drkonqi and the windows build

[Harald Sitter]

On Thu, Sep 17, 2020 at 4:17 PM Hannah von Reth <vonreth at kde.org> wrote:
>
> Hi,
>
> looks like some quoted text is missing.
>
> I have no idea bout the state of drkonqi on windows,
> ages ago a gsoc student did *things* but I don't think anybody was ever
> able to use it.
>
> I can remember a more recent communication that people where trying
> again to get it to work.

Ah sorry.

The original:
> drkonqi is part of plasma, yet it gets built on windows. shall we stop
> doing that? windows keeps not having expected plasma requirements
> because the rest of plasma is not on windows
>
> https://build.kde.org/job/Plasma/job/drkonqi/job/kf5-qt5%20WindowsMSVCQt5.14/
>
> this already happened a couple weeks ago and was then temporarily
> reverted but now it fails again. uselessly I might add. why we even
> want to have drkonqi on windows at all is equally questionable as the
> platform already has a far superior crash reporting system.

I guess the question is mostly.... do we think that pursuing drkonqi
on windows is useful at all?

The way I see it we'd either want to use the windows error reporting
system (to send reports off to microsoft and then get them from there)
or we'd need a fair amount of extra code in drkonqi to tie into
local-only WER (I'm assuming that's a thing based on the WER API).
Since we distribute the binaries we only need a mini dump and the
exact build version to then grab our debug symbols for that build and
do server-side tracing, so the actual live-tracing feature of drkonqi
is fairly uninteresting IMHO.

HS