Where communities thrive


  • Join over 1.5M+ people
  • Join over 100K+ communities
  • Free without limits
  • Create your own community
People
Repo info
Activity
  • Aug 16 11:46
    siquylee opened #2281
  • Aug 15 21:34
    jgcodes2020 opened #2280
  • Aug 14 18:38
    PrestigeDevop edited #2279
  • Aug 14 18:34
    PrestigeDevop opened #2279
  • Aug 10 07:41
    Serg-Norseman synchronize #2243
  • Aug 07 12:26
    allsorts46 commented #2277
  • Aug 07 12:18
    allsorts46 synchronize #2277
  • Aug 07 10:31
    Serg-Norseman commented #2243
  • Aug 07 10:15
    Miepee commented #2255
  • Aug 07 09:08
    tewkanz commented #2255
  • Aug 07 03:10
    allsorts46 opened #2277
  • Aug 07 01:39

    cwensley on 2.7.1

    (compare)

  • Aug 07 01:39

    cwensley on 2.7.1

    (compare)

  • Aug 07 01:37

    cwensley on develop

    Fix publish wildcard (compare)

  • Aug 06 22:34

    github-actions[bot] on 2.7.1

    (compare)

  • Aug 06 21:46

    cwensley on develop

    Add macOS platform templates Merge pull request #2276 from c… (compare)

  • Aug 06 21:46
    cwensley closed #2276
  • Aug 06 21:40
    cwensley commented #2243
  • Aug 06 21:33
    cwensley opened #2276
  • Aug 06 21:33
    cwensley labeled #2276
LuciferSam86
@LuciferSam86
Hi, just a question. Is the addin required for building Eto.Forms apps compatible with VS 2022? Thanks!
Curtis Wensley
@cwensley
Hey @LuciferSam86, no, you do not need the addin. It adds templates, xaml autocomplete, and a previewer
Curtis Wensley
@cwensley
Otherwise, Eto can be consumed as nuget packages, including the templates using the dotnet command line.
Miepee
@miepee:matrix.org
[m]
Is it possible to make some runtime dependencies on linux (e.g webkit if you use it, libappindicator etc.) Optional and to not cause a runtime crash?
It would make a few things more convenient, e.g newer Debian which changed the appindicator file would still be able to run the application, just not with all bells and whistles.
Miepee
@miepee:matrix.org
[m]
:point_up: Edit: Is it possible to make some runtime dependencies on linux/gtk (e.g webkit if you use it, libappindicator etc.) Optional and to not cause a runtime crash?
It would make a few things more convenient, e.g newer Debian which changed the appindicator file would still be able to run the application, just not with all bells and whistles.
Curtis Wensley
@cwensley
It is possible, we just need to detect which one is available and use the correct version, vs. just switching out for another lib that isn't ubiquitous.
Miepee
@miepee:matrix.org
[m]
well i meant more in the scenario that nothing is available
Curtis Wensley
@cwensley
The only way not to crash in that scenario is just don't use it. E.g. there's no way the WebView will work without the required depenencies
We could add some detection so it doesn't add the WebViewHandler if the dependencies don't exist, so you could do something like if (Platform.Supports<WebView>() { /* now it's safe to use WebView */ }
Miepee
@miepee:matrix.org
[m]
That would be neat, yeah.
Miepee
@miepee:matrix.org
[m]
I just encountered that while working on some flatpak related stuff.
Did you already do any work in that department? Would like to make a possible template public for Eto programs.
Curtis Wensley
@cwensley
Not sure what you're asking entirely, the WebViewHandler is always added so the Platform.Supports<> method will always return true. We just need to do a check first before adding the WebViewHandler in Eto.Gtk/Platform.cs
Same for the LinuxTrayIndicatorHandler
Miepee
@miepee:matrix.org
[m]
Sorry my bad, I completely misphrased what I meant <-<
The second line was regarding whether any flatpak related stuff for Eto was done yet. In case you're not familar with them, they're a distribution mechanism for linux programs.
Curtis Wensley
@cwensley
Yes I'm aware of flatpaks, I'm not quite a fan of them due to the added complexity of installing flatpak related dependencies.
If they were trying to make sandboxing very complicated, they certainly solved that problem.
Miepee
@miepee:matrix.org
[m]
That's because Linux as a whole is complicated :)
They wanted to create a sandboxing environment which supports:
  • GNU stuff (i.e. glibc and their coreutils)
  • non-GNU stuff (i.e. musl and busybox or other stuff)
  • All kinds of init systems (i.e., systemD, runnit, openRC etc.)
  • Different display protocols (i.e. X11, Wayland and any potential future ones)
  • Needs proper definitions in which programs and parts of the filesystem it needs to be able to access and which not
  • Needs to properly be able to invoke filepickers and such, while being sandboxed (as usually those lie outside of the sandbox)
  • Needs to properly communicate with the different DEs (to know which filepicker it needs to invoke)
    And probably a whole bunch of more stuff.
Curtis Wensley
@cwensley
That's fair. At any rate, I'd welcome any changes to help support running in flatpak if necessary.
Miepee
@miepee:matrix.org
[m]
Oh, something that could help a bit more for that, are there any public binaries for the test suite, or do I need to self-compile?
My current test application runs fine as of right now, but it'd be a good idea to have that tested against something something that uses all features of Eto, rather than only a few
Curtis Wensley
@cwensley
You can get all the test app binaries from the actions, such as this one for 2.7.0: https://github.com/picoe/Eto/actions/runs/2285279754#artifacts
Miepee
@miepee:matrix.org
[m]
Sweet, thanks!
Curtis Wensley
@cwensley
Any time!
Miepee
@miepee:matrix.org
[m]

I got a bit distracted by the testing application; it seems that a few things are broken on either GTK+Linux or Wayland.
Stuff that doesn't work (correctly), some exceptions the program catches, and some exceptions that crash the program outright.

Should I test those later on a more recent commit and submit issues for them?

As examples; the icon doesn't get displayed in either title or taskbar, the datepicker instantly closes, the "bitmap" thing produces an exception, a lot of stuff that uses the GTK/GNOME dialogs instead of current DE's, and something else related to child windows that straight up crashed the program.
(Also all non-flatpak related, this is all natively)
Curtis Wensley
@cwensley
Icons on the taskbar need to be handled by using a config file. Icons not showing on the title bar are probably a theme issue.
Miepee
@miepee:matrix.org
[m]
I'm on kde, not GNOME
Kde always showed them fine
Curtis Wensley
@cwensley
Either way, it's not provided by the program
afaik
Miepee
@miepee:matrix.org
[m]
To be specific it shows a program icon in both locations; which is the default Wayland icon.
Could investigate more tomorrow what exactly causes those kinds of bugs and then raise issues for them.
Datepicker not working for example is somewhat bad
Andrin Bertschi
@abertschi
Hi there. I have a GridView inside a Scrollable container. when the GridView is focused and I scroll, I want the parent to scroll and not the GridView itself. Is there a way to fix this? (Platform Windows).
philstopford
@philstopford
@cwensley : I don't seem to be able to update to 2.7.1 because there's a conflict with the recommended upgrade to the System.Security.Cryptography version from 6.0 to 6.0.1. Is that known as an issue?
Curtis Wensley
@cwensley
Hm, not that I'm aware of..
Eto doesn't reference that nuget package (directly) so I'm not sure where that would come from..
philstopford
@philstopford
As soon as I try to upgrade Eto nugets, it blows up citing that apparent dependency
Curtis Wensley
@cwensley
Which platform nuget are you using?
philstopford
@philstopford
WPF, GTK, Mac
Curtis Wensley
@cwensley
There were no nuget package changes from 2.7.0 to 2.7.1 fwiw.
philstopford
@philstopford
Mac64
I'm still poking it to see if I can find a sequence of operations that might side-step it. It's very odd.
Curtis Wensley
@cwensley
There was only one nuget package change in 2.7.0 that changed Windows7APICodePack-Shell to Microsoft-WindowsAPICodePack-Shell in WPF and WinForms.
What version are you upgrading from? 2.7.0?
philstopford
@philstopford
2.7.0, yes
Curtis Wensley
@cwensley
Hm, ok well there's no differences then.
afaik.
philstopford
@philstopford
I'll assume there's some unique breakage locally and continue to bang my head on the desk trying to sort it out :D
Curtis Wensley
@cwensley
Ok, good luck! cleaning out all your obj files might help..