--- Matt Zimmerman <mdz@debian.org> wrote:> On Fri, Jul 20, 2001 at 05:15:21PM -0700, Josh Coalson wrote: > > > I rethought it and it seemed like a bad idea to post > > a big file, so you can get it here: > > > > >http://cvs.sourceforge.net/cgi-bin/viewcvs.cgi/flac/junk/flac-1.0-src-candidate.tar.bz2> > With the attached patch, a complete "make distcheck" should work, > including the > self-tests. It adds missing directories to some SUBDIRS variables, > includes > some files that were otherwise missing from the distribution, fixes > CFLAGS to > use $(srcdir), adds an EXTRA_DIST where appropriate, adds some > missing > Makefiles to AC_OUTPUT, etc. > > The only functional change is that src/test_* is built by default, > even without > --enable-debug. This is consistent with GNU standards and other > auto* users, > and will allow the usual 'make check' idiom to work without fiddling > with > configure flags.Ok, cool. Actually I just did a sourceforge release so I went back and hid the source release until I get your patches integrated in. The binary releases are unaffected because I use the Makefile.lite system. I almost added man/ to SUBDIRS et al but then I thought it might break for people that don't have docbook-to-man and cause confusion.> The self-tests are finishing up now (boy there are a lot of noise > tests), so I > may have more patches if the rest of the distcheck fails for some > reason.OK, I'll wait a bit before redoing the source release. Josh __________________________________________________ Do You Yahoo!? Make international calls for as low as $.04/minute with Yahoo! Messenger http://phonecard.yahoo.com/
On Fri, Jul 20, 2001 at 08:10:51PM -0700, Josh Coalson wrote:> --- Matt Zimmerman <mdz@debian.org> wrote: > > > > The only functional change is that src/test_* is built by default, even > > without --enable-debug. This is consistent with GNU standards and other > > auto* users, and will allow the usual 'make check' idiom to work without > > fiddling with configure flags. > > Ok, cool. Actually I just did a sourceforge release so I went back and hid > the source release until I get your patches integrated in. The binary > releases are unaffected because I use the Makefile.lite system. > > I almost added man/ to SUBDIRS et al but then I thought it might break for > people that don't have docbook-to-man and cause confusion.What it will do is build flac.1 and include it in the source distribution (even though it is auto-generated). This ensures that no additional tools are needed to generate the man page.> > The self-tests are finishing up now (boy there are a lot of noise tests), > > so I may have more patches if the rest of the distcheck fails for some > > reason. > > OK, I'll wait a bit before redoing the source release.Self-tests are still going on my Pentium II. I'm also running them (at very low priority) on the ia64 that I have access to. -- - mdz
On Fri, Jul 20, 2001 at 08:10:51PM -0700, Josh Coalson wrote:> --- Matt Zimmerman <mdz@debian.org> wrote: > > The self-tests are finishing up now (boy there are a lot of noise tests), > > so I may have more patches if the rest of the distcheck fails for some > > reason. > > OK, I'll wait a bit before redoing the source release.The noise tests finished up OK, but test_bins.sh failed. I'm not really sure what that test is supposed to do, but here's what happened: Testing bins... ../../test_files/bins/b00: encode...ERROR during encode of ../../test_files/bins/b00 FAIL: ./test_bins.sh ==================1 of 3 tests failed ================== Should there be a test_files directory shipped with the distribution? -- - mdz
--- Matt Zimmerman <mdz@debian.org> wrote:> On Fri, Jul 20, 2001 at 08:10:51PM -0700, Josh Coalson wrote: > > > --- Matt Zimmerman <mdz@debian.org> wrote: > > > The self-tests are finishing up now (boy there are a lot of noise > tests), > > > so I may have more patches if the rest of the distcheck fails for > some > > > reason. > > > > OK, I'll wait a bit before redoing the source release. > > The noise tests finished up OK, but test_bins.sh failed. I'm not > really sure > what that test is supposed to do, but here's what happened: > > Testing bins... > ../../test_files/bins/b00: encode...ERROR during encode of > ../../test_files/bins/b00 > FAIL: ./test_bins.sh > ==================> 1 of 3 tests failed > ==================> > Should there be a test_files directory shipped with the distribution?hmm. test_files is in a separate cvs module since it is big. I never really meant it to be shipped, so conditionally disabling test_bins.sh sounds like a good idea. Josh __________________________________________________ Do You Yahoo!? Make international calls for as low as $.04/minute with Yahoo! Messenger http://phonecard.yahoo.com/
On Fri, Jul 20, 2001 at 11:51:29PM -0400, Matt Zimmerman wrote:> On Fri, Jul 20, 2001 at 08:10:51PM -0700, Josh Coalson wrote: > > > --- Matt Zimmerman <mdz@debian.org> wrote: > > > The self-tests are finishing up now (boy there are a lot of noise tests), > > > so I may have more patches if the rest of the distcheck fails for some > > > reason. > > > > OK, I'll wait a bit before redoing the source release. > > The noise tests finished up OK, but test_bins.sh failed. I'm not really sure > what that test is supposed to do, but here's what happened: > > Testing bins... > ../../test_files/bins/b00: encode...ERROR during encode of ../../test_files/bins/b00 > FAIL: ./test_bins.sh > ==================> 1 of 3 tests failed > ==================> > Should there be a test_files directory shipped with the distribution?If test_files is for some separate testing framework and the tests should be disabled in the distributed version, you should have test_bins.sh check for the existence of the test_files directory and exit with code 77 if it does not exist, e.g. test -d $(BINS_PATH) || exit 77 That will cause the test to be skipped, rather than fail. -- - mdz