Still Failing to Start Sometimes on XP x86

[DaveH1]

There is still an intermittent startup failure happening on Windows XP x86 with 126.0.6478.260 R6,
The previous version was the same.

Starting up the browser, it sometimes apparently crashes and no window appears.
It doesn't happen every time, I would say about one time in four.

When the problem happens, when the browser is run, either nothing appears to happen, or it appears briefly in the taskbar and then vanishes. Immediately running it again will often then apparently run correctly.

If the browser window does not open, there are still several chrome.exe tasks running in the Task Manager.
Sometimes it won't just start when it's tried again, and in that case these processes have to be manually killed before the browser will start again, and in that case it then usually does start correctly

[win32ss]

Based on various reports from Windows XP users, some further internal changes will be needed, including the introduction of keyed events. This should improve stability.

[DaveH1]

Thanks Shane!

[MWF2]

Meanwhile at 64bit browser (64 bit of xp) same thing
41ea783d-ca5f-4e9b-87cb-379e58640393.dmp

[DaveH1]

Interesting to hear that the 64 bit version is doing it as well.
I should have mentioned that the browser is rock solid once it has started successfully, I've never had it crash while it's running.
It's a startup problem only.

[win32ss]

Here are test versions of 32 and 64 bit progwrp.dll.
progwrp_32.zip
progwrp_64.zip

They will write a debugging log file to indicate failure due to invalid parameters in the InitOnce API.

[DaveH1]

Thanks Shane, I will install the 32 bit version and let you know how it goes!
Do you need feedback every time that the browser fails to start properly, or should I let the log build up to capture a few instances before I submit it?
Cheers, Dave.

[win32ss]

The log will only be written to on failure, and only in one of the two possible ways of failure. The other one cannot be easily captured by a log but should not happen anyway because it involves a function that should always be successful.

[zap369]

Everything that DaveH1 wrote is exactly how supermium is behaving on my end but I'm still using 256.R5 on win xp sp3 32bit.

When there are still three chrome.exe tasks running even though supermium is seemingly shut down, one of those tasks is running at 50% CPU. Since I'm running on a dual core, when I forget to kill these tasks the computer does slow down. I might roll back to R4 and wait until this is fixed since DaveH1 reported it being an issue on 260.R6 too.

Before I rollback to R4 I'll test R6 and report back on here.

[zap369]

So I just installed 260.R6 and on the first boot it crashed leaving several chrome.exe tasks running one of which at 50% CPU. After killing those the browser booted up normally.

Instead of rolling back to R4 I'll continue testing R6 to see if this is the only bug on my end.

Thanks for all the work.

[DaveH1]

Well I've been using the debug version of progwrp.dll (1.1.0.5024) for a few days now, and I haven't had a single startup crash!
The log file has remained empty throughout.
Looks like the debug version might actually have fixed the problem, at least on my system!

[zap369]

Yeah, so far so good on my end too.