Re: DBI development release ready to ship
On 06/16/10 19:37, Tim Bunce wrote: Done. But there's no need to wait for those results as there are some windows failures on 1.611_91 that need fixing: http://matrix.cpantesters.org/?dist=DBI%201.611_91;reports=1;os=MSWin32 As short explained in irc, Cwd::abs_path from PathTools 3.31 seems to die when the given file doesn't exists: Cwd::abs_path ("C:\\senseless-dir\\senseless.csv"); Calculating the absolute path was done twice there, that's why I removed it ($searchdir should be an absolute path name anyway). What remains: $dbh->do (q{select * from "non/existing/dir/file.csv"}); will die for the same reason some lines earlier. There're now 2 ways we can go: 1) we could prove existence of the directory before calling abs_path and croak on our own 2) we can fix the bug in Cwd and require fixed PathTools for DBI-1.612 Jens
Re: DBI development release ready to ship
On Wed, Jun 16, 2010 at 03:18:15PM +0200, Jens Rehsack wrote: > 2010/6/16 Tim Bunce : > > On Wed, Jun 16, 2010 at 08:41:27AM +0200, Jens Rehsack wrote: > >> 2010/6/16 Tim Bunce : > >> > On Tue, Jun 15, 2010 at 09:10:38PM +0200, Jens Rehsack wrote: > >> >> Hi Tim, > >> >> > >> >> from my point of view, the current trunk of DBI is ready to be shipped > >> >> as next development release. > >> > > >> > Not in MANIFEST: t/51dbm_file.t > >> > Not in MANIFEST: t/52dbm_complex.t > >> > > >> > Should they be in the MANIFEST? > >> > >> They should, my fault. I'll do it now. > >> > >> > > Because it should be uploaded together with the development release of > >> > > SQL::Statement, we should synchronize our uploads. > >> > > >> > Please tell me more about this. Why synchronize? > >> > >> DBI has test-depends (and recommendation) for SQL::Statement 1.28 > >> and S::S has test-depends on DBI 1.612. > > > > I think the DBI shouldn't have test-depends for SQL::Statement. > > The recommendation is fine though. > > It has a test-depends - but the tests are skipped if S::S is not available. > My though was in the direction "I want to have the feedback from CPAN > testers regarding the new S::S". I don't see a test-depends on S::S in the DBI's Makefile.PL. Only a recommends (plus a conflicts for SQL::Statement => 1.27). > >> Some tests in DBI work only with SQL::Statement (and vice versa). > > > > Okay, but they're skipped if SQL::Statement isn't defined (right?) > > so the tests don't strictly *depend* on SQL::Statement. > > No, they're simply skipped (or fallback to DBI::SQL::Nano) without > SQL::Statement. I want to see if there're more incompatibilities like > that one from Birmingham.pm. Those are not discovered without SQL::Statement > :) > > Can you upload a DBI-1.611_92 now (with 2 more tests and the fixed Changes)? Done. But there's no need to wait for those results as there are some windows failures on 1.611_91 that need fixing: http://matrix.cpantesters.org/?dist=DBI%201.611_91;reports=1;os=MSWin32 Tim.
Re: DBI development release ready to ship
2010/6/16 Tim Bunce : > On Wed, Jun 16, 2010 at 08:41:27AM +0200, Jens Rehsack wrote: >> 2010/6/16 Tim Bunce : >> > On Tue, Jun 15, 2010 at 09:10:38PM +0200, Jens Rehsack wrote: >> >> Hi Tim, >> >> >> >> from my point of view, the current trunk of DBI is ready to be shipped >> >> as next development release. >> > >> > Not in MANIFEST: t/51dbm_file.t >> > Not in MANIFEST: t/52dbm_complex.t >> > >> > Should they be in the MANIFEST? >> >> They should, my fault. I'll do it now. >> >> > > Because it should be uploaded together with the development release of >> > > SQL::Statement, we should synchronize our uploads. >> > >> > Please tell me more about this. Why synchronize? >> >> DBI has test-depends (and recommendation) for SQL::Statement 1.28 >> and S::S has test-depends on DBI 1.612. > > I think the DBI shouldn't have test-depends for SQL::Statement. > The recommendation is fine though. It has a test-depends - but the tests are skipped if S::S is not available. My though was in the direction "I want to have the feedback from CPAN testers regarding the new S::S". >> Some tests in DBI work only with SQL::Statement (and vice versa). > > Okay, but they're skipped if SQL::Statement isn't defined (right?) > so the tests don't strictly *depend* on SQL::Statement. No, they're simply skipped (or fallback to DBI::SQL::Nano) without SQL::Statement. I want to see if there're more incompatibilities like that one from Birmingham.pm. Those are not discovered without SQL::Statement :) Can you upload a DBI-1.611_92 now (with 2 more tests and the fixed Changes)? Jens
Re: DBI development release ready to ship
On Wed, Jun 16, 2010 at 08:41:27AM +0200, Jens Rehsack wrote: > 2010/6/16 Tim Bunce : > > On Tue, Jun 15, 2010 at 09:10:38PM +0200, Jens Rehsack wrote: > >> Hi Tim, > >> > >> from my point of view, the current trunk of DBI is ready to be shipped > >> as next development release. > > > > Not in MANIFEST: t/51dbm_file.t > > Not in MANIFEST: t/52dbm_complex.t > > > > Should they be in the MANIFEST? > > They should, my fault. I'll do it now. > > > > Because it should be uploaded together with the development release of > > > SQL::Statement, we should synchronize our uploads. > > > > Please tell me more about this. Why synchronize? > > DBI has test-depends (and recommendation) for SQL::Statement 1.28 > and S::S has test-depends on DBI 1.612. I think the DBI shouldn't have test-depends for SQL::Statement. The recommendation is fine though. > Some tests in DBI work only with SQL::Statement (and vice versa). Okay, but they're skipped if SQL::Statement isn't defined (right?) so the tests don't strictly *depend* on SQL::Statement. Tim.
Re: DBI development release ready to ship
On Wed, Jun 16, 2010 at 09:03:09AM +0200, Jens Rehsack wrote: > 2010/6/16 Tim Bunce : > > On Tue, Jun 15, 2010 at 09:10:38PM +0200, Jens Rehsack wrote: > >> Hi Tim, > >> > >> from my point of view, the current trunk of DBI is ready to be shipped > >> as next development release. > > > > Great. > > I really get blushed since you named me twice in Changes - but > shouldn't Merijn get an entry, too? Yes, it was a rushed (non)copy-n-paste error. Fixed already. Thanks. Tim.
Re: DBI development release ready to ship
2010/6/16 Tim Bunce : > On Tue, Jun 15, 2010 at 09:10:38PM +0200, Jens Rehsack wrote: >> Hi Tim, >> >> from my point of view, the current trunk of DBI is ready to be shipped >> as next development release. > > Great. I really get blushed since you named me twice in Changes - but shouldn't Merijn get an entry, too? Jens
Re: DBI development release ready to ship
2010/6/16 Tim Bunce : > On Tue, Jun 15, 2010 at 09:10:38PM +0200, Jens Rehsack wrote: >> Hi Tim, >> >> from my point of view, the current trunk of DBI is ready to be shipped >> as next development release. > > Not in MANIFEST: t/51dbm_file.t > Not in MANIFEST: t/52dbm_complex.t > > Should they be in the MANIFEST? They should, my fault. I'll do it now. > > Because it should be uploaded together with the development release of > > SQL::Statement, we should synchronize our uploads. > > Please tell me more about this. Why synchronize? DBI has test-depends (and recommendation) for SQL::Statement 1.28 and S::S has test-depends on DBI 1.612. Some tests in DBI work only with SQL::Statement (and vice versa). > Tim. >
Re: DBI development release ready to ship
On Tue, Jun 15, 2010 at 09:10:38PM +0200, Jens Rehsack wrote: > Hi Tim, > > from my point of view, the current trunk of DBI is ready to be shipped > as next development release. Not in MANIFEST: t/51dbm_file.t Not in MANIFEST: t/52dbm_complex.t Should they be in the MANIFEST? Tim.
Re: DBI development release ready to ship
Also, FYI, please don't use tim.bu...@gmail.com - I use it as a read-only archive and (occasionally) route outbound smtp through it. Any email sent only there is very unlikely to be seen. Tim.
Re: DBI development release ready to ship
On Tue, Jun 15, 2010 at 09:10:38PM +0200, Jens Rehsack wrote: > Hi Tim, > > from my point of view, the current trunk of DBI is ready to be shipped > as next development release. Great. > Because it should be uploaded together with the development release of > SQL::Statement, we should synchronize our uploads. Please tell me more about this. Why synchronize? Tim.
DBI development release ready to ship
Hi Tim, from my point of view, the current trunk of DBI is ready to be shipped as next development release. Because it should be uploaded together with the development release of SQL::Statement, we should synchronize our uploads. Jens