[prev in list] [next in list] [prev in thread] [next in thread] 

List:       kde-core-devel
Subject:    Re: Is there (going to be) an auto-retracer service for KDE?
From:       Harald Sitter <sitter () kde ! org>
Date:       2021-04-29 11:25:30
Message-ID: CAEc+18GLZwirGDkpVkhWV3rXvJShE4FyCOrMYdBk1-k6qbatyA () mail ! gmail ! com
[Download RAW message or body]

On Tue, Apr 27, 2021 at 2:16 PM Halla Rempt <boud@valdyas.org> wrote:
> 
> On Monday, 26 April 2021 13:07:17 CEST Harald Sitter wrote:
> > Also going off on a tangent: On windows I understand the store already
> > has crash tracking and all that stuff implemented, I expect the same
> > is true for OSX?
> 
> Yes, but without proper debug symbols the traces are not all that useful -- though \
> still good enough to tell me that display drivers on Windows are the main source of \
> crashes.

Our apps that I have access to all seem to have symbols uploaded. That
being said the "stack trace" isn't too hot in general:

> Frame    Image    Function    Offset
> 0    Qt5Core.dll    qt_message_fatal    0x0000000000000018
> 1    Qt5Core.dll    QMessageLogger::fatal    0x0000000000000071
> 2    Qt5Gui.dll    init_platform    0x0000000000000861

Which, I mean, is a call trace, but that's all it is. Fortunately one
can actually get a hold of the minidump with a huge caveat:

https://docs.microsoft.com/en-us/windows/uwp/publish/health-report

> CAB files will only be available when the failure occurred on a computer using a \
> Windows Insider build, so not all failures will include the CAB download option.

With that in mind I guess the answer is that crash reports from
systems that aren't in insider mode aren't terribly useful in general
outside the general sense that something may be crashing a whole lot
and which functions are involved.
Crash reports from insiders OTOH have this zip file available which
contains the WER [1] metadata and the minidump which one can load into
VS [2] or windbg [3]. Combined with auto-retrieval [4] that surely
should give all the building blocks to generate rich(er) traces.

This really doesn't seem quite as useless as you make it sound. It's
still a bit meh to need insider users to get viable data though. What
might be interesting there is if an app could opt into WER local
dumping and then have a dialog on next run "yo, krita crashed last
time, how about you send us the crash" which would then behave like
what I described for desktop linux (upload minidump to us, we trace
server-side) thereby bypassing microsoft entirely. Albeit this has all
the same challenges as desktop linux too then. Someone needed to write
the weby-techy bits of receiving dumps and then tracing them and
manage a crash record database.

[1] https://docs.microsoft.com/en-us/windows/win32/wer/windows-error-reporting
[2] https://docs.microsoft.com/en-us/visualstudio/debugger/using-dump-files
[3] https://docs.microsoft.com/en-us/windows-hardware/drivers/debugger/debugger-download-tools
 [4] https://docs.microsoft.com/en-us/windows/uwp/monetize/access-analytics-data-using-windows-store-services


HS


[prev in list] [next in list] [prev in thread] [next in thread]