Stepan Kasal

Show:

What's happening?

  • Comment: Do not check the version number, use library versioning

    The problem was observed in Fedora GNU/Linux: https://bugzilla.redhat.com/show_bug.cgi?id=496308.

    2009-04-17 18:21:05 UTC in PycURL

  • Do not check the version number, use library versioning

    At the end of function initpycurl(), that is at the end of src/pycurl.c, the version of the libcurl used at runtime is checked to be >= than the one used to build PycURL. That's harmless sanity check for a system build from scratch, but it is not friendly for packagers of any binary distribution, e.g. traditional binary GNU/Linux distributions (as oposed to source distros). The version...

    2009-04-17 18:18:13 UTC in PycURL

  • Comment: flex-2.5.34 (test-pthread) fails on /**/ sometimes

    The problem is that flex miscompiles itself. (The same holds for 1849805, see my today's comment there.) flex-2.3.34 is not able to correctly process one of its sources, scan.l. When I checkout 2.5.34 from CVS and build it, the testsuite passes. But in this case, scan.l has been processed by my installed flex-2.5.33. Then I rebuild scan.c by the just built flex: rm scan.c; make...

    2008-01-15 09:58:45 UTC in flex: the fast lexical analyser

  • Comment: flex-2.5.34 (test-pthread) fails on /**/ sometimes

    Ooops, the interface eats white space, let's show the difference this way: --- scanner-ok.l 2008-01-15 10:47:57.000000000 +0100 +++ scanner-fail.l 2008-01-15 10:49:18.000000000 +0100 @@ -1,3 +1,3 @@ -/* */ + /* */ %% %%.

    2008-01-15 09:53:35 UTC in flex: the fast lexical analyser

  • Comment: flex-2.5.34 (test-c++-multiple-scanners) fails on //

    The problem is that flex miscompiles himself: flex-2.3.34 is not able to correctly process one of its sources, scan.l. When I checkout 2.5.34 from CVS and build it, the testsuite passes. But in this case, scan.l has been processed by my installed flex-2.5.33. Then I rebuild scan.c by the just built flex: rm scan.c; make LEX=./flex flex and voila, flex is not able to compile...

    2008-01-15 09:17:16 UTC in flex: the fast lexical analyser

  • $(docdir) conflicts with CVS Autoconf

    CVS Autoconf introduces docdir, which defaults to $prefix/share/doc/packagename xmldocs.make also defines docdir, which conflicts with this variable. (I looked at current CVS.) Automake issues a warning. According to my tests with Automake 1.9.6, both definitions are outputted, and the "docdir = @docdir@" is lower, so things won't work as expected. Autoconf only obeys GNU Coding...

    2006-03-15 13:19:32 UTC in ScrollKeeper

  • moderator page with details misses a submit button

    Working with version 2.1.5. I moderate a few lists. When I go from summary page to "wiev the detail of individual postings", I get a page which has a submit button at the bottom. Similar button should be on the top, as with the Summary page. Besides being an inconvenience, this can also be confusing: I checked "Discard all msgs marked as deferred" on the top, then moved to the end...

    2004-07-30 13:30:40 UTC in Mailman

About Me

  • 2004-07-23 (5 years ago)
  • 1089759
  • kasal (My Site)
  • Stepan Kasal

Send me a message