Chris 'BinGOs' Williams [Wed, 20 Nov 2013 20:51:29 +0000 (20:51 +0000)]
Merge branch 'bingos/perl-5.19.6' into blead
Father Chrysostomos [Wed, 20 Nov 2013 14:03:09 +0000 (06:03 -0800)]
t/op/sub.t: Suppress warnings properly
We enable lexical warnings further up in the file, so $^W won’t do
anything to suppress the warnings from the glob assignment, only
those from XSLoader.
Chris 'BinGOs' Williams [Wed, 20 Nov 2013 14:53:31 +0000 (14:53 +0000)]
Add new release to perlhist
Chris 'BinGOs' Williams [Wed, 20 Nov 2013 14:35:59 +0000 (14:35 +0000)]
Finalise perldelta
Chris 'BinGOs' Williams [Wed, 20 Nov 2013 14:32:12 +0000 (14:32 +0000)]
Update Module-CoreList for v5.19.6
Chris 'BinGOs' Williams [Wed, 20 Nov 2013 12:48:25 +0000 (12:48 +0000)]
Release engineering work on perldelta
Daniel Dragan [Tue, 19 Nov 2013 21:25:51 +0000 (16:25 -0500)]
add Intel C++ Compiler for Win32 support
-most fixes involve code detecting Perl on VC, and changing it to ICC
is another kind of VC, but ICC's version isn't this "other kind of VC"'s
version number, call the partner VC to find out the "VC version number"
of ICC
not yet done
-no Intel C specific optimization flags
-long doubles and C99
-ccversion behavior might not be ideal/rethink
Tony Cook [Wed, 20 Nov 2013 04:11:15 +0000 (15:11 +1100)]
[perl #120043] fix some warnings
with an update for perldiag
Tony Cook [Wed, 20 Nov 2013 04:09:44 +0000 (15:09 +1100)]
update perldiag for the change in conversion specifiers
Lukas Mai [Sat, 28 Sep 2013 22:12:44 +0000 (00:12 +0200)]
fix a few warnings (format strings, unused variable)
During compilation gcc complains about the following:
perl.c:4970: warning: format '%u' expects argument of type 'unsigned int', but argument 2 has type 'U32' [-Wformat=]
perl.c:5075: warning: format '%u' expects argument of type 'unsigned int', but argument 2 has type 'I32' [-Wformat=]
mg.c:1972: warning: format '%ld' expects argument of type 'long int', but argument 2 has type 'ssize_t' [-Wformat=]
pp_ctl.c:2610: warning: unused variable 'mark' [-Wunused-variable]
regexec.c:2275: warning: format '%ld' expects argument of type 'long int', but argument 3 has type 'int' [-Wformat=]
This patch fixes all of them.
Tony: warning: unused variable 'mark' was fixed in
481c819b
Daniel Dragan [Wed, 13 Nov 2013 04:07:16 +0000 (23:07 -0500)]
remove PL_patchlevel from S_minus_v
|SvPVX(vstringify(PL_patchlevel))| is the same string as
|"v" PERL_VERSION_STRING|. No need to go through the overhead of using a
version object. Instead of creating a SV, then further manipulating it.
Create and manipulate it at the same time with Perl_newSVpvf_nocontext or
newSVpvn. "num_len >= level_len " will be folded away by the compiler,
previously, since both lengths came from SVs, they were not const
technically. A very smart compiler might see strncmp/strnEQ takes all
const arguments, and will optimize that away also. VC 2003 didnt do that.
Change SvREFCNT_dec to SvREFCNT_dec_NN for obvious reasons.
There should not be any user visible changes to -v with this patch.
switches.t contains a regexp for -v already so no further tests were added.
This patch is part of #116296.
Daniel Dragan [Tue, 19 Nov 2013 01:10:04 +0000 (20:10 -0500)]
regcomp.c extern -> EXTERN_C
Otherwise re wont build on Win32 VC6 C++ mode.
Daniel Dragan [Tue, 19 Nov 2013 04:26:05 +0000 (23:26 -0500)]
ioctl on perlhost platforms take a char*, not void*
As of commit
0cb9638729 IOCtl from perlhost is prototyped to take a char *.
In the later commit
2986a63f7e5 the netware commit is introduced that adds
a void* cast without explination, but today PerlLIOIOCtl is prototyped as
char* in nwperlhost.h (nwperlhost.h didnt exist at
2986a63f7e5 ). In
commit
6e22d04617 Win32 starts to use the void * cast from netware.
Using a void * cast breaks a VC 2003 C++ mode build of IO.xs because of
different types. Switch to a char * cast that matches the vtable prototype.
Daniel Dragan [Tue, 19 Nov 2013 20:37:58 +0000 (15:37 -0500)]
better perldelta description for #120091/#118059
Father Chrysostomos [Tue, 19 Nov 2013 05:53:43 +0000 (21:53 -0800)]
Move <-- HERE arrow for ‘Switch condition not recognized’
$ ./perl -Ilib -e '/(?(1(?#...)))/'
Switch condition not recognized in regex; marked by <-- HERE in m/(?(1( <-- HERE ?#...)))/ at -e line 1.
$ ./perl -Ilib -e '/(?(1x(?#...)))/'
Switch condition not recognized in regex; marked by <-- HERE in m/(?(1x(?#...) <-- HERE ))/ at -e line 1.
With the first one-liner, the arrow in the error message is pointing
to the first offending character.
With the second one-liner, the arrow points to the comment following
the offending character.
The logic for positioning the character is a little odd. The idea is
supposed to be something like:
if current_character++ is not ')'
croak with the arrow right before current_character
But nextchar() is used instead of ++, and nextchar() skips trailing
whitespace and comments after incrementing the current parse position.
We already have code right here to revert back to the previous parse
position and then increment it by one character, for the sake of UTF8.
Indeed, it behaves differently if you add a non-ASCII character under
‘use utf8’:
$ ./perl -Ilib -e 'use utf8; /é(?(1x(?#...)))/'
Switch condition not recognized in regex; marked by <-- HERE in m/?(?(1x <-- HERE (?#...)))/ at -e line 1.
So what this commit does is extend that backtrack logic to happen all
the time, not just with UTF8.
Father Chrysostomos [Tue, 19 Nov 2013 05:30:45 +0000 (21:30 -0800)]
perldiag: Update descr for ‘Switch condition not recognized’
Originally, the message ‘Switch (?(number%c not recognized’ only
applied to /(?(1junk))/. perl-5.6.0-657-gb45f050 changed it to read
‘Switch condition not recognized’.
perl-5.8.0-8771-g0a4db38 added (?(<...>)), (?('...')), (?(DEFINE)) and
(?(R...)). All of these can trigger that messages if there is junk
before the first closing parenthesis (e.g., /(?(<name>junk))/, causing
a mismatch between the cases where the error occurs and the descrip-
tion in perldiag:
(F) If the argument to the (?(...)if-clause|else-clause) con-
struct is a number, it can be only a number. The <-- HERE shows
whereabouts in the regular expression the problem was discovered.
See perlre.
(which itself was already confusingly worded).
Whether these should be changed to use the more standard ‘Unknown
switch condition’ message I don’t want to deal with right now.
In any case, the description is only sometimes relevant, so this com-
mit just copies the description from ‘Unknown switch condition’.
Father Chrysostomos [Tue, 19 Nov 2013 14:26:19 +0000 (14:26 +0000)]
perldelta for
7583957163
Signed-off-by: Chris 'BinGOs' Williams <chris@bingosnet.co.uk>
Chris 'BinGOs' Williams [Tue, 19 Nov 2013 13:29:25 +0000 (13:29 +0000)]
Update Module-Build to CPAN version 0.4202
[DELTA]
0.4202 - Tue Nov 19 12:48:19 CET 2013
[BUG FIXES]
- Don't merge prereqs from meta to mymeta [Leon Timmermans]
Steve Hay [Tue, 19 Nov 2013 12:21:56 +0000 (12:21 +0000)]
Tony Cook [Tue, 19 Nov 2013 05:45:13 +0000 (16:45 +1100)]
S_already_defined no longer uses its gv parameter, remove it
Tony Cook [Tue, 19 Nov 2013 05:01:29 +0000 (16:01 +1100)]
Matthew Horsfall (alh) [Tue, 8 Oct 2013 16:56:08 +0000 (12:56 -0400)]
Optimise 'if ($a || $b)' and 'unless ($a && $b)' early exit
An OP_AND/OP_OR in void context provides a short circuit
through ->op_other that can be used if AND/OR ops contained
within it jump out early. Use that short circuit.
Previously:
$ ./perl -Ilib -MO=Concise -e 'if ($aa || $bb) {}'
8 <@> leave[1 ref] vKP/REFC ->(end)
1 <0> enter ->2
2 <;> nextstate(main 3 -e:1) v:{ ->3
- <1> null vK/1 ->8
6 <|> and(other->7) vK/1 ->8
- <1> null sK/1 ->6
4 <|> or(other->5) sK/1 ->6 <-- Not optimised
- <1> ex-rv2sv sK/1 ->4
3 <$> gvsv(*aa) s ->4
- <1> ex-rv2sv sK/1 ->-
5 <$> gvsv(*bb) s ->6
- <@> scope vK ->-
7 <0> stub v ->8
Now:
$ ./perl -Ilib -MO=Concise -e 'if ($aa || $bb) {}'
8 <@> leave[1 ref] vKP/REFC ->(end)
1 <0> enter ->2
2 <;> nextstate(main 3 -e:1) v:{ ->3
- <1> null vK/1 ->8
6 <|> and(other->7) vK/1 ->8
- <1> null sK/1 ->6
4 <|> or(other->5) sK/1 ->7 <-- Short circuited
- <1> ex-rv2sv sK/1 ->4
3 <$> gvsv(*aa) s ->4
- <1> ex-rv2sv sK/1 ->-
5 <$> gvsv(*bb) s ->6
- <@> scope vK ->-
7 <0> stub v ->8
Chris 'BinGOs' Williams [Tue, 19 Nov 2013 00:13:41 +0000 (00:13 +0000)]
Update Module-Build to CPAN version 0.4201
[DELTA]
0.4201 - Mon Nov 18 23:23:25 CET 2013
[BUG FIXES]
- Prefer META.json over META.yml [Leon Timmermans]
Victor Efimov [Tue, 19 Nov 2013 00:18:26 +0000 (01:18 +0100)]
Check symlink status before setting File::Find::fullname to undef.
Problem reported by James Avera in RT #120388. Patches supplied by Victor
Efimov, then adapted to new testing functions in ext/File-Find/t/find.t.
Tony Cook [Mon, 18 Nov 2013 22:50:25 +0000 (09:50 +1100)]
[perl #120543] work around a C++ library bug in VC 2003
VC 2003 makes overloaded versions of ldexp() available when including
<math.h> which should only be visible when including <cmath>.
The cast ensures the ldexp(double, int) signature is used.
Chris 'BinGOs' Williams [Mon, 18 Nov 2013 19:16:26 +0000 (19:16 +0000)]
Update HTTP-Tiny to CPAN version 0.038
[DELTA]
0.038 2013-11-18 12:56:26 America/New_York
[FIXED]
- Fixed a bug where authentication parameters in the URL would override
an existing Authorization header
David Mitchell [Mon, 18 Nov 2013 16:48:48 +0000 (16:48 +0000)]
XS::Typemap: silence compiler warning.
xsubpp will give us a RETVAL var whether we use it not.
David Mitchell [Mon, 18 Nov 2013 15:41:45 +0000 (15:41 +0000)]
threads: silence some compiler warnings.
From 5.8.9, the MGVTBL struct has a 'svt_local' member
Declare some vars volatile within a function that uses jongjmp
Mark the unused return value from PerlLIO_write() as void(), even
though it won't shut up gcc.
David Mitchell [Mon, 18 Nov 2013 15:24:04 +0000 (15:24 +0000)]
POSIX: silence some compiler warnings
David Mitchell [Mon, 18 Nov 2013 15:07:23 +0000 (15:07 +0000)]
SDBM_File: fix 'set but not used' warning
the NEXTKEY() method has a key arg, but sdbm_nextkey() doesn't.
So don't bother mapping the key SV arg to a datum_key type.
David Mitchell [Mon, 18 Nov 2013 14:52:57 +0000 (14:52 +0000)]
PerlIO::via: fix compiler warning
In
GV *gv = newGVgen(HvNAME_get(s->stash))
HvNAME_get() can return a null value (don't know whether it ever will in
these circumstances), while newGVgen() expects a non-null arg.
So calculate HvNAME_get() first, and bail if it's null.
David Mitchell [Mon, 18 Nov 2013 14:37:22 +0000 (14:37 +0000)]
PerlIO::scalar: silence some compiler warnings
Father Chrysostomos [Mon, 18 Nov 2013 14:19:33 +0000 (06:19 -0800)]
perlhacktips: Two spaces after dots
It was not consistent throughout.
Father Chrysostomos [Mon, 18 Nov 2013 14:12:03 +0000 (06:12 -0800)]
perlhacktips: Fix pod formatting
Matthew Horsfall (alh) [Mon, 30 Sep 2013 17:23:22 +0000 (13:23 -0400)]
Add more examples of perl/gdb usage.
Also add doc about gdb's ptype, and make examples more clear.
Father Chrysostomos [Mon, 18 Nov 2013 14:01:56 +0000 (06:01 -0800)]
Fix ‘panic: memory wrap’ in reg_scan_name
reg_scan_name was not checking for end-of-string. If the character it
read were not a word character, it would then increment the current
position (RExC_parse), so that the <-- HERE marker in the error mes-
sage would point to the bad character.
If we try to split a regexp like /(?</ into two pieces when the cur-
rent position is off the end like this:
( ? < \0
^
then the first ‘half’ of the regexp, before the <-- HERE marker is
"(?<\0" (including the trailing null), and the second ‘half’ is of
negative length. Negative string lengths are what cause ‘panic: mem-
ory wrap’.
$ ./perl -Ilib -e '/(?</'
panic: memory wrap at -e line 1.
This commit takes advantage of the fact that, ever since
1f4f6bf1,
RExC_parse == name_start has never been true after a call to
reg_scan_name. This is how reg_scan_name now signals EOS.
Father Chrysostomos [Mon, 18 Nov 2013 13:57:50 +0000 (05:57 -0800)]
perldiag: Clarify memory wrap
Father Chrysostomos [Mon, 18 Nov 2013 13:45:19 +0000 (05:45 -0800)]
Require closing paren for (?&...
$ ./perl -Ilib -le 'print "aaz" =~ /(?<a>a)(?&a-z/'
a
(?&name) was actually allowing any non-word character to terminate it.
Chris 'BinGOs' Williams [Mon, 18 Nov 2013 15:13:01 +0000 (15:13 +0000)]
Update IPC-Cmd to CPAN version 0.90
[DELTA]
0.90 Mon Nov 18 15:08:15 GMT 2013
Misc:
* skip some problematic tests when PERL_CORE
Tony Cook [Mon, 18 Nov 2013 06:01:07 +0000 (17:01 +1100)]
Tony Cook [Mon, 18 Nov 2013 05:53:46 +0000 (16:53 +1100)]
Tony Cook [Mon, 18 Nov 2013 05:44:40 +0000 (16:44 +1100)]
[perl #120535] Add UTF8 flag to B::HV->ARRAY keys
Tony Cook [Mon, 18 Nov 2013 04:20:16 +0000 (15:20 +1100)]
un-TODO the test for 120535, Reini's patch fixes it
Reini Urban [Wed, 13 Nov 2013 19:18:09 +0000 (13:18 -0600)]
Add UTF8 flag to keys returned by B::HV->ARRAY
hv_iternextsv() ignores the HEK flags, but we do care about the UTF8 flag
at least when returning hash keys from B::HV->ARRAY
Tony Cook [Mon, 18 Nov 2013 04:16:31 +0000 (15:16 +1100)]
TODO test for 120535
Tony Cook [Mon, 18 Nov 2013 00:00:17 +0000 (11:00 +1100)]
fix expected failure text for new croak.t test
Father Chrysostomos [Sun, 17 Nov 2013 21:09:26 +0000 (13:09 -0800)]
[Merge] perldiag stuff
I noticed that some entries in perldiag had (F ...) with warnings
categories. Those categories apply only to warnings, not to
fatal errors.
So I looked to see why diag.t was not already catching it, since I
thought I had programmed it to.
It turned out that it was skipping any function calls that lacked
aTHX_, which included yyerror and anything with _nocontext.
So I fixed that and got pages and pages of failures, unsurprisingly.
I went through, fixing them, and inevitably tweaking other things.
In one case, I even changed the message output, removing an extraneous
dot before ‘at file line 24’.
Father Chrysostomos [Sun, 17 Nov 2013 19:45:33 +0000 (11:45 -0800)]
perldiag: Don’t let <-- HERE be wrapped
The meaning is much clearer if the <--
HERE marker doesn’t wrap like that.
Father Chrysostomos [Sun, 17 Nov 2013 14:44:05 +0000 (06:44 -0800)]
perldiag: Remove categories from fatal errors
These are *warnings* categories, so they don’t apply here.
Father Chrysostomos [Sun, 17 Nov 2013 14:42:45 +0000 (06:42 -0800)]
diag.t: Fix copy-and-paste error in os2 exception
Father Chrysostomos [Sun, 17 Nov 2013 14:40:14 +0000 (06:40 -0800)]
diag.t: Don’t check ‘calls’ in column 0
because they may be function declarations instead. It was getting
confused by win32_croak_not_implemented, which does not have ");"
in it before the first string.
Father Chrysostomos [Sun, 17 Nov 2013 14:39:36 +0000 (06:39 -0800)]
diag.t: More win32 exceptions
Father Chrysostomos [Sun, 17 Nov 2013 14:32:02 +0000 (06:32 -0800)]
diag.t: Support multiline perldiag entries better
I.e., don’t require a space before the newline.
Father Chrysostomos [Sun, 17 Nov 2013 14:29:37 +0000 (06:29 -0800)]
diag.t: Let diag.t be run on specific source files
./perl -Ilib t/porting/diag.t regcomp.c
will now run it just on regcomp.c.
Father Chrysostomos [Sun, 17 Nov 2013 14:23:07 +0000 (06:23 -0800)]
diag.t: Another exception
Father Chrysostomos [Sun, 17 Nov 2013 14:22:24 +0000 (06:22 -0800)]
diag.t: Skip x2p
Father Chrysostomos [Sun, 17 Nov 2013 14:22:07 +0000 (06:22 -0800)]
diag.t: win32 exceptions
Father Chrysostomos [Sun, 17 Nov 2013 14:18:12 +0000 (06:18 -0800)]
diag.t: Support win32_croak_not_implemented
Father Chrysostomos [Sun, 17 Nov 2013 14:17:58 +0000 (06:17 -0800)]
diag.t: Another exception
Father Chrysostomos [Sun, 17 Nov 2013 14:15:08 +0000 (06:15 -0800)]
Sprinkle diag_listed_as
Father Chrysostomos [Sun, 17 Nov 2013 14:10:40 +0000 (06:10 -0800)]
Remove extraneous dot from ‘Missing right brace...’
$ ./perl -Ilib -e '/\N{/'
Missing right brace on \N{} or unescaped left brace after \N. at -e line 1, within pattern
Execution of -e aborted due to compilation errors.
That dot before ‘at -e’ is confusing.
Father Chrysostomos [Sun, 17 Nov 2013 14:03:57 +0000 (06:03 -0800)]
perldiag: Clarify an entry
‘5.12’ is meaningful to insiders, but it may not be understood by
readers in general.
Father Chrysostomos [Sun, 17 Nov 2013 13:58:08 +0000 (05:58 -0800)]
perldiag: Document ‘\N{} in character class restricted...’
Father Chrysostomos [Sun, 17 Nov 2013 13:46:15 +0000 (05:46 -0800)]
perldiag: Document ‘Zero length \N{}’
Father Chrysostomos [Sun, 17 Nov 2013 13:38:07 +0000 (05:38 -0800)]
perldiag: Miscapitalisation; clarity
Father Chrysostomos [Sun, 17 Nov 2013 13:32:45 +0000 (05:32 -0800)]
perldiag: Clean up ‘Sequence (...’ messages
Make sure they are all documented. Sprinkle diag_listed_as where
appropriate.
Father Chrysostomos [Sun, 17 Nov 2013 06:25:32 +0000 (22:25 -0800)]
perldiag: Reword an entry slightly
to make it easier to understand
Father Chrysostomos [Sun, 17 Nov 2013 06:09:48 +0000 (22:09 -0800)]
perldiag: Add some ‘in regex’ variants
Father Chrysostomos [Sun, 17 Nov 2013 06:02:16 +0000 (22:02 -0800)]
diag.t: Fix FAIL and vFAIL support
Until recently, it was treating them like Perl_croak et al., expecting
there to be an aTHX_.
Now that I fixed that, it turns out that FAIL and vFAIL (which were
being ignored, since all the calls lacked aTHX_) never worked anyway.
If they are matched by $source_msg_call_re, they end up in the wrong
if/else branch; namely, the one for handling packWARN, and they skip
over the one for tacking ‘in regex’ on to the end of the message.
Father Chrysostomos [Sun, 17 Nov 2013 05:49:06 +0000 (21:49 -0800)]
diag.t: Exceptions from plan9.c
Father Chrysostomos [Sun, 17 Nov 2013 05:45:30 +0000 (21:45 -0800)]
perldiag: Document ‘Don't know how to get file name’
Father Chrysostomos [Sun, 17 Nov 2013 05:35:19 +0000 (21:35 -0800)]
perldiag: Update Assertion to match reality
Father Chrysostomos [Sun, 17 Nov 2013 02:15:59 +0000 (18:15 -0800)]
diag.t: Exceptions, mostly from os2.c
Father Chrysostomos [Sat, 16 Nov 2013 23:59:23 +0000 (15:59 -0800)]
diag.t: gv_override is not a croak function
Father Chrysostomos [Sat, 16 Nov 2013 23:58:20 +0000 (15:58 -0800)]
op.c: diag_listed_as for Using !~ with %s
Father Chrysostomos [Sat, 16 Nov 2013 23:55:30 +0000 (15:55 -0800)]
Stop diag.t from skipping yyerror messages
yyerror does not take aTHX_, but diag.t was expecting aTHX_ on all
warn/croak calls.
Yes, this causes massive failures.
Father Chrysostomos [Sat, 16 Nov 2013 23:12:58 +0000 (15:12 -0800)]
perldiag: \%d better written as $%d
splain can’t necessarily find it if we hard-code the number:
$ ./perl -Ilib -Mdiagnostics -we '()="" =~ s/()()/\1/r'
\1 better written as $1 at -e line 1 (#1)
(W syntax) Outside of patterns, backreferences live on as variables.
The use of backslashes is grandfathered on the right-hand side of a
substitution, but stylistically it's better to use the variable form
because other Perl programmers will expect it, and it works better if
there are more than 9 backreferences.
$ ./perl -Ilib -Mdiagnostics -we '()="" =~ s/()()/\2/r'
\2 better written as $2 at -e line 1.
$
%d is more intuitive than %c here, so I used %d in the pod instead of
copying the C code that generates it.
Father Chrysostomos [Sat, 16 Nov 2013 23:07:51 +0000 (15:07 -0800)]
perldiag: rewrap an entry for better splain output
James E Keenan [Sun, 17 Nov 2013 15:58:07 +0000 (16:58 +0100)]
Fix POD formatting error.
James E Keenan [Sun, 17 Nov 2013 14:27:14 +0000 (15:27 +0100)]
perldelta addition for the previous commit.
James E Keenan [Mon, 11 Nov 2013 20:11:14 +0000 (21:11 +0100)]
Convert to using Test::More. Provide each test with a description.
For: RT #120503. This entailed:
Replacing each instance of Check() and CheckDie() with appropriate
Test::More functions.
Replacing MkDir() and touch() with mkdir_ok() and create_file_ok() to
create a more Test::More-ish interface; create symlink_ok() to handle
similar circumstances.
Providing those three wrapper subroutines with default descriptions,
but allowing a custom description to be passed as an argument.
Highlighting more important unit tests found at the end of subsections
and blocks by prepending 'COMPLETE:' to their descriptions.
More inline comments to guide future maintainers.
Some repositioning of code in order to narrow the scope of variables.
Replacement of hard tabs with spaces; some regularization of indents
and other code tidying.
No tests were added or removed, so the planned number of tests on a
given platform will be unchanged.
TODO:
Some descriptions still need improvement.
There are quite a few File-Find features mentioned in the POD which
are not referenced at all in this test file. They need to have new
tests written for them.
Coverage analysis.
Father Chrysostomos [Sat, 16 Nov 2013 14:52:26 +0000 (06:52 -0800)]
perldelta for the previous commit
Father Chrysostomos [Sat, 16 Nov 2013 14:50:49 +0000 (06:50 -0800)]
Fix memory leak with recursive sub redefinition
See the thread starting at
<
20131115042923.7514.qmail@lists-nntp.develooper.com>.
Commits
7004ee4937 and
a61818571 changed subroutine redefinition to
null the GvCV slot before freeing the CV, so that destructors won’t
see GvCV without a reference count. That turns a double free into a
memory leak.
Kent Fredric explains it nice and clearly:
> sub foo{} # A
> bless \&foo;
> DESTROY { *foo = sub {} # C }
> eval "sub foo{} "; # B
>
> Previous behaviour was:
>
> B replaces A, triggers DESTROY, which triggers C replacing A, and this
> invoked a double free, because C , triggering the removal of A,
> happened while A still existed ( ?? )
>
> So the change fixes this, so that A is removed from the symbol table
> before DESTROY triggers , so that C is creating a "new" symbol,
> effectively, and the problem is that C is then clobbered by the B
> replacing the slot, after the DESTROY.
So C leaks.
This commit fixes it by changing the SvREFCNT_dec into SAVEFREESV,
essentially delaying the DESTROY until after the subroutine redefini-
tion is complete.
This does mean that C is what ends up in the glob afterwards; but as
long as perl’s own bookkeeping is thrown off, we can leave it to the
user (the Perl programmer) to handle the consequences of naughty
destructors.
Father Chrysostomos [Sat, 16 Nov 2013 13:18:59 +0000 (05:18 -0800)]
gv.c: Remove redundant mro_method_changed_in()
Core subs are already accounted for by method lookup, so vivifying
them does not change what gets inherited. Hence, there is no need
to flush caches here.
Father Chrysostomos [Sat, 16 Nov 2013 05:40:45 +0000 (21:40 -0800)]
Unwrap a perldiag entry for better splain output
Before:
$ ./perl -Ilib -Mdiagnostics -e 'warn "Missing '\'']'\'' in prototype for %s : %s"'
Missing ']' in prototype for %s : %s at -e line 1 (#1)
(W illegalproto) A grouping was started with [ but never closed with
].
After:
$ ./perl -Ilib -Mdiagnostics -e 'warn "Missing '\'']'\'' in prototype for %s : %s"'
Missing ']' in prototype for %s : %s at -e line 1 (#1)
(W illegalproto) A grouping was started with [ but never closed with ].
Father Chrysostomos [Sat, 16 Nov 2013 05:38:43 +0000 (21:38 -0800)]
perldiag: ‘my sub’ is no longer ‘reserved’ syntax
Father Chrysostomos [Sat, 16 Nov 2013 05:35:04 +0000 (21:35 -0800)]
Alphabetise perldiag
Father Chrysostomos [Sat, 16 Nov 2013 05:35:46 +0000 (21:35 -0800)]
perldiag: remove space from otherwise blank line
This enables splain to find the following entry.
David Mitchell [Sat, 16 Nov 2013 00:01:32 +0000 (00:01 +0000)]
Revert "make perl core quiet under -Wfloat-equal"
A suggested way of avoiding the the warning on nv1 != nv2
by replacing it with (nv1 < nv2 || nv1 > nv2), has too many issues
with NaN. [perl #120538].
I haven't found any other way of selectively disabling the warning,
so for now I'm just reverting the whole commit.
This reverts commit
c279c4550ce59702722d0921739b1a1b92701b0d.
Chris 'BinGOs' Williams [Fri, 15 Nov 2013 23:03:14 +0000 (23:03 +0000)]
Update DB_File to CPAN version 1.831
[DELTA]
1.831 15 November 2013
* C99 comment is a nogo
RT #90383
Chris 'BinGOs' Williams [Fri, 15 Nov 2013 14:56:16 +0000 (14:56 +0000)]
Update IPC-Cmd to CPAN version 0.88
[DELTA]
0.88 Fri Nov 15 14:44:03 GMT 2013
Misc:
* skip some problematic tests on Cygwin
Daniel Dragan [Fri, 15 Nov 2013 08:21:53 +0000 (03:21 -0500)]
[perl #120544] fix "goto" around var init error in av_extend_guts
This fixes a Win32 Visual C 6 in C++ mode syntax error. See #120544 for
the details.
Father Chrysostomos [Fri, 15 Nov 2013 13:31:22 +0000 (05:31 -0800)]
Father Chrysostomos [Fri, 15 Nov 2013 13:28:24 +0000 (05:28 -0800)]
perldelta for
c5f75dbad5e3
Father Chrysostomos [Fri, 15 Nov 2013 13:04:39 +0000 (05:04 -0800)]
Increase $fields::VERSION to 2.17
Father Chrysostomos [Fri, 15 Nov 2013 06:13:45 +0000 (22:13 -0800)]
Revamp fields documentation
• Actually give an example of a typed lexical.
• Also show the syntax parenthetically in the pod, since the phrase
‘typed lexical’ is not exactly well known.
• Don’t use dev version numbers.
• Don’t talk about pseudo-hashes so pervasively, as the fact that they
are used is more an implementation detail now than a feature (except
for phash).
• If we talk about restricted hashes, we should link to Hash::Util.
• -w is not the only way to turn warnings on.
Also mention in perlfunc that ‘my __PACKAGE__’ and ‘my CONSTANT’ work,
so that the only mentions are not in perl561delta and Lexical::Types.
Father Chrysostomos [Fri, 15 Nov 2013 04:30:30 +0000 (20:30 -0800)]
Father Chrysostomos [Fri, 15 Nov 2013 01:53:02 +0000 (17:53 -0800)]
perldelta for
4571f4a728 and, of course,
bc81b34d0e
4571f4a728 alone made things worse, causing a double free. :-)
Father Chrysostomos [Fri, 15 Nov 2013 01:51:04 +0000 (17:51 -0800)]
Another perldelta to-do
I’m sure Dave Mitchell can explain this better than I.
Father Chrysostomos [Thu, 14 Nov 2013 22:29:51 +0000 (14:29 -0800)]
[perl #120463] s/// and tr/// with wide delimiters
$ perl -Mutf8 -e 's αaαα'
Substitution replacement not terminated at -e line 1.
What is happening is that the first scan goes past the delimiter at
the end of the pattern. Then a single byte is compared (the previous
character against the first byte of the opening delimiter) to see
whether the parser needs to step back one byte before scanning the
second part.
That means you can do the equivalent of s/foo/|bar|g if / is replaced
with a wide character:
$ perl -l -Mutf8 -e '$_ = "a"; s αaα|b|; print'
b
This commit fixes it by giving toke.c:S_scan_str an extra parameter,
so it can tell the callers that need this (scan_subst and scan_trans)
where to start scanning the replacement.