https://bugs.winehq.org/show_bug.cgi?id=54769
Bug ID: 54769 Summary: major regression in dnSpy startup time Product: Wine Version: 8.5 Hardware: x86-64 OS: Linux Status: UNCONFIRMED Severity: normal Priority: P2 Component: -unknown Assignee: wine-bugs@winehq.org Reporter: galtgendo@o2.pl Distribution: ---
(OK, this will be quite a bit different than what I've planned when I started to write this)
Let's start with an oddity: only 32bit versions of dnSpy work, 64bit fail to show anything besides the recently popular stack corruption line (it does not terminate upon it, but fails to do anything else either).
The problem (initially observed in 6.1.8, but dnSpyEx 6.3.0 has the same problem):
I'm not quite sure when this happened, but IIRC it was at most two releases back, maybe it was the latest one.
Before, upon startup there was a minor delay before the window showed up and every now and then (but relatively rarely) the window got filled with visual garbage until it was refreshed.
Now, it *seems* corruption has been dealt with (though that needs more time to be verified), but there's a crippling startup delay (over 2-3 min) between when a blank window shows up and when it's filled with content.
https://bugs.winehq.org/show_bug.cgi?id=54769
Rafał Mużyło galtgendo@o2.pl changed:
What |Removed |Added ---------------------------------------------------------------------------- Keywords| |regression
https://bugs.winehq.org/show_bug.cgi?id=54769
--- Comment #1 from Rafał Mużyło galtgendo@o2.pl --- ...well, it seems that what had before manifested as visual garbage, now is simply blanking the window (that is turning it fully black) ...
Not sure if that counts as progress...
https://bugs.winehq.org/show_bug.cgi?id=54769
--- Comment #2 from Austin English austinenglish@gmail.com --- Please run a regression test: https://wiki.winehq.org/Regression_Testing
Given that there seem to be multiple issues, I'd suggest focusing on bisecting which commit makes it take so much longer to start.
That commit may or may not be the same commit that caused the black screen. If it's not, a second bisect/bug report will be needed.
https://bugs.winehq.org/show_bug.cgi?id=54769
--- Comment #3 from Rafał Mużyło galtgendo@o2.pl --- ...
OK, it seems I was wrong about what the problem actually was, but things are quite quirky.
My window manager is set to "focus follows mouse" policy with a 0.8s delay.
If I move the mouse in and out of the window a couple times (to a different window), slowly enough for the focus change to take place, after a random number of tries (but relatively low) the window eventually gets refreshed. Alt-tabbing doesn't *seem* to make a difference though.
So, it's not so much a startup delay, but some very odd race which conditions recently changed...
https://bugs.winehq.org/show_bug.cgi?id=54769
Zeb Figura z.figura12@gmail.com changed:
What |Removed |Added ---------------------------------------------------------------------------- Status|UNCONFIRMED |NEEDINFO CC| |z.figura12@gmail.com Ever confirmed|0 |1
--- Comment #4 from Zeb Figura z.figura12@gmail.com --- This bug report is a mess, missing a lot of information and reporting multiple unrelated problems at once, but regardless, I can't reproduce any of the mentioned problems with dnSpy 6.1.8 and upstream Wine (9.0-rc1), with dotnet48 and d3dcompiler_47 installed via winetricks. Running both dnSpy.exe and dnSpy-x86.exe works fine for me.
whatsit@camazotz:~/Downloads$ sha1sum dnSpy-netframework.zip 439f401712125220b05ba544d97f80f6dec43628 dnSpy-netframework.zip
If you can reproduce any of these problems with upstream Wine, please describe in detail how you set up the prefix, how you are running dnSpy, which binary you downloaded, and anything else unusual about your setup, and please also perform a bisect as was requested.
https://bugs.winehq.org/show_bug.cgi?id=54769
--- Comment #5 from Rafał Mużyło galtgendo@o2.pl --- See https://github.com/dnSpyEx/dnSpy - a fork, that's still being updated; original has ceased being a few years ago.
Also, your wine-mono installation must be weird (augmented by upstream dotnet ?) for the framework version to not at least complain.
That delay seems indeed fixed in 8.21 - well, maybe earlier, I didn't need to use dnSpy lately.
As for 64bit, that problem had been confirmed (either in a different bug here or in the chat) as a known problem with 64bit implementation of some function in wine-mono. (after checking the logs, it was bug 52396 - supposedly fixed now).
Yet, now it seemingly hangs after showing the window in the decompiling stage...and it happens in 32bit and 64bit both. It eventually gets through, but it takes a bit over a minute. So, even though old problem seems fixed, there's still more to do there. It might still be a regression, if a different one.
https://bugs.winehq.org/show_bug.cgi?id=54769
Zeb Figura z.figura12@gmail.com changed:
What |Removed |Added ---------------------------------------------------------------------------- Status|NEEDINFO |RESOLVED Resolution|--- |FIXED
--- Comment #6 from Zeb Figura z.figura12@gmail.com --- (In reply to Rafał Mużyło from comment #5)
See https://github.com/dnSpyEx/dnSpy - a fork, that's still being updated; original has ceased being a few years ago.
You mentioned in comment 0 that the bug was reproducible with 6.1.8.
Also, your wine-mono installation must be weird (augmented by upstream dotnet ?) for the framework version to not at least complain.
I mentioned in comment 4 that I installed dotnet48 and d3dcompiler_47 through winetricks.
That delay seems indeed fixed in 8.21 - well, maybe earlier, I didn't need to use dnSpy lately.
Okay, resolving fixed then.
https://bugs.winehq.org/show_bug.cgi?id=54769
Alexandre Julliard julliard@winehq.org changed:
What |Removed |Added ---------------------------------------------------------------------------- Status|RESOLVED |CLOSED
--- Comment #7 from Alexandre Julliard julliard@winehq.org --- Closing bugs fixed in 9.0-rc2.
-
WineHQ Bugzilla