This is an automated email from the git hooks/post-receive script. fsfs pushed a change to annotated tag upstream/0.10 in repository libio-pty-easy-perl.
at 921caca (tag) tagging c945c1f01c6d1517120bc26686deba05bebd0253 (commit) replaces upstream/0.09 tagged by Florian Schlichting on Wed Oct 19 12:15:19 2016 +0200 - Log ----------------------------------------------------------------- Upstream version 0.10 Florian Schlichting (1): New upstream version 0.10 Jesse Luehrs (31): localize $@ and $? in the destructor change the class to be globref-based update some tests that accessed the hashref directly i guess i do want this to be defined not sure where this pod came from add a few accessor methods, to make the api cleaner namespace the hash slots i'm using, as noted in the IO::Handle docs update Changes stop explicitly depending on core modules add IO::Pty::HalfDuplex to SEE ALSO bump version to 0.06 allow disabling the automatic $self->set_raw on spawn document the new raw option actually, the parent should never be automatically set raw update Changes switch to Dist::Zilla update Changes remove vim fold markers reformat a bunch of comments fix up doc links break some circular references changelog and bump version cleanups close the pty before killing the underlying process cleanups, dzil stuff, etc changelog changelog fix these docs changelog skip flaky test for now changelog Stefan O'Rear (1): Fix BSD by draining the kernel pty buffer before waiting doy (28): add failing test version bump to 0.04 $self->close when the object is destroyed don't fail on multiple close calls minor nitpicks clean up signals a bit better move the wait() call from the SIGCHLD handler into kill(), and use IGNORE for SIGCHLD instead cleaner logic in is_active don't reset SIGWINCH if we didn't set it in the first place close the child ptys after tests that spawn things in them system() stuff should be fixed don't eat cpu while waiting for a process to die make test 004 timeout rather than hanging on systems where it's broken use waitpid rather than wait, and don't mess with SIGCHLD at all why was i using require here meh, no idea why this is necessary a few minor cleanups reformat Changes update changelog for 0.04 no need for a changelog in README if i already have Changes update my email address various other 0.04 boilerplate updates call new directly, not indirectly need to preserve $pty->read returning undef on timeout add a test for read returning undef on timeout add a couple tests for write(), some failing we use POSIX too version bump to 0.05 jluehrs2 (42): probably should get this thing into version control... probably should have a while ago in fact add cpan stuff move the module into place, add some boilerplate pod stuff clean up the pod stuff change to carp for error messages add actual documentation add vim folds we don't need to explicitly specify a max for writing, since we can use the string length add some more to the synopsis make read() consistent with the documentation typo pod also requires newlines before commands various doc changes die if spawning a subprocess fails doc fixes we have tests write a readme getting ready for a release fix the changelog syswrite defaults to sending the whole string s/PTY/pty/g other doc fixes old tag: 0.01 fix is_active to not rely on SIGCHLD, since it seems to create race conditions that i don't understand make kill block on the subprocess death by default, and allow sending signals other than TERM make spawn wait for the child to die before returning if there was an error don't know where this line came from comment formatting don't allow a subprocess to be spawned if one is already running this should have been a croak missed some PTY->pty changes this test also needs to make sure that the subprocess is dead before exiting (hasn't been an issue yet, but the situation is no different from 010) synopsis edit fixme use the new kill() interface in the tests rather than the perl builtin doc changes for 0.02 old tag: 0.02 we can't handle the term size changes if STDIN isn't a tty, since then we have nowhere to get the size changes from this test is unnecessary, since sysread will return EOF if the process is dead... the pty object doesn't hold onto the stuff that came out on stdout print something on success and failure, so that the read call doesn't block, and make sure that the spawned script survives long enough for us to read from it update for version 0.03 TAG 0.03 ----------------------------------------------------------------------- No new revisions were added by this update. -- Alioth's /usr/local/bin/git-commit-notice on /srv/git.debian.org/git/pkg-perl/packages/libio-pty-easy-perl.git _______________________________________________ Pkg-perl-cvs-commits mailing list Pkg-perl-cvs-commits@lists.alioth.debian.org http://lists.alioth.debian.org/cgi-bin/mailman/listinfo/pkg-perl-cvs-commits