CPAN-Reporter
view release on metacpan or search on metacpan
- Fixed: some *.PL files prompt without autoflush, which works
normally, but appears to silently hang when piping to a tee.
CPAN::Reporter now uses Devel::Autoflush in PERL5OPT to always turn
on autoflush
- Fixed: distributions with odd names wouldn't be recorded in history;
now they are recorded as 'DISCARD'
1.15 Mon May 19 11:21:47 EDT 2008
- Changed: now requires Test::Reporter 1.40
- No other changes from 1.14_xx series
1.14_04 Thu Apr 17 06:37:26 EDT 2008
- Documented: added section on transport config option
1.14_03 Thu Apr 17 06:37:26 EDT 2008
- Uploaded without Changes file updated
1.14_02 Sun Apr 6 09:45:39 EDT 2008
- Changed: improved diagnostics messages for 'transport' failures
1.14_01 Sat Apr 5 18:30:44 EDT 2008
- Changed: if 'transport' configuration option is not valid according
to Test::Reporter, no report will be sent instead of falling back
to Net::SMTP
1.13 Mon Mar 24 09:22:44 EDT 2008
- Changed: Build.PL or Makefile.PL that returns without error but
does *not* create either 'Makefile' or 'Build' will be marked as
'DISCARD' for the PL phase instead of 'PASS'.
- Fixed: Prereq check would be confused if loading modules also
printed anything to STDOUT (somehow CPAN.pm is a culprit); added
"select DEVNULL" before requiring modules as a workaround
1.12 Mon Mar 10 08:55:48 EDT 2008
- Added: Win32::FsType() to special variables section of report
- Added: diagnostic warning if PrereqCheck output can't be parsed
1.11 Mon Feb 25 18:47:03 EST 2008
- Bug fix: successful PL's where prereqs happen not to be satisfied
were mistakenly being marked as 'DISCARD' for the PL phase. Fixed so
that PL's that PASS will stay that way regardless of prerequisites.
(This does not prevent a later make or test failure from being
downgraded,
- Bug fix: skipfiles patterns should match case-insensitive (Cantrell);
also, documented lack of starting anchor
- Bug fix: command_timeout option would not accept zero as valid; now
zero is a valid option and disables command_timeout
- more minor boilerplate tweaks
1.10 Mon Feb 25 10:07:28 EST 2008
I hereby dub 1.10 as the "I hate Windows" release
- Bug fix (sort of): On Win32, child processes were not timing out.
Changed from Win32::Process to Win32::Job to fix that. However,
discovered that process termination on Win32 can deadlock in some
cases, so timeout testing on Win32 will now skip unless
$ENV{PERL_AUTHOR_TESTING} is set. Added warning to documentation.
- Bug fix: On a Win32 virtual machine (VirtualBox), handles for output
capture were not inherited by child processes, which caused Test::Harness
to crash for some reason.
1.09 Mon Feb 18 22:10:55 EST 2008
- Simplified and shortened boilerplate
- Bug fix (minor): better detection of make vs Build for progress
messages
- Test fix: skip interrupt testing on MSWin32 if Win32::Process
is not installed
1.0801 Sun Feb 10 00:27:05 EST 2008
- Test fix: adjust timeout timings in t/14_command_timeout.t to try to
avoid failures on heavily loaded machines
1.08 Sat Jan 26 18:52:45 EST 2008
- No changes from 1.07_06 (seems to pass smoke tests)
- Summary of new features since 1.06:
- Added have_tested() function to CPAN::Reporter::History
- Added 'cc_skipfile' and 'send_skipfile' advanced config options
- Added 'command_timeout' advanced config option
- Added support for PERL_CPAN_REPORTER_DIR and
PERL_CPAN_REPORTER_CONFIG environment variables
1.07_06 Fri Jan 25 11:27:28 EST 2008
- Added 'command_timeout' config to halt commands after a period of time
- Added detection of commands killed with a signal; reports are discarded
(N.B. this may change in a future version)
- Bug fix: processes timing out could still hang with child processes
(killing 'make test' still leaves the harness running); fixed by making
'make test' its own process group and killing the process group; hoping
that's portable across non-Win32; Win32 timeouts use Win32::Process and
do the right thing already
- Downgraded lots of progress messages from CPAN's 'mywarn' to 'myprint'
(from red to blue if you use CPAN's colored output)
1.07_05 Thu Jan 24 08:27:48 EST 2008
- Added PERL_CPAN_REPORTER_DIR and PERL_CPAN_REPORTER_CONFIG environment
variables to specify alternate locations for default configuration
directory and file (RT#30314)
- Removed File::Temp as a full prerequisite; now only used in testing
(Requested by Adam Foxson to extract features to CPAN::Testers::*)
1.07_04 Sun Jan 20 22:06:56 EST 2008
- Test fix: stopping cc on blead broke lots of other tests. Fixed those.
Built my own blead perl so I can actually check these things in the
future instead of relying on Andreas' blead smoking
- Test fix: testing skipfile wasn't being created properly on Perl 5.005
1.07_03 Sat Jan 19 22:16:44 EST 2008
- Test fix: set sample history file writeable after copy
1.07_02 Sat Jan 19 20:33:55 EST 2008
- Added have_tested() function to CPAN::Reporter::History; provide ability
to search the CPAN::Reporter history file
- Bug fix: PL files that warned "OS Unsupported" and exited with 0 will
now be flagged as NA; this bends the CPAN Testers "rules" slightly, but
"OS Unsupported" is a hack anyway, so this should help things DWIM
- Test fix: more changes to timeout testing to avoid false failures
- Authors will never be CC'd on reports when running maint/blead perl
(anything with $Config{perl_patchlevel} set)
- Reports discarded due to missing prerequisites will now be recorded
in the history file with a 'grade' of DISCARD (unless a duplicate
DISCARD already exists)
- Changed "edit_report" prompt to clarify this can be used to review
reports in addition to just editing them
- Shortened duplicate report warning to fit on one line
1.07_01 Tue Jan 8 01:41:02 EST 2008
- Added "cc_skipfile" and "send_skipfile" advanced config options to
( run in 0.788 second using v1.01-cache-2.11-cpan-995e09ba956 )