Closed Silent-Hunter closed 8 years ago
Hi, a little more info about those crashes? Bug reports like that can make me aware of the fact that there is an issue, but they will help me very little actually fixing it.
I need core dumps, distribution, pidgin logs, basically any information about your environment and the crashes I can get.
Absolutely and thank you for replying to this so quickly! I was a little frustrated earlier, but I'm more calm and able to troubleshoot now.
What command should I use to get this information? I have the debug symbols built into Pidgin, so it prints a LOT of information to the console. Does Pidgin save a log in /tmp or somewhere? Or shall I use gdb?
pidgin -d will start in debug mode, the log would be useful. Otherwise start it with gdb and use it until it crashes. If it crashes, enter the command "bt" in gdb
Alright! I will not have time for this tonight, because it can take a while before it does it, so I will do that tomorrow. Thank you for your help, and thanks for the plugin, until now it's been lovely!
On Sun, Nov 22, 2015 at 5:15 PM, mj notifications@github.com wrote:
pidgin -d will start in debug mode, the log would be useful. Otherwise start it with gdb and use it until it crashes.
— Reply to this email directly or view it on GitHub https://github.com/majn/telegram-purple/issues/162#issuecomment-158808094 .
OK, I've been using it for over six hours with no problems, so I don't know what to do. Either starting it with -d fixed it somehow, or it was a fluke. I'll get back to you if it ever happens again.
It crashed! Where do I find the log file that starting Pidgin with -d created?
In your terminal, when you execute pidgin
Btw, depending on your linux distribution there may be other places where crash reports are stored. If you use something based on Gnome3 the tool gnome-abrt shows you the last crashes including stack trace.
Oh crap, well the terminal didn't log far enough, I thought it would store it in a file. Can I use the usual funnel method to get it into a file?
Just do the following:
gdb pidgin
When gdb is loaded start the application with:
run
Once it crashes you can get the backtrace with, copy/paste the output here.
bt
@majn in future, might be easier to point to https://developer.pidgin.im/wiki/GetABacktrace which has instructions about the debug logging and backtracing
Yeah it didn't crash all day. I would close this, I'm not sure what happened but it's stable now, or at least seems so. Can running something with gdb change its behaviour?
Well, it crashed, does this tell you anything?
---Type
---Type
---Type
Sorry about all the trouble.
It seems to often happen on closing a tab actually.
#5 0x00007ffff5d5e060 in purple_imgstore_unref () from /usr/lib64/libpurple.so.0
together with "crash on closing a tab" sounds a lot like there's a double-free (or "not allocated using g_malloc or whatever glib expects") of an image. As far as I remember, tgp knows three types of images:
Your telegram-purple version would be nice to know
How do I check the version? I just checked out the repo and built it. On Nov 27, 2015 5:30 AM, "mj" notifications@github.com wrote:
Your telegram-purple version would be nice to know
— Reply to this email directly or view it on GitHub https://github.com/majn/telegram-purple/issues/162#issuecomment-160107811 .
Execute git log HEAD^..
and tell us the result. That's the most detailed version information one can obtain.
commit 6cfeeb4835481e544a05a81f253dec8e6f63ca84 Merge: 09339c7 2e3951f Author: mj mtthsjntsch@gmail.com Date: Sat Nov 21 09:27:22 2015 +0100
Merge pull request #159 from tuxmaster/master
Fix rpm spec file.
commit 2e3951f5a5a6fc3050923c6a019aa22e12f7daed Author: tuxmaster frank@familie-büttner.de Date: Wed Nov 18 15:05:10 2015 +0100
Add missing space.
commit f7b01583fc34c8125d52d6785a5cf891e9f7202d Author: tuxmaster frank@familie-büttner.de Date: Wed Nov 18 15:03:40 2015 +0100
Add required version for libgcrypt.
commit 2f5c70b04bed231fdce035464bef511d38500712 Author: tuxmaster frank@familie-büttner.de Date: Tue Nov 17 18:53:09 2015 +0100
-Fix date for rpm spec file.
There ya go! I hope it helps.
On Fri, Nov 27, 2015 at 8:05 AM, Ben Wiederhake notifications@github.com wrote:
Execute git log HEAD^.. and tell us the result. That's the most detailed version information one can obtain.
— Reply to this email directly or view it on GitHub https://github.com/majn/telegram-purple/issues/162#issuecomment-160135800 .
Then you have a new version (not the newest, but new enough) from master.
This bug is bad, and we're interested in resolving it. (Just to reassure you :P)
Do you recall what you were sending each other? Stickers, or other images?
Well, no stickers, there was one the other day but there was no crash that day. Images yes, but the crashes are unrelated to those, seemingly. It seems like it's related to using Ctrl+W to close a tab.
On Fri, Nov 27, 2015 at 10:02 AM, Ben Wiederhake notifications@github.com wrote:
Then you have a new version (not the newest, but new enough) from master.
This bug is bad, and we're interested in resolving it. (Just to reassure you :P)
Do you recall what you were sending each other? Stickers, or other images?
— Reply to this email directly or view it on GitHub https://github.com/majn/telegram-purple/issues/162#issuecomment-160155723 .
@Silent-Hunter: Sorry for my lack of response. If you have sent images, then the theory "tried to double-free an image-ref" makes sense.
Does this still occur in 1.2.5?
It doesn't happen anymore, no. Thanks for fixing that!
On Sat, Feb 20, 2016 at 10:17 AM Ben Wiederhake notifications@github.com wrote:
@Silent-Hunter https://github.com/Silent-Hunter: Sorry for my lack of response. If you have sent images, then the theory "tried to double-free an image-ref" makes sense.
Does this still occur in 1.2.5?
— Reply to this email directly or view it on GitHub https://github.com/majn/telegram-purple/issues/162#issuecomment-186623271 .
I was having problems with random crashes, so I just updated less than an hour ago, and now it also disconnects randomly, as well as not receiving messages and still crashing. I'm not sure what exactly is happening, but since no one else has reported this, can I assume it's only happening to me?