This week on the Perl 6 mailing lists
"...I think you're the path toward sanity."
-- Larry Wall, in 'Patterns'
Earlier, Luke Palmer asked for a clear definition of when math should use floating points and when it should be integer-based. This led to a great deal of discussion.
This week, a subthread started by TSa continued from Darren Duncan's previous post which proposed distinct operators for integer division/modulus and non-integer division/modulus. TSa felt that modulus should be defined by the Euclidean definition. Mark J. Reed proposed to define it in the conventional way. Several other people replied to the thread.
Luke Palmer suggested removing two items from the smart match table in
S03, namely the Hash and Array with Any on the right side. He preferred
~~ operator, which would allow him to define his own
patterns. In short, with his proposal,
~~ would be singly-dispatched
based upon the right argument, which would determine how the left was
Larry Wall replied that he was concerned about the consistency between compile time and run time, but that was in fact also an argument against an anonymous Any on the right.
Jonathan Lang had some suggestions on how to implement reverse test instances. There was a brief discussion between Jonathan and Larry on this subject. Ashley Winters also offered a comment.
Larry Wall committed a change to S03 which affects smartmatching. He noted that it is still subject to change. Nicholas Clark and Luke Palmer offered some suggestions, which Larry responded to. TSa also commented on the thread.
In this commit to S03, Larry Wall clarified how the Grammar pattern finds
its top rule. Jerry Gay questioned the use of
top as a lowercase
reserved word; he suggested
Another commit to S03 by Larry Wall modified smartmatching semantics. Darren Duncan noted a few possible errors, and Luke Palmer added some thoughts. This led Larry to consider a different approach, to avoid confusion.
This commit to S03
by Larry Wall included renaming
.exists, replacing it with
.contains; and removing
Joe Gottman proposed an infix minmax operator, which could be used as a reduction operator.
In ticket [perl #41196], James Keenan reported that the
t/tools/README mentioned in the MANIFEST did not exist. He included
a patch to remove the entry from the MANIFEST.
James Keenan created ticket [perl #41197] to report problems with a failing make. He elaborated on the problem, which happened just a couple of hours after his last successful build.
Lee Duhem proposed a facility for string differences in Test::More which could be used to perform tasks like ignoring insignificant whitespace. This would allow tests to continue passing even if the source had been changed to remove undesired whitespace.
Nicholas Clark felt that patches should be checked against the test
suite, and that there were dangers in relaxing the tests to allow
a broader definition of a string. Bernhard Schmalhofer replied that
t/examples/past.t was one of the few examples where ignoring
whitespace would be useful, and noted that it is now being handled with
a regular expression.
Xi Wang suggested a change in
io/io_win32.c to prevent a crash.
In ticket [perl #41198], Bernhard Schmalhofer pointed out that several
files are in the repository and mentioned in MANIFEST.generated, but that
they are added to MANIFEST by
Gay explained that the files belonged in MANIFEST, not MANIFEST.generated.
Paul Cochrane fixed this in r16482.
In ticket [perl #41201], Paul Cochrane noted that there is a temporary
Configure.pl which should be implemented properly.
Klaas-Jan Stol was working on implementing a PIR grammar which allows more complex expressions to be written. He submitted a patch of the grammar to date.
Ticket [perl #40905] was addressed with r16498. This task called for certain files to be exempted from coding standards.
This commit, by Paul Cochrane, adds a section explaining which files can be excluded from coding standards tests.
In ticket [perl #41214], Patrick R. Michaud pointed out that
make languages-test generates some files which aren't removed in
make clean or
make realclean. He stated that, at
make clean should remove them, but that perhaps the
tests should clean them up. Nicholas Clark commented that a pitfall
with the tests doing the cleanup is that the files aren't deleted if
the tests crash. Patrick agreed that in either case,
should remove them.
In ticket [perl #39905], Will Coleda requested that compiled PIR files
report error lines against the original
Jerry Gay created ticket [perl #41217] to report that he had been unable
to eliminate the warnings which were being generated in
In ticket [perl #41218], Jerry Gay reported some warnings coming from the imcc lexer code.
Andy Dougherty replied to Will Coleda about a test file which no longer exists. He stated that he did not mind closing the ticket ([perl #38584]), because the problem was probably not about Punie, but how Punie and PGE interacted. He included some information about his recent attempts to repeat the tests.
Allison Randal asked Andy if there was any way that someone could get access to the machine he was running the tests on, as they were not able to duplicate the problems.
In ticket [perl #41224], Jerry Gay noted that Test::Builder and Test skip differently. He wondered which was the correct way to handle things. He felt that Perl's 'ok' was better than Parrot's 'not ok' because the Perl module had been around for a while. chromatic agreed with Jerry and changed the Parrot module in r16553.
In ticket [perl #41226], Jerry Gay suggested that people could look through DEPRECATED.pod and deprecate things which have been scheduled for deprecation for some time.
Jerry Gay complained that the perlcritic coding standards test was using
too many resources, and asked in ticket [perl #41230] if someone
could look in to this problem. Paul Cochrane reported that he thought
there was a memory issue relating to
critique() and that the memory
was only freed up when the tests for a given policy are complete. Paul
elaborated on what what he had attempted to address the issue.
In ticket [perl #41231], Jerry Gay noted that coding standards must apply to core code, but not to languages.
Earlier, chromatic submitted a patch to implement
get_name() as specified
in PDD21. Jerry Gay created ticket [perl #41235] to track the status
of the patch, which he would like to see added before release 0.4.8.
chromatic said that he hadn't applied it because a deprecation cycle
was needed for the
Jerry Gay created ticket [perl #41237] to address an item in DEPRECATED.pod about PMC Class name IDs. He felt that either it should use one syntax or the other, but not both.
Jerry Gay created ticket [perl #41238] as a placeholder for an item in DEPRECATED.pod, namely removing unused PMCS PerlString, PerlUndef, etc.
Jerry Gay made a list of some of the core PMCs which he feels are undertested. The ticket [perl #41239] suggests that more tests be created.
Will Coleda requested additional information on ticket [perl #41241], about Tcl not running on Win32.
Steve Peters reported that Parrot can now be compiled with Borland C++ on Win32, but that this had linking problems. The ticket is [perl #41243].
Jerry Gay reminded people to update copyright notices from 2006 to 2007. This was ticket [perl #41244].
In ticket [perl #41247], Jarkko Hietaniemi reported a failure on
tge/parser. Patrick R. Michaud reported that it had been
fixed in r16570.
In ticket [perl #41248], Jarkko Hietaniemi reported a failure with
pge_util on the tru64 platform. Patrick R. Michaud reported that it was
fixed in r16570.
Jarkko Hietaniemi reported a core dump in
t/pmc/io_1.pir on tru64, in ticket
In ticket [perl #41250], Jarkko Hietaniemi informed about failures
p5regex/p5rx on the tru64 platform. Patrick R. Michaud reported that
he believed the failures were resolved in r16593 and asked for a confirmation.
Ticket [perl #41251] holds a report from Jarkko Hietaniemi
about a core dump caused by
Jarkko Hietaniemi reported a core dump on tru64 from
This was done in ticket [perl #41253].
In ticket [perl #41254], Jarkko Hietaniemi gave information about a
core dump from
library/pg on tru64.
In ticket [perl #41255], Jarkko Hietaniemi reported core dumps from
t/pmc/pmc_5.pasm with tru64.
In ticket [perl #41256], Jarkko Hietaniemi reported that many tests
t/pmc/complex fail because of nonportable assumptions about floating
Jarkko Hietaniemi created ticket [perl #41257] to report a core
t/pmc/pmc_5.pasm on tru64.
Klaas-Jan Stol looked at why Tcl cannot build on windows. He was unable to fix it, but passed on what he had gleaned from his experience. Matt Diephouse appreciated the information, and committed a fix to resolve the issue. Jerry Gay confirmed that r16596 worked.
Isaac Freeman was wondering how communication between the embedding program and the scripts run by the interpreter is implemented. He offered to implement it if it was likely to be used. Jeff Horwitz pointed Isaac to some code Jeff wrote over a year ago, which could be used as a starting point.
Gabriele Renzi worked problem 13 from Ovid's list of 99 Perl 6 examples. The code was submitted, and the question of if it should be committed was asked. Ovid replied that if there's a bit, there should be a commit.
This summary was prepared using Mail::Summary::Tools, available on CPAN.
If you appreciate Perl, consider contributing to the Perl Foundation to help support the development of Perl.
Thank you to everyone who has pointed out mistakes and offered suggestions for improving this series. Comments on this summary can be sent to Ann Barcomb, email@example.com.
This summary can be found in the following places: