Locations refactory to use native Gio.AppIcon's (for GNOME 42 support) #1656

3v1n0 · 2022-01-24T05:43:11Z

Refactor locations handling to have a more native Shell integration.

A part the various improvements and features, the main trigger for this was (as per the main commit message):

As per gjs 1.72 it won't be possible to inject an interface vfunc as it
used to work in previous versions, this would make gnome-shell to crash
as we may try to get the location "apps" ID which is now null (and not
customId anymore).

So, let's re-implement the things in a cleaner way by actually using
real custom AppInfo's that are still based on Gio.DesktopAppInfo
(beause the shell will still need to be such type), but that are
implementing the needed virtual functions of the parent interface.

As per this, we can finally use native shell mount operations to handle
mount options such as mounting encrypted disks.

timocapa · 2022-03-10T21:20:47Z

not sure if this is the right place for this, but some issues ive ran into so far:

dock doesn't hide properly and app names when hovered over don't disappear most of the time, making it actually kinda annoying to use

sorry if this is unrelated here (since im not sure what this commit does)

stuarthayhurst · 2022-03-17T15:51:08Z

not sure if this is the right place for this, but some issues ive ran into so far:

dock doesn't hide properly and app names when hovered over don't disappear most of the time, making it actually kinda annoying to use

sorry if this is unrelated here (since im not sure what this commit does)

If you're referring to this branch specifically, I'm getting the same issues, using the standard release from EGO (Version 71), so I don't think it's related to this

timocapa · 2022-03-20T01:34:32Z

not sure if this is the right place for this, but some issues ive ran into so far:
dock doesn't hide properly and app names when hovered over don't disappear most of the time, making it actually kinda annoying to use
sorry if this is unrelated here (since im not sure what this commit does)

If you're referring to this branch specifically, I'm getting the same issues, using the standard release from EGO (Version 71), so I don't think it's related to this

I'm quite confident this was introduced in a Gnome change a while ago, thought I'd mention it anyway

Etaash-mathamsetty · 2022-03-27T17:20:04Z

could this get merged?

schtobia · 2022-03-31T05:14:20Z

I can confirm this fix for GNOME 42.0-2 as present in Debian sid.

fancypantalons · 2022-03-31T20:33:37Z

Confirmed here as well. Thank you @3v1n0 !

rastersoft · 2022-04-01T21:31:16Z

There is a bug in this patch: in the file docking.js, the property this._monitorIndex is replaced with this.monitorIndex, but there are several references kept to this._monitorIndex, specially in the Desktop Icons integration (which, BTW makes the integration to fail).

phush0 · 2022-04-07T07:44:27Z

There is a bug in animation minimizing window. It will draw two animations one go to the left upper edge of screen and one minimize to dock. It is most visible with minimize Nautilus window

martindmtrv · 2022-04-08T15:17:11Z

Running this branch on my fresh Gnome 42 install all seems swell! Only issue I had was with the trash icon it had a white background. When disabling the "show trash icon" then re-enabling it seems to have fixed that issue

OctoRocket · 2022-04-11T05:28:43Z

Is this pull ready to go? Can't wait to use it on my computer!

Windows changes may not be processed at the same time we got a fm1Client change now, so we should monitor apps signals instead of rely on what happens on the backend

We don't care about exposing windows-changed signal if the file manager app is wrapped to isolate devices. As we're handling signaling manually anyways

Try to just read the trash item count file info before enumerating the trash. It may be just cheaper and will work anyways

Since we've a native trash app info now we can handle the logic inside that. We need to handle data destruction from parent app wrapper though

We used to control volumes and volumes differently, while a mount can always refer to its parent volume, so we can just handle both cases using a more abstract class to handle all the mountable volumes so that in case an unmounted volume gets mounted we can avoid creating a new app but just update the state of the old one. This will allow also to control more devices as the ones that are in a not-mounted state

For some implementations we're getting the volume set after the mount so let's ignore the volume in such case. LP: #1960898

Some devices may not be visible in the launcher even if we can technically list them, but it was not possible to mount them before because we were using gio open only and we needed a mount location. This is not the case anymore now, so we can just support showing all the devices, whether they're mounted or not.

Fixes micheleg#1338

It's just faster to wait the last event not being cancelled in an idle instead of doing them after the idle is ready.

If the user already requested to mount or umount a device, there's no point to permit calling this again till the operation is not completed. Notify about

So, for example still present the mount/unmount operations on devices that are being mounted/unmounted but do not leave the user activate them

…-manager When a new filemanager window is opened, we may assign it to nautilus while we're still receiving its location infos, causing it to be temporarily associated to the file manager app, instead to the actual location app. Causing some blinking, so let's use some quick timeout to wait the dbus events to come before proceeding with the assignments.

An application can take some time to start (especially when performing busy operations such as mounting a remote device), so in this case it's better to mark the application as starting and only when we've windows set as running. This can be now easily implemented for the case

Sadly nautilus doesn't provide an API to open new windows when a location is already opened, so using `gio open` can lead to focusing an already opened location. As per this, manually handle opening new windows if the command-line supports --new-window parameter.

Because `g_file_monitor_set_rate_limit` doesn't work. It's a stub in glib! Fixes: https://launchpad.net/bugs/1962699

We were already filtering them out, but this follows upstream change: https://gitlab.gnome.org/GNOME/gnome-shell/-/commit/c02ca54943

We use the same upstream wording here, to avoid having to change it in multiple places, so it has to be dynamic. Depending on the version we're running on. Fixes micheleg#1697

Apparently gjs was wrong in generating the camelCase name for variables in Turkish (at least) as there 'i'.toLocaleUpperCase() is `İ`. So let's workaround this, meanwhile gjs is properly fixed everywhere: https://gitlab.gnome.org/GNOME/gjs/-/merge_requests/742 Fixes micheleg#1687

3v1n0 · 2022-04-29T14:47:53Z

@rastersoft I don't see your comment here, however the mentions of this._monitorIndex present are present only on functions/vfuncs overrides.

So they refer to the overridden type.

rastersoft · 2022-04-29T14:54:54Z

@3v1n0 Yes, I discovered it after posting the message, that's why I deleted it. Sorry for the noise.

3v1n0 requested a review from ewlsh January 24, 2022 05:43

3v1n0 force-pushed the locations-refactory branch from 0a39266 to ca44729 Compare January 24, 2022 05:44

This was linked to issues Jan 24, 2022

dash-to-dock is not showing mounted network volumes and devices #1338

Closed

Trash Icon not reporting content #1375

Closed

Use custom icons for mounted devices #1508

Closed

3v1n0 changed the title ~~Locations refactory to use native Gio.AppIcon's~~ Locations refactory to use native Gio.AppIcon's (for GNOME 42 support) Jan 24, 2022

3v1n0 force-pushed the locations-refactory branch 2 times, most recently from e786dfc to 40d0858 Compare January 24, 2022 16:32

This was referenced Feb 4, 2022

Add 42 since it works there #1666

Closed

GNOME 41 round of fixes #1655

Merged

3v1n0 force-pushed the locations-refactory branch from 40d0858 to 5d26575 Compare February 23, 2022 16:57

weimzh mentioned this pull request Mar 4, 2022

Issues under GNOME 42 beta. #1679

Closed

emansom mentioned this pull request Mar 25, 2022

[DNM]metadata.json: Add 42 to shell version list #1686

Closed

3v1n0 mentioned this pull request Mar 30, 2022

Error "Main.panel._leftCorner" on Gnome 42 and gjs 1.72.0 (Debian sid) #1691

Closed

3v1n0 linked an issue Mar 30, 2022 that may be closed by this pull request

Error "Main.panel._leftCorner" on Gnome 42 and gjs 1.72.0 (Debian sid) #1691

Closed

3v1n0 force-pushed the locations-refactory branch from 5d26575 to 1385d9e Compare April 11, 2022 13:26

3v1n0 linked an issue Apr 11, 2022 that may be closed by this pull request

"Main.panel._rightCorner is undefined" on Gnome 42 #1702

Closed

3v1n0 mentioned this pull request Apr 11, 2022

Show mounted volumes and devices crashes GNOME #1701

Closed

3v1n0 linked an issue Apr 11, 2022 that may be closed by this pull request

Show mounted volumes and devices crashes GNOME #1701

Closed

3v1n0 mentioned this pull request Apr 11, 2022

Extension not working when using Turkish locale #1687

Closed

3v1n0 and others added 16 commits April 28, 2022 17:36

locations: Monitor windows-changed from apps and not from fm1Client

4173bb9

Windows changes may not be processed at the same time we got a fm1Client change now, so we should monitor apps signals instead of rely on what happens on the backend

locations: Ignore native file manager windows-changed signal emissions

d56bef7

We don't care about exposing windows-changed signal if the file manager app is wrapped to isolate devices. As we're handling signaling manually anyways

locations: Try to get trash status from metadata info before iterating

7074fc7

Try to just read the trash item count file info before enumerating the trash. It may be just cheaper and will work anyways

locations: Move handling of trash logic into TrashAppInfo

1b5b9a6

Since we've a native trash app info now we can handle the logic inside that. We need to handle data destruction from parent app wrapper though

locations: Do not try to add null volumes from valid mounts

f1e45bc

For some implementations we're getting the volume set after the mount so let's ignore the volume in such case. LP: #1960898

locations: Add support for showing network volumes

9321e24

Fixes micheleg#1338

locations: Use ChildCancellable to avoid queuing multiple trash updates

1b7039f

It's just faster to wait the last event not being cancelled in an idle instead of doing them after the idle is ready.

locations: Do not try to perform (un)mount actions concurrently

666a55d

If the user already requested to mount or umount a device, there's no point to permit calling this again till the operation is not completed. Notify about

appIcons: Show menu item as insensitive if the app is busy

adb8f28

So, for example still present the mount/unmount operations on devices that are being mounted/unmounted but do not leave the user activate them

locations: Reintroduce manual rate limiting for trash monitoring

0f97f60

Because `g_file_monitor_set_rate_limit` doesn't work. It's a stub in glib! Fixes: https://launchpad.net/bugs/1962699

locations: Always ignore OR windows in windows-backed app implementation

ae11cb7

We were already filtering them out, but this follows upstream change: https://gitlab.gnome.org/GNOME/gnome-shell/-/commit/c02ca54943

3v1n0 force-pushed the locations-refactory branch from 6c032bc to ae11cb7 Compare April 28, 2022 15:38

appIcons/prefs: Follow upstream terminology for Favorite or Pinned apps

d378382

We use the same upstream wording here, to avoid having to change it in multiple places, so it has to be dynamic. Depending on the version we're running on. Fixes micheleg#1697

3v1n0 force-pushed the locations-refactory branch from 095905a to d378382 Compare April 28, 2022 15:59

3v1n0 mentioned this pull request Apr 28, 2022

Move "Favorites" menu labels to the dashtodock domain #1711

Closed

3v1n0 force-pushed the locations-refactory branch from f3d9d1a to a6447f2 Compare April 28, 2022 19:44

3v1n0 linked an issue Apr 28, 2022 that may be closed by this pull request

Extension not working when using Turkish locale #1687

Closed

Zorono approved these changes Apr 29, 2022

View reviewed changes

ashb approved these changes May 4, 2022

View reviewed changes

emansom mentioned this pull request May 13, 2022

Quick fix to allow use on Gnome 42 #1740

Closed

micheleg merged commit 51b0e21 into micheleg:master May 15, 2022

LoneFenris mentioned this pull request Jun 5, 2022

Gnome 42 - What is happening with Dash to Dock? #1730

Closed

Provide feedback

Saved searches

Use saved searches to filter your results more quickly

Uh oh!

Locations refactory to use native Gio.AppIcon's (for GNOME 42 support) #1656

Locations refactory to use native Gio.AppIcon's (for GNOME 42 support) #1656

Uh oh!

3v1n0 commented Jan 24, 2022 •

edited

Loading

Uh oh!

timocapa commented Mar 10, 2022

Uh oh!

stuarthayhurst commented Mar 17, 2022

Uh oh!

timocapa commented Mar 20, 2022

Uh oh!

Etaash-mathamsetty commented Mar 27, 2022

Uh oh!

schtobia commented Mar 31, 2022

Uh oh!

fancypantalons commented Mar 31, 2022

Uh oh!

rastersoft commented Apr 1, 2022

Uh oh!

phush0 commented Apr 7, 2022

Uh oh!

martindmtrv commented Apr 8, 2022

Uh oh!

OctoRocket commented Apr 11, 2022

Uh oh!

3v1n0 commented Apr 29, 2022

Uh oh!

rastersoft commented Apr 29, 2022

Uh oh!

Uh oh!

Locations refactory to use native Gio.AppIcon's (for GNOME 42 support) #1656

Locations refactory to use native Gio.AppIcon's (for GNOME 42 support) #1656

Uh oh!

Conversation

3v1n0 commented Jan 24, 2022 • edited Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

Uh oh!

timocapa commented Mar 10, 2022

Uh oh!

stuarthayhurst commented Mar 17, 2022

Uh oh!

timocapa commented Mar 20, 2022

Uh oh!

Etaash-mathamsetty commented Mar 27, 2022

Uh oh!

schtobia commented Mar 31, 2022

Uh oh!

fancypantalons commented Mar 31, 2022

Uh oh!

rastersoft commented Apr 1, 2022

Uh oh!

phush0 commented Apr 7, 2022

Uh oh!

martindmtrv commented Apr 8, 2022

Uh oh!

OctoRocket commented Apr 11, 2022

Uh oh!

3v1n0 commented Apr 29, 2022

Uh oh!

rastersoft commented Apr 29, 2022

Uh oh!

Uh oh!

3v1n0 commented Jan 24, 2022 •

edited

Loading