This was a quiet week -- summer approaches -- but a few interesting points were raised. New warnings, portability points, and miscellaneous bugs are covered in this summary.
Following-up to bug report #22599,
Robin Barker added a new warning,
in the printf category : Newline in left-justified string for printf.
It's issued if a string to be left-justified by printf() or sprintf() (via the
%-ns format) contains a newline.
Ton Hospel reports (bug #22613) that if the loop variable used in a for() loop is
$\ (the output record separator),
then this new value doesn't in fact affect the behavior of print().
Andy Dougherty reports some test failures for ExtUtils with maintperl on Solaris 8. They're caused by a difference between GNU make and Sun make, in handling commands split over several lines.
Alan Ferrency investigated bug #18849 : on some recent versions of FreeBSD (or other *BSD, including Mac OS X), if the SIGCHLD handler is set to 'IGNORE' and a child process is forked, then all subsequent system() calls made in the parent process will hang until the previously forked child has exited. The confusion comes from the SA_NOCLDWAIT flag, how it's defined, implemented and supported by different UNIX vendors. See the thread for the portability round-up (Tru64, Solaris, Linux, AIX, IRIX). This ended in a patch (#19765) specific to BSDish OSes.
Aaron J Mackey finds that an untie() from within a FETCH or STORE is broken with perl 5.8.0. It worked in previous perls. He proposed a patch that reverts to the internals of untie() as they were in 5.6.x, but the root of the problem is not clear.
Bug: http://www.xray.mpe.mpg.de/mailing-lists/perl5-porters/2003-06/msg00365.html Patch: http://www.xray.mpe.mpg.de/mailing-lists/perl5-porters/2003-06/msg00422.html
Rafael removed the
$* special variable. See last week's summary for why. Using it still triggers a warning, though.
Anthony Heading manages to get a panic error out of the regular expression compiler. The offending regexp is
Steve Grazzini found out that a recent change in the formatting of the filetest entries in perlfunc broke perldoc.
Dan Kogai tested perl on FreeBSD 5.1. He suggested to use perl's malloc() instead of the system's one, which is too slow in perl's context.
Ilya Zakharevich proposed a lot of little OS/2 patches. This notably includes a test fix for MakeMaker. He also worked on perl's malloc().
This week's discussions were summarized by Rafael Garcia-Suarez. Weekly summaries are published on http://use.perl.org/ and/or via a mailing list, which subscription address is firstname.lastname@example.org. Comments, corrections, additions, and suggestions are welcome.