TONYC TPF grant report #9

[Hours]         [Activity]
2013/08/19      Monday
 2.72           #119321 - remove stringification, test, track down
                Devel::TrackSIG strangeness (local issue)
=====
 2.72

2013/08/20      Tuesday
 0.38           #116132 - reject with comment
 0.35           #117649 - reject with comment
 1.58           #85104 - write tests, try to have external signals work at
                all on win32
 2.23           #85104, #116264 - testing, write-up
=====
 4.54

2013/08/21      Wednesday
 1.83           #85104 - write test using fork emulation, testing
 0.60           #113054 - extend test mechanism
 2.48           #113054 - test cross platform, apply, resolve
=====
 4.91

2013/08/22      Thursday
 1.02           #119351 - make a patch and test against Devel-Cover 1.06
 0.65           #3330 - split off the variable names into 119415, close
                3330
 0.35           #115894 - review and reject ticket too
 1.82           #78066 - attempt to reproduce in a number of ways, comment
=====
 3.84

2013/08/23      Friday
 0.32           #119413 - comment
 0.45           #115928 - comment, ask Tux about Configure
 0.55           #86686 - apply
 1.28           #85296 - apply, perldelta both
 0.67           #119429 - review
 1.23           #119429 - update bump-perl-version, research, apply
=====
 4.50

Which I calculate is 20.51 hours.

Approximately 14 tickets were worked on, and 4 patches were applied.

I applied the 3 patches I adapted last week for Math-BigInt([perl
#86686], [perl #85296]) and for find2perl [perl #113054].

I spent a fair amount of time attempting to write a test for [perl
#85104] for Win32. The original complaint for this ticket was that $!
wasn't preserved across signal handlers, this had been fixed in
5.14.0, but the ticket remained open since $^E wasn't being preserved
on platforms other than VMS.

I tried sending SIGINT (and SIGBREAK) cross process, but in most cases
Windows would refuse to deliver them, or would kill the entire process
group.  I ended up using pseudo-fork and an emulated NUM signal which
reproduced the original failure and confirmed the fix.

• TONYC TPF grant report #9 by Tony Cook