python2.4 testsuite failures & python2.4 on the buildd's
Some question about buildd configurations (python2.4_2.4.3-3 builds), if some tests should be disabled when running the testsuite on the buildds, or if these are things, where configurations can be changed on the buildd's. - test_ioctl wants /dev/tty, is this device available or not readable (mips at least)? test succeeds otherwise. - test_tcl hangs on some buildds (mipsel, but not mips, alpha, s390, amd64), passes on others. Why? - test_socket fails on amd64 (excelsior). configuration error? works ok on a local machine. test test_socket failed -- Traceback (most recent call last): File "/build/buildd/python2.4-2.4.3/Lib/test/test_socket.py", line 449, in testSockName self.assertEqual(name, ("0.0.0.0", PORT+1)) AssertionError: ('83.143.242.5', 50008) != ('0.0.0.0', 50008) - test_codeccallbacks fails with a bus error on ia64 and hppa, unreproducible on local machines. other test failures: - test_bz2 on alpha (decoding error) - test_bsddb3 on amd64 (locking error) Thanks for your input. Matthias -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED]
Re: RFS: python-urwid - Console UI Library for Python
On Sun, 2006-04-23 at 11:28 -0400, Ian Ward wrote: > I've never tried using setup.py to generate documentation. Could you > point me to an example of a module that does this? I don't really know enough about distutils or of any modules that do this. I think it would be fine to distribute both the compiled html and the source in the tarball, and expect people to run a separate command if they want to recompile the html. > It would seem more reasonable to add templayer to the build-depends > and include the pre-built docs with the standard package. That is the idea. > Let me know if I should submit templayer for inclusion. Please do. > For 0.9.3, feel free to drop the docs for now since they are not > DFSG-Free until we get these source-building issues sorted out. I'll fix the packaging and make a new orig.tar.gz. > That works. It's registered as python-urwid in freshmeat. Thanks. -- bye, pabs http://wiki.debian.org/PaulWise signature.asc Description: This is a digitally signed message part
Re: RFS: python-urwid - Console UI Library for Python
Paul Wise wrote: On Sat, 2006-04-22 at 12:03 -0400, Ian Ward wrote: What is the source code for tutorial.html? Debian requires source code where it exists (see the DFSG/social contract). Perhaps you could release 0.9.3.1 with the source code for tutorial.html (and reference.html if it has source). Preferably, the setup.py should build these from their source. I've never tried using setup.py to generate documentation. Could you point me to an example of a module that does this? The script that generates the tutorial has an external dependency: another module I've written called "templayer". If the doc source files are included, would all users installing Urwid have to get templayer and build the docs as part of the install? It would seem more reasonable to add templayer to the build-depends and include the pre-built docs with the standard package. Let me know if I should submit templayer for inclusion. For 0.9.3, feel free to drop the docs for now since they are not DFSG-Free until we get these source-building issues sorted out. When I do make a new release, does debian-python need any kind of notification? Nope, just the maintainer(s). If you use freshmeat, we can subscribe to the freshmeat releases. That works. It's registered as python-urwid in freshmeat. Ian -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED]
Re: python2.4 testsuite failures & python2.4 on the buildd's
On Sun, Apr 23, 2006 at 10:56:49AM +0200, Matthias Klose wrote: > Some question about buildd configurations (python2.4_2.4.3-3 builds), > if some tests should be disabled when running the testsuite on the > buildds, or if these are things, where configurations can be changed > on the buildd's. > - test_ioctl wants /dev/tty, is this device available or not > readable (mips at least)? test succeeds otherwise. Buildds are not guaranteed to have an associated tty, so this test would fail. (I think policy might require the availability of a tty for maintainer scripts, but it's definitely not required during package building...) > - test_codeccallbacks fails with a bus error on ia64 and hppa, > unreproducible on local machines. Try running your build under prctl --unaligned=signal to reproduce. The Debian buildds run with non-default unaligned trap handling in order to catch programming errors that *can* be fixed up in the kernel, but only at a high cost. > other test failures: > - test_bz2 on alpha (decoding error) Have you reproduced this failure outside the buildds, or do you need someone to give it a try? -- Steve Langasek Give me a lever long enough and a Free OS Debian Developer to set it on, and I can move the world. [EMAIL PROTECTED] http://www.debian.org/ signature.asc Description: Digital signature