Bug#753135: Bug#750557: Bug#754114: Bug#750557: pgadmin3: segmentation fault
Le 27.10.2014 12:58, Christoph Berg a écrit : Version: 1.20.0~beta2-1 Re: Damyan Ivanov 2014-10-27 20141027112735.gb18...@ktnx.net -=| Bogdan Vatra, 27.10.2014 12:16:16 +0200 |=- On Monday 27 October 2014 10:53:01 Christoph Berg wrote: Re: To bmorel 2014-09-11 20140911093144.ga7...@msg.df7cb.de there's now 1.20.0~beta2-1 in unstable. Could you give that a try again? I don't see the crash in beta2 here. Also, would you say this bug makes pgadmin3 unusable, or is the rest of the functionality working just fine? I'm pondering downgrading the severity of this bug to just important so we at least have a mostly-working version of pgadmin3 in jessie. It seems the crash if fixed, I can't reproduce it anymore. Same here. No crashes. Ok, that's even better, thanks for confirming! I'll close the bug. Sorry, I successfully reproduced the bug, or at least, a sigsegv. I am using a stable version of Debian, but since the upgrade needed a more recent version of libjpeg and libjpeg-turbo, I have installed them (from debian package). I also have updated pgadmin3-data, of course. Right-cliking on an entry in the menu still makes pgadmin having a segfault. Started from terminal: $ pgadmin3 15:33:00: Debug: Failed to connect to session manager: SESSION_MANAGER environment variable not defined 15:33:01: Debug: Adding duplicate image handler for 'PNG file' Erreur de segmentation The 2 first messages appears at start, the Erreur de segmentation (which means segmentation fault) comes when I right click on an entry (even when I'm not connected) after an event made the context menu to stop appearing (tried to click somewhere else and to use ESC key, same results). I notice a small glitch that may be related. After opening the database maintenance dialog, hovering over any of the checkboxes or the first radio button (VACUUM) changes the mouse pointer to a nw-se resizing arrow, which is not the standard resizing arrow for the graphical environment (GNOME) but something else. Not sure about the bug's severity. I use unstable, so I am not even sure the bug still applies to the version in jessie. There was an update of wx3.0 some weeks ago that may as well sbe fixing this bug. Thanks for bringing beta2 to unstable, ... and now hopefully also to jessie :) Christoph -- To UNSUBSCRIBE, email to debian-bugs-dist-requ...@lists.debian.org with a subject of unsubscribe. Trouble? Contact listmas...@lists.debian.org
Bug#754114: Bug#750557: Bug#754114: Bug#750557: pgadmin3: segmentation fault
Le 27.10.2014 15:43, berenger.mo...@neutralite.org a écrit : Le 27.10.2014 12:58, Christoph Berg a écrit : Version: 1.20.0~beta2-1 Re: Damyan Ivanov 2014-10-27 20141027112735.gb18...@ktnx.net -=| Bogdan Vatra, 27.10.2014 12:16:16 +0200 |=- On Monday 27 October 2014 10:53:01 Christoph Berg wrote: Re: To bmorel 2014-09-11 20140911093144.ga7...@msg.df7cb.de there's now 1.20.0~beta2-1 in unstable. Could you give that a try again? I don't see the crash in beta2 here. Also, would you say this bug makes pgadmin3 unusable, or is the rest of the functionality working just fine? I'm pondering downgrading the severity of this bug to just important so we at least have a mostly-working version of pgadmin3 in jessie. It seems the crash if fixed, I can't reproduce it anymore. Same here. No crashes. Ok, that's even better, thanks for confirming! I'll close the bug. Sorry, I successfully reproduced the bug, or at least, a sigsegv. Hum... in facts, I was in a mix of stable, stable-bpo and testing. My problems might come from that mix, instead of from an actual bug in pgadmin3 (but it only show it pgadmin, and I'm sometimes using another wx3 application which is not impacted by it). I guess my last comment can be ignored. -- To UNSUBSCRIBE, email to debian-bugs-dist-requ...@lists.debian.org with a subject of unsubscribe. Trouble? Contact listmas...@lists.debian.org
Bug#750557: pgadmin3: segmentation fault
Le 11.09.2014 11:43, Bogdan Vatra a écrit : On Thursday 11 September 2014 11:31:44 Christoph Berg wrote: Hi guys, I've just uploaded 1.20.0~beta1 to unstable. Could you check if you still see the pgadmin3 crashes there? It seems to work for me now. Packages on apt.postgresql.org should be available shortly in the *-pgdg-testing suites. Christoph Hi, updated and tested, but no joy ;-(. Check http://paste.kde.org/pnrcj16nt BogDan. Same here, still segfaulting after right clicks. -- To UNSUBSCRIBE, email to debian-bugs-dist-requ...@lists.debian.org with a subject of unsubscribe. Trouble? Contact listmas...@lists.debian.org
Bug#750557: [Pkg-postgresql-public] Bug#750557: pgadmin: have random crashes ( memory corruption or double free )
Le 04.06.2014 15:48, Christoph Berg a écrit : Control: tag -1 moreinfo Re: bmorel 2014-06-04 20140604132642.9797.58545.reportbug@D2SI-P07-2.D2SI Package: pgadmin3 Version: 1.18.1-3 Severity: important File: pgadmin Tags: upstream Dear Maintainer, *** Reporter, please consider answering these questions, where appropriate *** * What led up to the situation? * What exactly did you do (or not do) that was effective (or ineffective)? * What was the outcome of this action? * What outcome did you expect instead? Can you provide any details on this? What did you do? What's the output? Can you provide a backtrace? Christoph Hello. Honestly, the crash is quite random. One thing which *might* be a cause could be that I am using a ssh connection: I first go on a distant server ( which does not have the database installed on it, but which is allowed to access the db ) so that pgadmin can then access the real DB server. I do not remember having that crash when working with my local database or with a database in a VM on my computer. About providing a backtrace... I have started pgadmin3 through my terminal, and the only thing it showed was memory corruption or double free. If you have some idea about how to have more informations about this crash, I would be happy to investigate further. -- To UNSUBSCRIBE, email to debian-bugs-dist-requ...@lists.debian.org with a subject of unsubscribe. Trouble? Contact listmas...@lists.debian.org
Bug#739161: marsshooter: incorrect dependency strength on fonts
Le 16.02.2014 14:53, Markus Koschany a écrit : Control: tags -1 moreinfo On 16.02.2014 11:37, bmorel wrote: Package: marsshooter Version: 0.7.5-2 Severity: normal File: /usr/games/mars marsshooter depends on 3 fonts ( fonts-gargi, fonts-wqy-microhei and fonts-tlwg-waree ) but works perfectly fine without them. ( I used #dpkg --force-all --purge fonts-gargi fonts-tlwg-waree fonts-wqy-microhei ) They should probably be only recommended ( note that I did not tried with fonts-dejavu, but it is probably the same, except that this font is probably really used by my system, of course ). Hello, it depends on your point of view and whether you are Chinese or Thai or not. M.A.R.S. supports many different languages and not all chars and glyphs can be provided by fonts-dejavu. That means if you switch to the Chinese language in the option menu, without those fonts nothing meaningful will be displayed on your screen. If I were Chinese I would then file a bug report against marsshooter. Debian Policy 7.2 is quite clear about the meaning of Depends and Recommends https://www.debian.org/doc/debian-policy/ch-relationships.html The Depends field should be used if the depended-on package is required for the depending package to provide a significant amount of functionality. I think the fonts provide this significant amount of functionality here because without them some languages will not be displayed correctly and the in-game menu will become non-functional. Please also note that these fonts are normally integrated in the upstream version of M.A.R.S. For reasons of deduplication and efficiency we only link to corresponding fonts packages in Debian, so that bugs in font packages can be fixed speedily and without having to interact with hundreds of packages separately. In short: I do not think this is a bug and the package must depend on these fonts to work correctly for all users. Please report back, if you agree with this answer. Regards, Markus I see your point. To be honest, I would not name what I reported a bug, but a suggestion ( since for me, a bug is a wrong behavior ). AFAIK, there is no suggestionreport tool in debian :) I did not meant that asian users should not be able to use correct fonts, but that an asian user will probably not need european fonts, and european people won't need asian ones. Plus, any user might simply prefer the font he installed himself on his computer instead of dejavu ( I quickly did a try without any fonts mars needed, it was able to find one and go with it. I think that it did not even changed font, which would mean that I have dejavu installed for nothing, so I'll investigate on how fonts are managed on my system later... ). It sounds like having an explicit dependency to xorg on graphical softwares. Almost no one have it, they rely on the system to already have it. Maybe that's wrong, I do not know, but it makes me think that the better solution would be a recommend dependency. However, if you prefer to keep the strong dependency, what about depending on at least one of those fonts? Opera does that, for fonts, and other packages likes xserver-xorg does that for other kinds of things. Some even does that with libraries (which probably is not as easy as for things like fonts, because of linkage). I think that this way would be a good compromise. People would still have a fonts required, but at least they would not have all fonts for nothing. -- To UNSUBSCRIBE, email to debian-bugs-dist-requ...@lists.debian.org with a subject of unsubscribe. Trouble? Contact listmas...@lists.debian.org
