Shlomi Fish [Sun, 30 Sep 2012 12:17:09 +0000 (14:17 +0200)]
[perl5db] Extract a function.
Shlomi Fish [Sun, 30 Sep 2012 11:42:13 +0000 (13:42 +0200)]
[perl5db] Remove a stray declaration.
Shlomi Fish [Sun, 30 Sep 2012 10:37:31 +0000 (12:37 +0200)]
[perl5db-refactor] Extract a method.
Shlomi Fish [Sun, 30 Sep 2012 00:51:07 +0000 (02:51 +0200)]
[perl5db-refactor] Extract _delete_all_breakpoints.
Shlomi Fish [Sun, 30 Sep 2012 00:36:09 +0000 (02:36 +0200)]
[perl5db] Remove leading ampersands in sub calls.
Shlomi Fish [Sun, 30 Sep 2012 00:30:27 +0000 (02:30 +0200)]
More refactoring.
Shlomi Fish [Sat, 29 Sep 2012 23:31:27 +0000 (01:31 +0200)]
[perl5db] Remove ampersand.
Shlomi Fish [Sat, 29 Sep 2012 23:24:07 +0000 (01:24 +0200)]
[perl5db-refactor] Merge two conditionals.
Shlomi Fish [Sat, 29 Sep 2012 23:14:56 +0000 (01:14 +0200)]
Add more.
Shlomi Fish [Sat, 29 Sep 2012 23:09:02 +0000 (01:09 +0200)]
Refactored a loop.
Shlomi Fish [Sat, 29 Sep 2012 22:54:56 +0000 (00:54 +0200)]
[perl5db] Avoid predeclaration.
Shlomi Fish [Sat, 29 Sep 2012 22:46:07 +0000 (00:46 +0200)]
[perl5db] Refactored more.
Shlomi Fish [Sat, 29 Sep 2012 22:26:38 +0000 (00:26 +0200)]
[perl5db-refactor] Remove leading ampersands.
From subroutine calls.
Shlomi Fish [Sat, 29 Sep 2012 22:18:40 +0000 (00:18 +0200)]
[perl5db] refactoring + minor bug-fix.
The regex h\s* matches for every string that starts with "h". It should
match only if h is the only thing so a \z is needed.
Shlomi Fish [Sat, 29 Sep 2012 22:08:22 +0000 (00:08 +0200)]
[perl5db-refactoring] Refactored parse_options.
Shlomi Fish [Sat, 29 Sep 2012 21:42:02 +0000 (23:42 +0200)]
[perl5db] Convert $| to ->autoflush(1).
Shlomi Fish [Sat, 29 Sep 2012 21:33:59 +0000 (23:33 +0200)]
[perl5db] Refactor the LineInfo function.
Shlomi Fish [Sat, 29 Sep 2012 21:18:01 +0000 (23:18 +0200)]
[perl5db-refactoring] Got rid of trailing whitespace.
Shlomi Fish [Sat, 29 Sep 2012 21:01:25 +0000 (23:01 +0200)]
[perl5db-Refactor] Convert $_ to a lexical.
Shlomi Fish [Sat, 29 Sep 2012 20:53:00 +0000 (22:53 +0200)]
[perl5db] Extract a function.
Shlomi Fish [Sat, 29 Sep 2012 20:42:30 +0000 (22:42 +0200)]
[perl5db] Remove some trailing space.
Shlomi Fish [Sat, 29 Sep 2012 20:34:52 +0000 (22:34 +0200)]
[perl5db] More refactoring.
1. Got rid of a trailing block comment because the block is shorter.
2. Now using something more sensible than s/^/perl/.
Shlomi Fish [Sat, 29 Sep 2012 20:21:53 +0000 (22:21 +0200)]
[perl5db] Convert @pods to a hash and moved it out.
This is part of the refactoring.
Shlomi Fish [Sat, 29 Sep 2012 19:29:14 +0000 (21:29 +0200)]
Move away from more "or do { ... }" statements.
Shlomi Fish [Sat, 29 Sep 2012 19:10:05 +0000 (21:10 +0200)]
[perl5db] Refactored do {...}s into ifs.
Shlomi Fish [Sat, 29 Sep 2012 18:33:50 +0000 (20:33 +0200)]
[perl5db] More refactoring.
1. Made the POD better.
2. Converted more && do to ifs.
3. More eq conversions.
4. Captures in lexical variables.
5. Using /ms instead of [\x00-\xFF].
Shlomi Fish [Sat, 29 Sep 2012 17:47:19 +0000 (19:47 +0200)]
More refactoring.
Shlomi Fish [Sat, 29 Sep 2012 17:26:27 +0000 (19:26 +0200)]
Convert some COND() && do to ifs.
Also converted "^" and "$" to the less ambiguous \A and \z and converted
an /^q$/ to the "eq" operator.
Here I am starting to refactor the perl debugger.
Shlomi Fish [Fri, 28 Sep 2012 18:03:12 +0000 (20:03 +0200)]
Add a test for the frame option.
Shlomi Fish [Thu, 27 Sep 2012 18:31:27 +0000 (20:31 +0200)]
[perl5db] o PrintRet=0 in void context.
Shlomi Fish [Thu, 27 Sep 2012 18:25:11 +0000 (20:25 +0200)]
Test o PrintRet=1 in void context.
Shlomi Fish [Thu, 27 Sep 2012 18:13:18 +0000 (20:13 +0200)]
[perl5db] Add another test.
Shlomi Fish [Thu, 27 Sep 2012 18:04:50 +0000 (20:04 +0200)]
[perl5db] Test o PrintRet in list context.
Shlomi Fish [Thu, 27 Sep 2012 17:50:05 +0000 (19:50 +0200)]
[perl5db] add a test for o PrintRet=0.
Shlomi Fish [Thu, 27 Sep 2012 17:40:31 +0000 (19:40 +0200)]
Add a test for o PrintRet.
Shlomi Fish [Tue, 25 Sep 2012 16:53:12 +0000 (18:53 +0200)]
Test o inhibit_exit=0.
Shlomi Fish [Fri, 21 Sep 2012 09:14:35 +0000 (12:14 +0300)]
Fix a bug with o AutoTrace.
There it was.
Shlomi Fish [Fri, 21 Sep 2012 08:54:39 +0000 (11:54 +0300)]
[perl5db] Add a test for t command with sub calls.
Shlomi Fish [Fri, 21 Sep 2012 08:36:00 +0000 (11:36 +0300)]
[perl5db] Test o AutoTrace .
It appears to work fine now. Strange.
Shlomi Fish [Fri, 21 Sep 2012 08:31:29 +0000 (11:31 +0300)]
[perl5db] Add a test for the 't' command.
Apparently, the 'o AutoTrace' command is broken since perl-5.16.x and I
want to get a good output for 't' first before fixing it.
Shlomi Fish [Thu, 20 Sep 2012 14:28:39 +0000 (17:28 +0300)]
[perl5db] Add a test for warnLevel=1.
Shlomi Fish [Wed, 19 Sep 2012 13:33:22 +0000 (16:33 +0300)]
[perl5db] Add test for dieLevel option.
Shlomi Fish [Tue, 18 Sep 2012 14:24:14 +0000 (17:24 +0300)]
[perl5db] Test the recallCommand option.
Shlomi Fish [Sat, 15 Sep 2012 14:18:37 +0000 (17:18 +0300)]
[perl5db] Test the M command.
Shlomi Fish [Sat, 15 Sep 2012 14:10:13 +0000 (17:10 +0300)]
Test more for the m staetement.
Shlomi Fish [Sat, 15 Sep 2012 13:54:24 +0000 (16:54 +0300)]
Test the m command.
Shlomi Fish [Sat, 15 Sep 2012 13:46:35 +0000 (16:46 +0300)]
[perl5db] Test the = command.
Shlomi Fish [Sat, 15 Sep 2012 13:29:55 +0000 (16:29 +0300)]
Correct the test blurbs.
Shlomi Fish [Sat, 15 Sep 2012 12:40:21 +0000 (15:40 +0300)]
Add a test for the H command.
Shlomi Fish [Sat, 15 Sep 2012 12:35:07 +0000 (15:35 +0300)]
[perl5db] Add a test for H -7.
Shlomi Fish [Thu, 13 Sep 2012 14:08:58 +0000 (17:08 +0300)]
[perl5db] Fix source cmd from typeahead.
With a test.
Shlomi Fish [Thu, 13 Sep 2012 13:52:54 +0000 (16:52 +0300)]
Add a test for the source command.
Shlomi Fish [Thu, 13 Sep 2012 11:41:59 +0000 (14:41 +0300)]
Extract a common subroutine.
Shlomi Fish [Thu, 13 Sep 2012 11:34:18 +0000 (14:34 +0300)]
Rename to something more meaningful.
Shlomi Fish [Thu, 13 Sep 2012 11:32:41 +0000 (14:32 +0300)]
Convert to DebugWrap.
Shlomi Fish [Thu, 13 Sep 2012 11:26:55 +0000 (14:26 +0300)]
Convert more to DebugWrap.
Shlomi Fish [Wed, 12 Sep 2012 14:13:32 +0000 (17:13 +0300)]
Convert the 61222 test to DebugWrap.
Shlomi Fish [Wed, 12 Sep 2012 14:08:40 +0000 (17:08 +0300)]
Extract _calc_threads_wrapper.
Shlomi Fish [Wed, 12 Sep 2012 14:06:27 +0000 (17:06 +0300)]
Remove some no-longer used code.
Shlomi Fish [Wed, 12 Sep 2012 14:04:49 +0000 (17:04 +0300)]
Prepend with an underscore.
To make it an internally-used subroutine.
Shlomi Fish [Wed, 12 Sep 2012 14:03:48 +0000 (17:03 +0300)]
Convert another test to DebugWrap.
Shlomi Fish [Wed, 12 Sep 2012 14:01:16 +0000 (17:01 +0300)]
Convert another test to DebugWrap.
Shlomi Fish [Wed, 12 Sep 2012 13:54:30 +0000 (16:54 +0300)]
Convert another test to DebugWrap.
Shlomi Fish [Wed, 12 Sep 2012 13:51:50 +0000 (16:51 +0300)]
Convert another test to DebugWrap.
Shlomi Fish [Wed, 12 Sep 2012 13:29:46 +0000 (16:29 +0300)]
Extract a constructor.
This will help in converting the other instances to DebugWrap.
Shlomi Fish [Wed, 12 Sep 2012 13:08:16 +0000 (16:08 +0300)]
Convert a test to DebugWrap.
Shlomi Fish [Wed, 12 Sep 2012 12:59:12 +0000 (15:59 +0300)]
Convert a perl5db.t test to DebugWrap.
Shlomi Fish [Wed, 12 Sep 2012 10:35:43 +0000 (13:35 +0300)]
Add a test for l -num.
Shlomi Fish [Wed, 12 Sep 2012 09:56:51 +0000 (12:56 +0300)]
Test the ! command.
Shlomi Fish [Wed, 12 Sep 2012 07:56:54 +0000 (10:56 +0300)]
Test the { * command.
Shlomi Fish [Wed, 12 Sep 2012 07:52:12 +0000 (10:52 +0300)]
Convert to a different quoting.
This was done so gvim won't be confused with the bracket-matching.
I'll report the problem to the perl.vim project.
Shlomi Fish [Wed, 12 Sep 2012 07:47:19 +0000 (10:47 +0300)]
Test for the { and { ? commands.
Shlomi Fish [Wed, 12 Sep 2012 07:22:26 +0000 (10:22 +0300)]
Add a test for the < and > commands together.
Daniel Dragan [Mon, 12 Nov 2012 05:04:00 +0000 (00:04 -0500)]
rmv context from Perl_croak_no_modify and Perl_croak_xs_usage
Remove the context/pTHX from Perl_croak_no_modify and Perl_croak_xs_usage.
For croak_no_modify, it now has no parameters (and always has been
no return), and on some compilers will now be optimized to a conditional
jump. For Perl_croak_xs_usage one push asm opcode is removed at the caller.
For both funcs, their footprint in their callers (which probably are hot
code) is smaller, which means a tiny bit more room in the cache. My text
section went from 0xC1A2F to 0xC198F after apply this. Also see
http://www.nntp.perl.org/group/perl.perl5.porters/2012/11/msg195233.html .
David Mitchell [Mon, 12 Nov 2012 10:36:55 +0000 (10:36 +0000)]
PADRANGE: don't assume targs are contiguous
Normally in something like my($x,$y,$z), the 3 targs are contiguous;
however eternal modules like Function::Parameters can break that
assumption, so don't assume it.
Yves Orton [Mon, 12 Nov 2012 08:35:57 +0000 (09:35 +0100)]
fix hash order dependency bug in new Module::Corelist::deprecated_in()
The code was doing
$version= List::Util::min(keys %$versions);
which introduces a hash order dependency as min() returns either the first or last
(i didnt check which) of the lowest items in the list, and considers 5.011000 and
5.011 to be equivalent. Depending on the hash order it would return
either one. Hash randomization revealed this bug immediately.
Changing the function to use List::Util::minstr() eliminates the
dependency.
$version= List::Util::minstr(keys %$versions);
Father Chrysostomos [Mon, 12 Nov 2012 06:16:35 +0000 (22:16 -0800)]
[perl #43077] Make goto &sub leave @_ alone
It is a little tricky, as we have to hang on to @_ while unwinding the
effects of local @_.
Father Chrysostomos [Sun, 11 Nov 2012 08:36:33 +0000 (00:36 -0800)]
leakfinder.pl: Many more exceptions
Father Chrysostomos [Fri, 9 Nov 2012 20:50:59 +0000 (12:50 -0800)]
Env.pm: Fix splice in list context
It was passing an extra argument to splice, causing splice @PATH, 0, 0
to append ":0" to the environment variable.
Father Chrysostomos [Fri, 9 Nov 2012 20:47:39 +0000 (12:47 -0800)]
Increase $Env::VERSION to 1.04
Father Chrysostomos [Fri, 9 Nov 2012 14:14:49 +0000 (06:14 -0800)]
Even more regexp charclass leaks
Karl Williamson [Sun, 11 Nov 2012 22:07:14 +0000 (15:07 -0700)]
toke.c: White-space only; no code changes
This indents a newly formed block, and changes one other line to line up
better.
Karl Williamson [Sun, 11 Nov 2012 21:47:53 +0000 (14:47 -0700)]
toke.c: Don't force \N{} into utf8 unnecessarily
regcomp.c no longer requires everything containing \N{} to be in UTF-8.
I'm not sure of the exact commit; it might even have been in 5.16. But
it was done by the time
86e88272fdabc40e3b168a3cc43af5e86284d01b was
done.
Therefore we can remove the temporary code that forced utf8, and replace
it with code that handles the non-utf8 case.
Note that outside patterns, \N{} still forces utf8. This is so that
Unicode semantics will be imposed on the string it resides in, no matter
how it is used. Patterns have a flag that indicates Unicode semantics,
so don't need to be in utf8.
Karl Williamson [Sun, 11 Nov 2012 21:45:19 +0000 (14:45 -0700)]
toke.c: Remove redundant checks
This code is redundant, as the whole string has been checked earlier for
valid utf8ness, or we assume that the internal Perl handler, charnames,
is correctly returning valid utf8.
Karl Williamson [Sun, 11 Nov 2012 21:28:17 +0000 (14:28 -0700)]
toke.c: Reorder declaration
This is in preparation for the next commit.
Karl Williamson [Thu, 8 Nov 2012 19:02:29 +0000 (12:02 -0700)]
toke.c: Use UTF-8 macros for portability
I noticed this code in toke.c that hand-rolled its own UTF-8 handling,
and is assuming an ASCII platform. There are macros that do the same
thing that this code does, and are portable.
Karl Williamson [Thu, 8 Nov 2012 19:04:34 +0000 (12:04 -0700)]
utf8.h: Add macro that handled malformed 2-byte UTF-8
The macro used traditionally to see if there is a two-byte UTF-8
sequence doesn't make sure that there actually is a second byte
available; it only checks if the first byte indicates that there is.
This adds a macro that is safe in the face of malformed UTF-8.
I inspected the existing calls in the core to the unsafe macro, and I
believe that none of them need to be converted to the safe version.
Karl Williamson [Thu, 8 Nov 2012 18:11:41 +0000 (11:11 -0700)]
util.c: Add some asserts()
These functions improperly have a signed parameter, when the value
should always be at least 0. We have a field report that one of them is
getting called with a negative value. Add asserts to catch this
condition (if compiling assertions is on) earlier.
Karl Williamson [Tue, 30 Oct 2012 16:15:38 +0000 (10:15 -0600)]
utf8.c: Remove redundant unlikely branches
The 2 lines removed in each function provide an early exit if the input
is malformed UTF-8. Other code executed later makes the same test.
But most inputs are going to be well-formed, so the test will almost
always fail, so will slow things up.
Karl Williamson [Fri, 9 Nov 2012 17:18:17 +0000 (10:18 -0700)]
Validate above-Latin1 characters in \N{} aliases
This completes the process of allowing users to define their own aliases
for \N{} in any language they choose. Names have some validation
applied so that they can't, for example, begin with something that is a
digit in some Unicode script. Tests and documentation are included in
this patch. The loop in toke.c that does the validation for
user-supplied translators is revamped, and the messages that are output
when there is an error are fixed to work with UTF-8.
Karl Williamson [Fri, 9 Nov 2012 16:46:37 +0000 (09:46 -0700)]
embed.fnc: Allow toke.c to call core_swash_init()
This internal function is allowed to just a few core functions
by #ifdef's in embed.fnc. Expand the list to include toke.c, as it will
be needed in a future commit.
Karl Williamson [Thu, 1 Nov 2012 18:06:32 +0000 (12:06 -0600)]
toke.c: Fail on malformed UTF-8 in \N{} input
The handler for \N{} can be user-supplied and charnames itself
shouldn't have to worry about malformed input. This changes toke.c to
check for malformed input before calling the \N{} handler.
Karl Williamson [Sun, 28 Oct 2012 16:29:10 +0000 (10:29 -0600)]
toke.c: Remove no-longer needed {...} and outdent correspondingly
No code changes
Brian Fraser [Sun, 28 Oct 2012 16:23:35 +0000 (10:23 -0600)]
toke.c: [RT#73022] Make \N{...} UTF-8 clean.
Karl Williamson [Sun, 28 Oct 2012 15:53:05 +0000 (09:53 -0600)]
toke.c Refactor S_get_and_check_backslash_N_name()
This code was recently factored out into a separate subroutine, and was
originally designed for a non-fatal deprecated warning. This
refactoring just goes immediately to failure when an illegal character
is found.
(It also changes the code to use Perl standard coding practices)
Karl Williamson [Fri, 26 Oct 2012 17:23:51 +0000 (11:23 -0600)]
perldiag: Add entry for unknown charname
Karl Williamson [Fri, 26 Oct 2012 17:06:18 +0000 (11:06 -0600)]
toke.c: Don't validate core charnames output
charnames now refuses to allow invalid aliases to be installed, so we
only need to validate those that come from a custom translator.
Karl Williamson [Fri, 26 Oct 2012 16:48:48 +0000 (10:48 -0600)]
charnames: Don't accept illegal :aliases
Now that improper names for characters are an error, we can forbid them
at definition time. For the time being allow a colon in the check that
continues to be run in toke.c. This will be removed in a future commit.
Karl Williamson [Thu, 8 Nov 2012 18:01:38 +0000 (11:01 -0700)]
regen/mk_PL_charclass.pl: Use mktables table for charname
This commit uses the mktables defined table for whether or not a
character is a legitimate charname continuation. This will allow it to
be kept in sync with other code that needs the definition.
The only change this makes is to delete "colon" from being a legitimate
continuation character. A colon was only accepted because it was used
in the paradigm for like "Greek: Alpha", and is not part of any
actual character name.
Karl Williamson [Thu, 8 Nov 2012 17:59:11 +0000 (10:59 -0700)]
mktables: Create tables for charname beginning and follow-on
mktables is changed to add two new tables, one that matches the first
character in a character names, and one that matches continuation
characters.