PySide Bugzilla Closed for New Bugs

PySide is now a Qt Add-on and uses the Qt Project's JIRA Bug Tracker instead of this Bugzilla instance. This Bugzilla is left for reference purposes.

Bug 524 - Debugging errors during work of createpackage.js on windows is hard
: Debugging errors during work of createpackage.js on windows is hard
Status: CLOSED FIXED
Product: PySide
Classification: Unclassified
Component: packaging
: 1.0.0 beta1
: All MS Windows XP/Vista/7
: P2 normal
Assigned To: Bruno Araujo
:
:
:
  Show dependency treegraph
 
Reported: 2010-12-03 16:05 EET by Vladimir Rutsky
Modified: 2010-12-20 11:26 EET (History)
9 users (show)

See Also:


Attachments
patch described in bug report (2.24 KB, patch)
2010-12-03 16:07 EET, Vladimir Rutsky
Details

Note You need to log in before you can comment on or make changes to this bug.
Description Vladimir Rutsky 2010-12-03 16:05:06 EET
In case of error createpackage.js close shell window with output of command
that produced error so there is no chance to see what exactly caused error.

E.g. I had mingw version of Qt SDK installed and all that createpackage.js
outputted was "Configuring ApiExtractor... Error configuring ApiExtractor"
(window with cmake that produced closed so fast that I had no chances to see
any  output with errors from cmake).

I propose use intead of:
  shell.Run("some_shell_command", 5, true))
construction like:
  shell.Run("cmd /c some_shell_command || (echo FAILED && pause && exit /b 1)",
5, true))

I added this kind of error handling during shell.run(), please see attached
patch.
Comment 1 Vladimir Rutsky 2010-12-03 16:07:10 EET
Created attachment 163 [details]
patch described in bug report
Comment 2 Matti Airas 2010-12-04 08:12:11 EET
Seems like a nice feature. I'm prioritizing P2 to get the patch in without
unnecessary delay.
Comment 3 Hugo Parente Lima 2010-12-06 10:31:50 EET
The current official Windows packages aren't created by those scripts, they are
created by Roman Lacko scripts instead, so IMO is better to get Roman scripts a
try, put them on git, etc.

I told him to do a merge request, I'll ping him again and maybe just ask for
the tarballs and push the files to the repository myself.
Comment 4 Hugo Parente Lima 2010-12-06 10:36:19 EET
hmmm I did read the bug description very fast, there is a patch :-), so I guess
we can apply your patch while the Roman Lacko scripts aren't on git yet.
Comment 5 Matti Airas 2010-12-07 06:19:01 EET
Hugo, I wasn't really sure how to handle this - whether we should forward the
issue to Roman or deal with it ourselves. Since the patch was against
createpackage.js which seemed to originate from you (or was it Renato?), I
thought we could as well take it. In any case, we probably should try to
prioritize bugs with patches. :-)
Comment 6 Vladimir Rutsky 2010-12-07 10:02:09 EET
I didn't succeed in building PySide on Windows using createpackage.js.
Script clones apiextractor repository, configures build, but fails with
compiling part (by running nmake), it says:

  NMAKE : fatal error U1064: MAKEFILE not found and no target specified
  Stop.

This happens because cmake don't generate Makefile, this probably can be fixed
by adding '-G "NMake Makefiles"' to cmake, as adviced by Stefan in mailing list
(http://www.mail-archive.com/pyside@lists.openbossa.org/msg01211.html), but I
didn't tried this.

I managed to build PySide using new tutorial posted by Roman Lacko at mailing
list (http://www.mail-archive.com/pyside@lists.openbossa.org/msg01165.html) (by
the way building required some adjustments, which probably I commit as bug
reports as soon as new tutorial will be published in wiki). Building PySide as
described in this tutorial don't require createpackage.js, and IMO
createpackage.js is obsolete and will be removed as soon as new tutorial will
be finished.

So IMO:
  1. We should wait release of new tutorial by Roman Lacko (and for now use
tutorial that he published on mailing list). 
  2. No improvements of createpackage.js script is needed because a) it will be
removed soon; b) it doesn't work. Anyway I see that my patch is already pushed
to master (e9a41fe5).
Comment 7 Hugo Parente Lima 2010-12-07 10:46:10 EET
It works, but it's obsolete and have some problems when the sources fail to
build.

Yes, I did need to compile PySide under Windows and took the chance to test
your patch in one go so I compiled PySide under Windows with createpackage.js
and your patch and it worked, them I pushed it to mainline, so now I don't know
if I can mark this bug as WONTFIX, INVALID or FIXED, because this script is
obsolete but the patch is already on mainline, Matti?
Comment 8 Matti Airas 2010-12-07 11:30:06 EET
Hugo, let's mark it fixed - the original issue of the bug was addressed and the
patch successfully applied.
Comment 9 Hugo Parente Lima 2010-12-07 15:14:17 EET
hmm... we must add this use case for our EM policies: "Trivial bugs with
patches included."
Comment 10 Matti Airas 2010-12-07 18:58:25 EET
(In reply to comment #9)
> hmm... we must add this use case for our EM policies: "Trivial bugs with
> patches included."

Yes, I agree. Let's discuss this tomorrow.
Comment 11 Hugo Parente Lima 2010-12-20 11:26:13 EET
Closign bugs after release of 1.0.0 beta2.