021-03-121-5/+16
|
* mime-actions: Open files as groups if not sandboxedAntónio Fernandes2021-03-121-23/+215
| | | | | | | | | | | | | | | | | | While sandboxed, we open files using the OpenURI portal, and we don't know which app is the default handler app for each file. As such, we have given up group-launching files with the same default handler when adapting nautilus to being sandboxed.[0] But this resulted in a feature regression in the non-sandboxed case, which is still the common case in production. Reinstate the code for the old behaviour[1], but keep the current behaviour when running in inside a flatpak sandbox. Closes https://gitlab.gnome.org/GNOME/nautilus/-/issues/117 [0] f5206a6daf0991d91e885a28bb66795a8ae12a41 [1] based on revert patch with revert conflicts resolved by hadess
* window: Handle NULL active_slot pointerAntónio Fernandes2021-03-121-3/+12
| | | | | | This pointer may be NULL. Usually this may happen only during window initialization and destruction. However, for robustness, make sure every use either handles a NULL pointer or asserts it's non-NULL.
* window: Add weak reference to the active slotAntónio Fernandes2021-03-121-3/+3
| | | | | | | | | | | | | | | We used to explicitly set the active slot when closing a tab. However, we now let GtkNotebook pick the next tab, and wait for ::switch-tab to set the active_slot field to the one GtkNotebook picked, thanks to commit 475684ac9e556b144da594bf25581560d4fa5a7f. However, if the closed tab was the only tab in this window, then ::switch-tab is never called, so active_slot becomes a dangling pointer, crashing the application when trying to close the window. Use a weak reference to ensure the pointer is set to NULL in that case. Fixes https://gitlab.gnome.org/GNOME/nautilus/-/issues/1759
* Update Korean translationChangwoo Ryu2021-03-121-978/+991
|
* Update Persian translationDanial Behzadi2021-03-101-765/+765
|
* Update Friulian translationFabio Tomat2021-03-101-771/+776
|
* Update German translationTim Sabsch2021-03-081-139/+138
|
View Lorry Controller Status