Paul Johnson [Mon, 14 May 2012 08:45:10 +0000 (09:45 +0100)]
Don't test that errno is still 0 after POSIX::f?pathconf
I think the best we can do with respect to the f?pathconf tests is to
make sure that the perl call doesn't die, and that the system call
doesn't fail. And it's arguable we should only be testing the former.
But since we've been testing more that this anyway, it's probably safe
to test both.
With respect to the sysconf call, I think we shouldn't test more than
that perl doesn't die. Any further testing would require different
tests based the argument being passed in. Before doing that, it's
probably worth considering the purpose of the tests. I don't think we
really want to test that POSIX has been implemented correctly, only that
our layer over it is correctly implemented.
This fixes RT #112866.
Karl Williamson [Mon, 14 May 2012 15:47:36 +0000 (09:47 -0600)]
perldelta: Remove duplicate paragraph
Ricardo Signes [Fri, 11 May 2012 22:00:03 +0000 (18:00 -0400)]
study as no-op is a bugfix, not performance enhancement
Father Chrysostomos [Fri, 11 May 2012 16:55:09 +0000 (09:55 -0700)]
perldelta: Add ‘(5.14.2)’ markers
Father Chrysostomos [Fri, 11 May 2012 16:50:20 +0000 (09:50 -0700)]
perldelta: Explain the ‘(5.14.1)’ markers
Father Chrysostomos [Fri, 11 May 2012 16:48:49 +0000 (09:48 -0700)]
perldelta: Use single quotes in C<>
C<> renders as "..." in nroff, so C<... "..." ...> ends up looking weird.
Karl Williamson [Fri, 11 May 2012 16:44:10 +0000 (10:44 -0600)]
perldelta: Use L<> to link to changed module pods
Spotted by Vincent Pit
Karl Williamson [Fri, 11 May 2012 16:35:13 +0000 (10:35 -0600)]
perldelta: Reorder to avoid pronoun confusion
Spotted by Zsbán Ambrus
Karl Williamson [Fri, 11 May 2012 16:29:31 +0000 (10:29 -0600)]
perldelta: typo
Spotted by Zsbán Ambrus
Karl Williamson [Fri, 11 May 2012 16:25:15 +0000 (10:25 -0600)]
perldelta: Add future deprecation text about \Q
Father Chrysostomos [Fri, 11 May 2012 16:30:25 +0000 (09:30 -0700)]
perldelta: misuse of commas
Father Chrysostomos [Fri, 11 May 2012 16:27:08 +0000 (09:27 -0700)]
perldelta: typo
Father Chrysostomos [Fri, 11 May 2012 16:26:43 +0000 (09:26 -0700)]
perldelta: [rt.cpan.org #0], not RT 0
Father Chrysostomos [Fri, 11 May 2012 16:24:14 +0000 (09:24 -0700)]
Rmv second ‘version’ in upgrade notices
Some of these were like this:
...from version 123 to version 456.
and some like this:
...from version 123 to 456.
Since the former is wordy, I’ve used the latter throughout.
Father Chrysostomos [Fri, 11 May 2012 16:20:46 +0000 (09:20 -0700)]
perldelta: Consistent fullstops for ‘upgraded from x to x’
Father Chrysostomos [Fri, 11 May 2012 16:18:44 +0000 (09:18 -0700)]
perldelta: consistent spaces after dots
Father Chrysostomos [Fri, 11 May 2012 16:17:10 +0000 (09:17 -0700)]
perldelta: consistent semicolons in CGI example
Father Chrysostomos [Fri, 11 May 2012 16:16:43 +0000 (09:16 -0700)]
perldelta: grammar
Father Chrysostomos [Fri, 11 May 2012 16:16:11 +0000 (09:16 -0700)]
perldelta: fix capitalisation
Karl Williamson [Fri, 11 May 2012 15:40:20 +0000 (09:40 -0600)]
perldelta: Mention 5.14.0, not 5.13.6
Karl Williamson [Fri, 11 May 2012 15:37:37 +0000 (09:37 -0600)]
perldelta: Correct statement
It was pointed out to me after I wrote the text in an earlier perldelta
that this one is extracted from, that it is extremely unlikely to run
out of memory; I had not bothered to really do the math.
Karl Williamson [Fri, 11 May 2012 15:36:45 +0000 (09:36 -0600)]
perldelta: correct statement
Karl Williamson [Fri, 11 May 2012 15:33:04 +0000 (09:33 -0600)]
perldelta: grammar
Ricardo Signes [Fri, 11 May 2012 14:06:39 +0000 (10:06 -0400)]
perldelta: slightly expand and clarify policy note
Ricardo Signes [Fri, 11 May 2012 12:18:05 +0000 (08:18 -0400)]
perldelta: break Pod:: deprecations onto two items
Ricardo Signes [Fri, 11 May 2012 12:07:25 +0000 (08:07 -0400)]
Revert "perl5160delta: The coreargs opcode is undeserving of mention"
This reverts commit
1061b56a7b2cc84a8ac96a405e5b8c185936605c.
This is a reversion of a reversion. The reversion in
1061b56a7b2cc was
a bizarre mistake made during merging some blead/release conflicts, and
rjbs sincerely apologizes!
Ricardo Signes [Fri, 11 May 2012 12:04:55 +0000 (08:04 -0400)]
add long-form keys for newer versions in CoreList
Dave Rolsky [Fri, 11 May 2012 07:28:33 +0000 (02:28 -0500)]
Various small grammar fixes in perldelta
Ricardo Signes [Fri, 11 May 2012 02:36:37 +0000 (22:36 -0400)]
perldelta: update "Updated Modules" with highlights
Ricardo Signes [Fri, 11 May 2012 01:24:47 +0000 (21:24 -0400)]
bump the CoreList version in CoreList for 5.16
Ricardo Signes [Thu, 10 May 2012 20:56:40 +0000 (16:56 -0400)]
skip the porting/utils.t unless in a git checkout
Today I tried to build 5.16.0-RC0 on my Linode and I got this:
ok 78 # skip utils/cpanp-run-perl executes code in a BEGIN block which fails for
empty @ARGV
not ok 79 - utils/cpanp compiles
# Failed test 79 - utils/cpanp compiles at porting/utils.t line 81
# got "defined(%hash) is deprecated at /usr/local/lib/perl5/site_perl/5.10.
0/Locale/Maketext/Lexicon.pm line 307.\n\t(Maybe you should just omit the define
d()?)\nutils/cpanp syntax OK\n"
# expected "utils/cpanp syntax OK\n"
Ugh. We really don't want this to happen to somebody else, because this
test is "do not let the developer break stuff" not "make sure the install
works."
Ricardo Signes [Thu, 10 May 2012 19:08:21 +0000 (15:08 -0400)]
add a changes_between function in Module::CoreList
Ricardo Signes [Thu, 10 May 2012 18:47:18 +0000 (14:47 -0400)]
point out "corelist --diff" in perldelta
Ricardo Signes [Tue, 1 May 2012 22:28:43 +0000 (18:28 -0400)]
add the --diff option to corelist
Ricardo Signes [Thu, 10 May 2012 18:37:22 +0000 (14:37 -0400)]
update Module::CoreList for 5.16.0
Ricardo Signes [Thu, 10 May 2012 18:34:27 +0000 (14:34 -0400)]
allow for .tgz dists in the CoreList updater
Ricardo Signes [Thu, 10 May 2012 18:17:04 +0000 (14:17 -0400)]
perldelta: the acknowledgements section!
Father Chrysostomos [Wed, 25 Apr 2012 05:23:34 +0000 (22:23 -0700)]
perl5160delta: The coreargs opcode is undeserving of mention
Ricardo Signes [Thu, 10 May 2012 13:38:02 +0000 (09:38 -0400)]
import perldelta from eb83ed8 into release branch
Leon Timmermans [Thu, 3 May 2012 12:19:31 +0000 (14:19 +0200)]
perldelta: Explain stdio/sfio future deprecation.
Ricardo Signes [Wed, 2 May 2012 12:27:24 +0000 (08:27 -0400)]
okay the links to CPAN modules in the perldelta
Ricardo Signes [Wed, 2 May 2012 12:27:06 +0000 (08:27 -0400)]
update .gitignore: we generate 5160delta now
Ricardo Signes [Wed, 2 May 2012 12:17:38 +0000 (08:17 -0400)]
regenerate uconfig.h
Ricardo Signes [Wed, 2 May 2012 02:33:19 +0000 (22:33 -0400)]
remove perl515*delta, add perl5160delta
Ricardo Signes [Wed, 2 May 2012 01:18:37 +0000 (21:18 -0400)]
bump version to 5.16.0 RC0
Done with:
./perl -Ilib Porting/bump-perl-version -i 5.15.9 5.16.0
...followed by a small edit to INSTALL and patchlevel.h.
Dominic Hargreaves [Wed, 9 May 2012 18:09:18 +0000 (19:09 +0100)]
add Test::More as a prereq to Makefile.PL
Tony Cook [Wed, 9 May 2012 18:04:28 +0000 (19:04 +0100)]
sometimes fork() isn't available
This was amended from the original Tony prepared in a parallel branch
Daniel Kahn Gillmor [Fri, 17 Feb 2012 22:29:14 +0000 (14:29 -0800)]
[rt.cpan.org #61577] sockdomain and socktype undef on newly accepted sockets
There appears to be a flaw in IO::Socket where some IO::Socket objects
are unable to properly report their socktype, sockdomain, or protocol
(they return undef, even when the underlying socket is sufficiently
initialized to have these properties).
The attached patch should cover IO::Socket objects created via accept(),
new_from_fd(), new(), and anywhere else whose details haven't been
properly cached.
No new code should be executed on IO::Socket objects whose details are
already cached and present.
Craig A. Berry [Wed, 9 May 2012 23:41:05 +0000 (18:41 -0500)]
Skip Carp tests on VMS.
They want IPC::Open3::open3, which is not currently working.
Father Chrysostomos [Wed, 9 May 2012 19:49:49 +0000 (12:49 -0700)]
perl5160delta: tweaks
sdio -> stdio
two spaces after dots
Leon Timmermans [Thu, 3 May 2012 12:19:31 +0000 (14:19 +0200)]
perldelta: Explain stdio/sfio future deprecation.
Ricardo Signes [Wed, 9 May 2012 01:11:12 +0000 (21:11 -0400)]
add a missing blink above =item to s2p.PL
Father Chrysostomos [Tue, 8 May 2012 15:26:54 +0000 (08:26 -0700)]
Fix test failure
Lesson learnt: After switching from threaded to unthreaded and fixing
the test, switch back again and re-run the test. :-)
Father Chrysostomos [Tue, 8 May 2012 03:43:18 +0000 (20:43 -0700)]
[perl #112780] Don’t set cloned in-memory handles to ""
PerlIO::scalar’s dup function (PerlIOScalar_dup) calls the base imple-
mentation (PerlIOBase_dup), which pushes the scalar layer on to the
new file handle.
When the scalar layer is pushed, if the mode is ">" then
PerlIOScalar_pushed sets the scalar to the empty string. If it is
already a string, it does this simply by setting SvCUR to 0, without
touching the string buffer.
The upshot of this is that newly-cloned in-memory handles turn into
the empty string, as in this example:
use threads;
my $str = '';
open my $fh, ">", \$str;
$str = 'a';
async {
warn $str; # something's wrong
}->join;
This has probably always been this way.
The test suite for MSCHWERN/Test-Simple-1.005000_005.tar.gz does some-
thing similar to this:
use threads;
my $str = '';
open my $fh, ">", \$str;
print $fh "a";
async {
print $fh "b";
warn $str; # "ab" expected, but 5.15.7-9 gives "\0b"
}->join;
What was happening before commit
b6597275 was that two bugs were can-
celling each other out: $str would be "" when the new thread started,
but with a string buffer containing "a" beyond the end of the string
and $fh remembering 1 as its position. The bug fixed by
b6597275 was
that writing past the end of a string through a filehandle was leaving
junk (whatever was in memory already) in the intervening space between
the old end of string and the beginning of what was being written to
the string. This allowed "" to turn magically into "ab" when "b" was
written one character past the end of the string. Commit
b6597275
started zeroing out the intervening space in that case, causing the
cloning bug to rear its head.
This commit solves the problem by hiding the scalar temporarily
in PerlIOScalar_dup so that PerlIOScalar_pushed won’t be able to
modify it.
Should PerlIOScalar_pushed stop clobbering the string and should
PerlIOScalar_open do it instead? Perhaps. But that would be a bigger
change, and we are supposed to be in code freeze right now.
Father Chrysostomos [Mon, 7 May 2012 21:53:20 +0000 (14:53 -0700)]
Increase $PerlIO::scalar::VERSION to 0.14
Ricardo Signes [Mon, 7 May 2012 15:03:37 +0000 (11:03 -0400)]
with 5.16.0, 5.12.x is security-only
H.Merijn Brand [Sun, 6 May 2012 11:11:03 +0000 (13:11 +0200)]
check for PA* in both branches of case
(thanks ilmari for spotting)
H.Merijn Brand [Sun, 6 May 2012 11:03:08 +0000 (13:03 +0200)]
Disable optimizer for 32bit PA-RISC builds on HP-UX
The (ANSI) C compiler fails to compile precompiled (.i) files when both
-g and -O (all +O1 and above) are given. When -g is requested, -O, +O,
and +Onolimit are removed from optimize flags
This #fail does not occur with the newer aCC compiler B3910B, which is
also used on HP-UX on Itanium.
The check/modification has to be done as late as possible, as the other
options, like -Duse64bitall and -DDEBUGING, will modify the variables
that need to be checked after hints/hpux.sh has been dealt with.
Steve Peters [Fri, 4 May 2012 15:51:06 +0000 (10:51 -0500)]
Add --libpods back as a non-functional option to pod2html.
When --libpods was removed, this broke backward compatiblility with
existing uses. This change adds back the option, but warns that
--libpods is no longer supported.
David Golden [Fri, 4 May 2012 15:02:26 +0000 (11:02 -0400)]
delete PERL_YAML_BACKEND and PERL_JSON_BACKEND in T/TEST
If these are set, Parse-CPAN-Meta and other things that depend
on it may fail.
Paul Johnson [Sun, 29 Apr 2012 18:27:37 +0000 (20:27 +0200)]
Correct variable name in example.
As noticed by Lawrence Statton <lawrence@cluon.com>
Jesse Vincent [Tue, 24 Apr 2012 23:02:34 +0000 (19:02 -0400)]
Bump the version of perl5db since the porting scripts care
Jesse Vincent [Tue, 24 Apr 2012 19:35:39 +0000 (15:35 -0400)]
we no longer have in-file changelogs, since we have a version control system
Jesse Vincent [Tue, 24 Apr 2012 19:05:55 +0000 (15:05 -0400)]
We now have version control and no longer need a changelog in perl5db
Karl Williamson [Fri, 27 Apr 2012 17:09:14 +0000 (11:09 -0600)]
utf8n_to_uvuni(): Fix broken malformation interactions
All code points whose UTF-8 representations start with a byte containing
either \xFE or \xFF are considered problematic because they are not
portable. There are many such code points that are too large to
represent on a 32 or even a 64 bit platform. Commit
eb83ed87110e41de6a4cd4463f75df60798a9243 failed to properly catch
overflow when the input flags to this function say to warn on, but
otherwise accept FE and FF sequences. Now overflow is checked for
unconditionally.
Father Chrysostomos [Fri, 27 Apr 2012 20:31:20 +0000 (13:31 -0700)]
Really increase $File::DosGlob::VERSION to 1.07
I honestly thought I had run the tests, but I suppose not.
Father Chrysostomos [Fri, 27 Apr 2012 16:43:07 +0000 (09:43 -0700)]
Increase $version::VERSION to 0.99
What we have in blead right now matches that CPAN release, so this
version bump *must* happen before 5.16.
Ricardo Signes [Fri, 27 Apr 2012 01:39:33 +0000 (21:39 -0400)]
disable codes_in_verbatim for Pod::Html
...otherwise all our verbatim blocks will change radically!
Karl Williamson [Thu, 19 Apr 2012 04:14:15 +0000 (22:14 -0600)]
is_utf8_char_slow(): Avoid accepting overlongs
There are possible overlong sequences that this function blindly
accepts. Instead of developing the code to figure this out, turn this
function into a wrapper for utf8n_to_uvuni() which already has this
check.
Karl Williamson [Thu, 19 Apr 2012 00:32:57 +0000 (18:32 -0600)]
perlapi: Update for changes in utf8 decoding
Karl Williamson [Mon, 23 Apr 2012 19:28:32 +0000 (13:28 -0600)]
utf8.c: White-space only
This outdents to account for the removal of a surrounding block.
Karl Williamson [Wed, 18 Apr 2012 23:36:01 +0000 (17:36 -0600)]
utf8.c: refactor utf8n_to_uvuni()
The prior version had a number of issues, some of which have been taken
care of in previous commits.
The goal when presented with malformed input is to consume as few bytes
as possible, so as to position the input for the next try to the first
possible byte that could be the beginning of a character. We don't want
to consume too few bytes, so that the next call has us thinking that
what is the middle of a character is really the beginning; nor do we
want to consume too many, so as to skip valid input characters. (This
is forbidden by the Unicode standard because of security
considerations.) The previous code could do both of these under various
circumstances.
In some cases it took as a given that the first byte in a character is
correct, and skipped looking at the rest of the bytes in the sequence.
This is wrong when just that first byte is garbled. We have to look at
all bytes in the expected sequence to make sure it hasn't been
prematurely terminated from what we were led to expect by that first
byte.
Likewise when we get an overflow: we have to keep looking at each byte
in the sequence. It may be that the initial byte was garbled, so that
it appeared that there was going to be overflow, but in reality, the
input was supposed to be a shorter sequence that doesn't overflow. We
want to have an error on that shorter sequence, and advance the pointer
to just beyond it, which is the first position where a valid character
could start.
This fixes a long-standing TODO from an externally supplied utf8 decode
test suite.
And, the old algorithm for finding overflow failed to detect it on some
inputs. This was spotted by Hugo van der Sanden, who suggested the new
algorithm that this commit uses, and which should work in all instances.
For example, on a 32-bit machine, any string beginning with "\xFE" and
having the next byte be either "\x86" or \x87 overflows, but this was
missed by the old algorithm.
Another bug was that the code was careless about what happens when a
malformation occurs that the input flags allow. For example, a sequence
should not start with a continuation byte. If that malformation is
allowed, the code pretended it is a start byte and extracts the "length"
of the sequence from it. But pretending it is a start byte is not the
same thing as it actually being a start byte, and so there is no
extractable length in it, so the number that this code thought was
"length" was bogus.
Yet another bug fixed is that if only the warning subcategories of the
utf8 category were turned on, and not the entire utf8 category itself,
warnings were not raised that should have been.
And yet another change is that given malformed input with warnings
turned off, this function used to return whatever it had computed so
far, which is incomplete or erroneous garbage. This commit changes to
return the REPLACEMENT CHARACTER instead.
Thanks to Hugo van der Sanden for reviewing and finding problems with an
earlier version of these commits
Karl Williamson [Wed, 18 Apr 2012 22:48:29 +0000 (16:48 -0600)]
utf8n_to_uvuni: Avoid reading outside of buffer
Prior to this patch, if the first byte of a UTF-8 sequence indicated
that the sequence occupied n bytes, but the input parameters indicated
that fewer were available, all n were attempted to be read
Karl Williamson [Wed, 18 Apr 2012 22:35:39 +0000 (16:35 -0600)]
utf8.c: Clarify and correct pod
Some of these were spotted by Hugo van der Sanden
Karl Williamson [Wed, 18 Apr 2012 22:20:22 +0000 (16:20 -0600)]
utf8.c: Use macros instead of if..else.. sequence
There are two existing macros that do the job that this longish sequence
does. One, UTF8SKIP(), does an array lookup and is very likely to be in
the machine's cache as it is used ubiquitously when processing UTF-8.
The other is a simple test and shift. These simplify the code and
should speed things up as well.
Karl Williamson [Wed, 18 Apr 2012 21:25:28 +0000 (15:25 -0600)]
utf8.h: Use correct definition of start byte
The previous definition allowed for (illegal) overlongs. The uses of
this macro in the core assume that it is accurate. The inacurracy can
cause such code to fail.
Christian Hansen [Wed, 18 Apr 2012 20:32:16 +0000 (14:32 -0600)]
utf8.h: Use correct UTF-8 downgradeable definition
Previously, the macro changed by this commit would accept overlong
sequences.
This patch was changed by the committer to to include EBCDIC changes;
and in the non-EBCDIC case, to save a test, by using a mask instead, in
keeping with the prior version of the code
Brian Fraser [Sat, 21 Apr 2012 01:09:56 +0000 (22:09 -0300)]
Make unicode label tests use unicode_eval.
A recent change exposed a faulty test, in t/uni/labels.t;
Previously, a downgraded label passed to eval under 'use utf8;'
would've been erroneously considered UTF-8 and the tests
would pass. Now it's correctly reported as illegal UTF-8
unless unicode_eval is in effect.
Father Chrysostomos [Wed, 25 Apr 2012 05:27:04 +0000 (22:27 -0700)]
perlsub: Fix new typo
Since the typo was added during code freeze, I hope I’m not going too
far in fixing it during code freeze. :-)
Father Chrysostomos [Wed, 25 Apr 2012 05:23:34 +0000 (22:23 -0700)]
perl5160delta: The coreargs opcode is undeserving of mention
Father Chrysostomos [Wed, 25 Apr 2012 05:22:38 +0000 (22:22 -0700)]
perl5160delta: 2 dots after spaces, please
Father Chrysostomos [Wed, 25 Apr 2012 05:19:47 +0000 (22:19 -0700)]
perl5160delta: Document wrap_op_checker
This was added in
e8570548af4, but was somehow missed from
perl5158delta.
Father Chrysostomos [Wed, 25 Apr 2012 05:18:17 +0000 (22:18 -0700)]
perl5160delta: Tweak punctuation
Father Chrysostomos [Wed, 25 Apr 2012 05:15:13 +0000 (22:15 -0700)]
perl5160delta: More details for C<utf8_to_uv*_buf>
I had a note to myself to make sure these were added. I don’t
remember why they were not already there. Perhaps the changes
were after 5.15.9.
Father Chrysostomos [Wed, 25 Apr 2012 05:08:16 +0000 (22:08 -0700)]
perl5160delta: Add missing C<>
Ricardo Signes [Wed, 25 Apr 2012 02:51:06 +0000 (22:51 -0400)]
C<> is not L<> and does not take two |-delim parts
reported by Tom Christiansen
Ricardo Signes [Wed, 25 Apr 2012 02:44:23 +0000 (22:44 -0400)]
minor unicode doc tweaks
reported by Tom Christiansen
Ricardo Signes [Wed, 25 Apr 2012 02:36:03 +0000 (22:36 -0400)]
fix some typos in perlsyn.pod
(both reported by Tom Christiansen)
Ricardo Signes [Wed, 25 Apr 2012 02:34:32 +0000 (22:34 -0400)]
point out __SUB__ in perlsub.pod
Ricardo Signes [Wed, 25 Apr 2012 02:23:18 +0000 (22:23 -0400)]
fix mode used to open /dev/tty in perlop example
(Thanks for reporting this, Tom Christiansen!)
Ricardo Signes [Wed, 25 Apr 2012 00:51:22 +0000 (20:51 -0400)]
perldelta: reimport the 5.15.x internals sections
Ricardo Signes [Wed, 25 Apr 2012 00:46:38 +0000 (20:46 -0400)]
perldelta: reimport the 5.15.x config/compile sections
this catches a few things that were omitted; the reordering
that resulted is neither intentional nor (as far as I can tell)
significant
Ricardo Signes [Wed, 25 Apr 2012 00:38:30 +0000 (20:38 -0400)]
perldelta: add note about HP-UX gcc support
Nicholas Clark [Tue, 24 Apr 2012 09:02:15 +0000 (11:02 +0200)]
Remove outdated todo items, and update and clarify others.
Nicholas Clark [Fri, 20 Apr 2012 08:51:54 +0000 (10:51 +0200)]
In the todo list, note who is already working on various suggestions.
Nicholas Clark [Fri, 20 Apr 2012 08:47:24 +0000 (10:47 +0200)]
Remove (?{...}) closures in regexps from the todo list.
Dave has this substantially done, and is working on finishing it off.
Nicholas Clark [Wed, 18 Apr 2012 19:06:21 +0000 (21:06 +0200)]
Remove "Properly Unicode safe tokeniser and pads" from the todo
Brian Fraser (during GSoC), Father Chrysostomos, Zefram and others have done
most of the tasks suggested in this todo, so remove it from consideration as
it's clearly stale. The only possible work outstanding is how toke.c behaves
when PL_rsfp is marked as returning UTF-8, and when source filters return
SVs with SvUTF8() true. So leave this in as a comment in the Pod so that it
does not get lost, pending proper investigation.
Nicholas Clark [Fri, 30 Mar 2012 19:50:28 +0000 (21:50 +0200)]
Remove 2 microperl suggestions from the todo list.
The --regen option added to Porting/checkcfgvar.pl makes creating uconfig.sh
pretty much automatic.
Whilst maintaining Makefile.micro could be made more automatic by templating
it based on MANIFEST, the benefits are probably never going to pay back the
cost of doing it, as we add (or remove) C and header files so rarely.
It would only really make sense as part of a system that did the same for
Makefile.SH, vms/descrip_mms.template and the multitude of Win* Makefiles.
Nicholas Clark [Fri, 30 Mar 2012 19:35:23 +0000 (21:35 +0200)]
Drop the iCOW plan from the todo.
Dave has detailed reasons why he doesn't think that this would work. So it's
a bad idea to suggest it as something people should work on.
Nicholas Clark [Fri, 30 Mar 2012 19:23:25 +0000 (21:23 +0200)]
Switching entersub ops from "XS" to "Perl" and back likely won't help.
Switching OPs goes against the grain of immutable optrees under ithreads -
it could probably be made to work, but the amount of extra effort would be
non-trivial.
Secondly, cache misses are the overriding speed hit on modern CPUs.
The code switching proposed would add complexity but only be able to change
the number of branch instructions encountered, which come a distant second
to cache misses as a speed hit. The code switching would not change the data
lookups needed, and hence not the cache misses they cause.