The crash reporter didn't work on my computer because of a missing URL
or something.
But as far as crashes go, I had heaps of them one day in particular and
then they stopped. On 3.4.9
On 18/07/19 6:19 PM, Nyall Dawson wrote:
On Thu, 18 Jul 2019 at 16:05, Andreas Neumann wrote:
Hi QGIS
Hi Juergen,
On 22/07/19 10:51, Jürgen E. Fischer wrote:
> Then they've never seen the crash dialog - which would be good. Otherwise it
> has the instructions:
>
> https://github.com/qgis/QGIS/blob/master/src/crashhandler/qgscrashdialog.cpp#L35
probably clear enough, thanks
cheers
--
Paolo
Hi,
On Mon, 22. Jul 2019 at 09:34:36 +0200, Paolo Cavallini wrote:
> On 20/07/19 10:13, Jürgen E. Fischer wrote:
> > I've implemented 5ed9efb3b60 to make use of this in master - the release
> > branches will follow.
> do we have simple instructions on how to get and send a valid log?
> I
Hi all,
On 20/07/19 10:13, Jürgen E. Fischer wrote:
> I've implemented 5ed9efb3b60 to make use of this in master - the release
> branches will follow.
do we have simple instructions on how to get and send a valid log?
I don't think this is widely known, even to interested users.
Cheers.
--
Hi Nathan,
On Thu, 18. Jul 2019 at 20:18:15 +1000, Nathan Woodrow wrote:
> Yeah I know it was a bit much to have them included. I will see what I can
> do with my install and hopefully get a good stack trace.
Oh well, several issues with the pdbs.
The setup.hints were missing from the release
Calvin,
A fair amount of crashers and stability issues have been resolved in 3.8.1,
which'll be out in a few days. Give that a try.
If you still find it unstable, actionable information describing crashes
and issues are needed (since you're a plugin dev, I'm sure that's familiar
to you ;) ). If
On Fri, 19 Jul 2019 at 09:26, Nyall Dawson wrote:
>
> On Thu, 18 Jul 2019 at 16:52, Nathan Woodrow wrote:
> >
> > Yeah, in this case, the stack trace is wrong because of the lack of
> > symbols. The symbols could be installed with QGIS at the increase of
> > package size but maybe we can do
On Fri, Jul 19, 2019 at 09:26:08AM +1000, Nyall Dawson wrote:
> On Thu, 18 Jul 2019 at 16:52, Nathan Woodrow wrote:
> >
> > Yeah, in this case, the stack trace is wrong because of the lack of
> > symbols. The symbols could be installed with QGIS at the increase of
> > package size but maybe we
Hmm pretty sure if is wrong. Without debug symbols it doesn't know what
code is running and just has a good guess. A stack trace with symbols will
show a completely different stack
On Fri., 19 Jul. 2019, 9:26 am Nyall Dawson, wrote:
> On Thu, 18 Jul 2019 at 16:52, Nathan Woodrow wrote:
> >
> >
On Thu, 18 Jul 2019 at 16:52, Nathan Woodrow wrote:
>
> Yeah, in this case, the stack trace is wrong because of the lack of symbols.
> The symbols could be installed with QGIS at the increase of package size but
> maybe we can do that for a while until we catch this issue.
I don't think the
QGIS 3.8 crashes most times for me on exiting and I have also found it to
be fairly unstable. I have had it corrupt one of the saved files so that
when it tried to load it, it would just hang. The georeferencing almost
always crashes or delivers strange results. I have not reported on it
because I
Thanks Jürgen,
Yeah I know it was a bit much to have them included. I will see what I can
do with my install and hopefully get a good stack trace.
Regards,
Nathan
On Thu, Jul 18, 2019 at 5:07 PM Jürgen E. Fischer wrote:
> Hi Nathan,
>
> On Thu, 18. Jul 2019 at 16:51:44 +1000, Nathan Woodrow
Hi,
Unfortunately I am of no good use currently to test this, as I just
recently switched jobs and currently only have access to Linux as
sysadmin. In Windows I currently don't have admin rights. And I don't
own a private copy of Windows. I am working on getting admin rights on
Windows as well,
Hi Nathan,
On Thu, 18. Jul 2019 at 16:51:44 +1000, Nathan Woodrow wrote:
> Jurgen, would we be able to include the debug symbols for the build on the
> normal installers for once release cycle while this is still an issue? That
> would be super handy to catch this and other issues that come up.
In the meantime, if we have fellow Andreas able to reproduce an exit crash,
let's help him install the debug package and get a valid trace?
Andreas, that is, if you're game.
Math
On Thu, Jul 18, 2019, 13:51 Nathan Woodrow wrote:
> Yeah, in this case, the stack trace is wrong because of the
Yeah, in this case, the stack trace is wrong because of the lack of
symbols. The symbols could be installed with QGIS at the increase of
package size but maybe we can do that for a while until we catch this issue.
At the moment installing the symbols for the normal installer versions of
QGIS is
While discussing this further with Nathan, disabling the crash reporter
might not be the right solution as it'd most likely bring the OS crash
dialog anyways.
We'd need for the crash reporter to detect specific traces and silently
fail when it detects a useless / harmless trace.
There's also the
Hi Nyall,
Thanks for the reply. I also experienced quite a lot of these crashes.
Something that often shows up in the stack trace is
proj_lpz_dist :
proj_lpz_dist :
QgsCoordinateTransform::transformPolygon :
QgsCoordinateTransform::transformPolygon :
On Thu, 18 Jul 2019 at 16:05, Andreas Neumann wrote:
>
> Hi QGIS core devs,
>
> Currently on the Github issues I have the (subjective) feeling that every
> third report is about QGIS crashing on closing the program. It seems to be a
> really wide-spread issue that a majority of our users have.
Hi QGIS core devs,
Currently on the Github issues I have the (subjective) feeling that
every third report is about QGIS crashing on closing the program. It
seems to be a really wide-spread issue that a majority of our users
have. It would be great for QGIS' reputation if we could fix this/these
20 matches
Mail list logo
