after I commit or add, the Close button terminates the whole application, not only the Commit or Add window

Did your commit succeed?

I've only noticed this when python comes across an error and presents an "ignore or close" dialogue.

Happens also in Win7 64-bit, Bazaar Explorer 1.0.1.

Commit succeeded but the application crashed and had to be started again.

Yes, I was able to commit... but the application closed instead of the
commit window only

2010/4/9 lsuser <email address hidden>

> Happens also in Win7 64-bit, Bazaar Explorer 1.0.1.
>
> Commit succeeded but the application crashed and had to be started
> again.
>
> ** Attachment added: "Bzr_Explorer_friendly_crash_Win7_64bit.png"
>
> http://launchpadlibrarian.net/43514255/Bzr_Explorer_friendly_crash_Win7_64bit.png
>
> --
> after I commit or add, the Close button terminates the whole application,
> not only the Commit or Add window
> https://bugs.launchpad.net/bugs/543913
> You received this bug notification because you are a direct subscriber
> of the bug.
>

--
Fábio Leitão
..-. .- -... .. --- .-.. . .. - .- --- ...-.-

What will happens if you press Esc button on the keyboard instead of Close on the dialog?

Something wrong here, can you tell us your OS version and also PyQt/Qt versions (available via Explorer main menu -> Help -> About).

Sorry for the delay...

Windows 7, Bzr 2.1.1 from standalone setup (what ever comes with it) and on Ubuntu 10.04.01 fully up to date with bzr and bzrtools together with bzrexplorer and whatever dependencies selected by the package mantainer.

Bazaar Explorer — Controle de Versão para Seres Humanos
Versão 1.0.1 "Marco Polo"

QBzr 0.18.6, bzrlib 2.1.1, PyQt 4.7.2, Qt 4.6.2, Python 2.6.5

Happened to me, too. I actually did this over and over again, because the actual behavior is so very different from the expected behavior.

Using Mac OS X 10.6 and Bazaar installed with the latest installer (v2.2.0 I believe).

I believe bzr-explorer is exhibiting similar behavior. After submitting a commit, the commit succeeded but the program crashed and closed altogether. I'll will include a backstrace as soon as I can, there a lot of debugging symbols that I need to install.

ProblemType: Bug
DistroRelease: Ubuntu 9.10
Package: bzr-explorer Version 1.1b1 "Bartolomeu Dias"
ProcVersionSignature: Ubuntu 2.6.31-22-generic-pae
Uname: Linux 2.6.31-22-generic-pae
Architecture: i386
Dependencies:
QBzr 0.19.0, bzrlib 2.2.0, PyQt 4.6, Qt 4.5.2, Python 2.6.4

I could install the debugging symbols for python-qt4 so this backtrace maybe imcomplete.

I meant to say that I could NOT install the debugging symbols for python-qt4 so this backtrace maybe incomplete.

I changed the status as new information (OS version, version of QBzr, bzrlib, PyQt, python) has been provided regarding this bug.

Either this is related to some Qt/PyQt bug, or I just don't know what's going on here.

Can anybody test the latest 2.2.1 release of standalone installer for Windows with Bazaar Explorer 1.1.1 (you can use the corresponding installer to update your version) and tell me if this error present there? I never got this error on Windows XP.

I can confirm this bug when used on Fedora 10 x64.
The versions are:
Bazaar Explorer version 1.1.2dev "Bartolomeu Dias"
QBzr 0.19.1, bzrlib 2.2.1, PyQt 4.6.2, Qt 4.5.3, Python 2.5.2

When home I'll check this on XP, Vista and Win7.

Checked that on Windows XP, Vista and 7. When using official Bazaar installer 2.2.1-3 NO such problem occurs.
Versions as reported by Bazaar Explorer::
Bazaar Explorer 1.1.1
QBzr 0.19.2
bzrlib 2.2.1
PyQt 4.5.2
Qt 4.5.2
Python 2.6.4

I tried several revisions of bzr-explorer on Fedora10_x64 with QBzr 0.19.1, bzrlib 2.2.1, PyQt 4.6.2, Qt 4.5.3, Python 2.5.2. When using revison 494 (tagged as version 1.1.1) everything works fine, and with the next revision (r495) this bug is there. So it looks like this regression was introduced in revision 495 after 1.1.1 release

Pavlo Shchelokovskyy пишет:
> I tried several revisions of bzr-explorer on Fedora10_x64 with QBzr
> 0.19.1, bzrlib 2.2.1, PyQt 4.6.2, Qt 4.5.3, Python 2.5.2. When using
> revison 494 (tagged as version 1.1.1) everything works fine, and with
> the next revision (r495) this bug is there. So it looks like this
> regression was introduced in revision 495 after 1.1.1 release

Thank you for testing!

after some more testing I've managed to find this bug in the 1.1.1 release also. It affects at least commit, add and update dialogs

Simple way to reproduce it - make two updates of any working tree, even when there is nothing to update. On the second update the close button closes the whole bzr-explorer application. This works to some extend with two different commands (add+commit closes the whole app, but update+commit does not).

Again the specs:
Fedora10_x64
Bazaar Explorer Version 1.1.1
QBzr 0.19.1, bzrlib 2.2.1, PyQt 4.6.2, Qt 4.5.3, Python 2.5.2

So the bug is older than 1.1.1, it just that 1.1.2dev exposes it better.

I still see bzr-explorer exhibiting the same behavior after installing updated packages. After submitting a commit, the commit succeeded but the program crashed and closed altogether.

ProblemType: Bug
DistroRelease: Ubuntu 9.10
Package: bzr-explorer Version 1.1.1 "Bartolomeu Dias"
ProcVersionSignature: Ubuntu 2.6.31-22-generic-pae
Uname: Linux 2.6.31-22-generic-pae
Architecture: i386
Dependencies:
QBzr 0.19.3, bzrlib 2.2.2, PyQt 4.6, Qt 4.5.2, Python 2.6.4

While running bzr explorer from a terminal and after bzr explorer crashed and closed completely, the following error was displayed:

Failed to decode using ascii, falling back to latin1
Segmentation fault

Book 'em Dano пишет:
> While running bzr explorer from a terminal and after bzr explorer
> crashed and closed completely, the following error was displayed:
>
> Failed to decode using ascii, falling back to latin1

This line most likely comes from qdiff, and maybe unrelated to crash.

> Segmentation fault

This is the real problem, but why it's crashed?

--
All the dude wanted was his rug back

After another set of upgrades the problem seems to be gone :)
Current configuration:
Fedora 12 x86_64
Bazaar Explorer 1.1.2, QBzr 0.19.3, bzrlib 2.2.2, PyQt 4.7.3, Qt 4.6.2, Python 2.6.2

So probably it was/is a bug in PyQt 4.6.xx

Either this problem related to Qt/PyQT (and I won't be surprised) or there is some sort of race condition.

To workaround the former please try to use the latest available installers with newer PyQt libs inside (or upgrade your Linux). To workaround the latter please disable automatic refresh in Explorer options.

I have to confess I been using bzr-explorer less since the issue, but I will try it again fully up-to-date.... How do I disable the automatic refresh in the Explorer options? I don't seem to find it (Natty 64)

15.05.2011 18:35, PresuntoRJ пишет:
> I have to confess I been using bzr-explorer less since the issue, but I
> will try it again fully up-to-date.... How do I disable the automatic
> refresh in the Explorer options? I don't seem to find it (Natty 64)

Tools - Preferences - Behavior - Automatically refresh status report

http://doc.bazaar.canonical.com/explorer/en/visual-tour-gnome.html#preferences

--
All the dude wanted was his rug back