This is a live mirror of the Perl 5 development currently hosted at https://github.com/perl/perl5
Jarkko Hietaniemi [Sat, 8 Aug 2015 20:39:25 +0000 (16:39 -0400)]
Give a better error if malloc.c is used without MYMALLOC
The build will die already on errors because of the missing ASSERT
macros (plus get warnings on the implicit declarations, i.e. missing
prototypes), but better to give a clear error.
FWIW, looking at hints/*, only HP-UX of the even remotely
common/current systems seems to use Perl's malloc anymore,
and then only if perlio is NOT being used.
(Other platforms that seemingly use Perl's malloc include ancient
SysVs like SCO 2.3.4, NCR Tower, Tektronix' UTek V, then Unicos and
Unicos/mk of Cray, and NEC's SUPER-UX.) (OS/2 still probably does
use it, given all the work ilyaz used to pour on it.)
Tony Cook [Mon, 10 Aug 2015 01:39:07 +0000 (11:39 +1000)]
bump $ExtUtils::ParseXS::VERSION to 3.29
Ed J [Sun, 14 Jun 2015 23:40:36 +0000 (00:40 +0100)]
Fix INCLUDE_COMMAND $^X for Perl In Space
Steve Hay [Fri, 7 Aug 2015 07:39:07 +0000 (08:39 +0100)]
CPAN-Meta-YAML is now synced with 0.017-TRIAL
Delete Changes file. We don't put IGNORABLEs in cpan/ folders.
Tony Cook [Fri, 7 Aug 2015 04:20:36 +0000 (14:20 +1000)]
remove the byte-order-mark introduced to sv.c by
5488d373
This causes the build to fail on NetBSD
James E Keenan [Wed, 5 Aug 2015 01:50:47 +0000 (21:50 -0400)]
t/00-report-prereqs.dd no longer part of CPAN-Meta-YAML.
Removed it from MANIFEST. Added two more files missing from MANIFEST.
Had to run ./perl -Ilib Porting/metadata to update META.yml.
Maintainer reports that the "patch is safe to be merged now"; hence,
merging notwithstanding '-TRIAL' comment.
For: RT #125748
Karen Etheridge [Tue, 4 Aug 2015 15:30:31 +0000 (08:30 -0700)]
Upgrade CPAN-Meta-YAML from 0.016 to 0.017-TRIAL
from https://cpan.metacpan.org/authors/id/D/DA/DAGOLDEN/CPAN-Meta-YAML-0.017-TRIAL.tar.gz
meta files omitted:
LICENSE
MANIFEST
META.json
META.yml
README
t/00-report-prereqs.t
xt/
Sisyphus [Wed, 5 Aug 2015 20:53:38 +0000 (16:53 -0400)]
double-double long double %a fixes
- reserve enough buffer space
- name the two different errors differently
- test around the problem spot
James E Keenan [Sun, 2 Aug 2015 13:15:52 +0000 (09:15 -0400)]
Bring list of directories requiring custom TAP up-to-date.
The list of directories in 'perltodo' which t/test.pl may not be used and in
which hand-rolled TAP is required was out-of-date. It did not reflect the
creation in 2012 of the t/opbasic directory. It failed to include the t/comp
directory.
Tony Cook [Mon, 3 Aug 2015 06:48:01 +0000 (16:48 +1000)]
[perl #120826] correct documentation of sv parameter of call_sv()
Aaron Priven [Wed, 5 Aug 2015 00:08:56 +0000 (20:08 -0400)]
Added {unlock,lock}_hashref_recurse to @EXPORT_OK
In Hash::Util, the functions lock_hashref_recurse and
unlock_hashref_recurse were omitted from the @EXPORT_OK array,
the synopsis in the POD, and the test for exported functions in
t/Util.t. This commit adds those functions in all three places.
For: RT #125730
James E Keenan [Thu, 30 Jul 2015 00:58:42 +0000 (20:58 -0400)]
Guarantee that, after unlocking, the hash is actually assignable.
Hash::Util::unlock_hashref_recurse(); Hash::Util::unlock_hash_recurse().
Thanks to report from Diab Jerius.
For: RT #125721
Ricardo Signes [Mon, 3 Aug 2015 02:44:21 +0000 (22:44 -0400)]
add a small note that find2perl is not in core
Daniel Dragan [Wed, 29 Jul 2015 07:31:23 +0000 (03:31 -0400)]
Safefree(NULL) reduction
locale.c:
- the pointers are always null at this point, see
http://www.nntp.perl.org/group/perl.perl5.porters/2015/07/msg229533.html
pp.c:
- reduce scope of temp_buffer and svrecode, into an inner branch
- in some permutations, either temp_buffer is never set to non-null, or
svrecode, in permutations where it is known that the var hasn't been set
yet, skip the freeing calls at the end, this doesn't eliminate all
permutations with NULL being passed to Safefree and SvREFCNT_dec, but
only some of them
regcomp.c
- dont create a save stack entry to call Safefree(NULL), see ticket for
this patch for some profiling stats
Tony Cook [Mon, 3 Aug 2015 01:41:42 +0000 (11:41 +1000)]
use a more sensible homepage value for XSLoader
Reini Urban [Fri, 13 Mar 2015 15:34:38 +0000 (16:34 +0100)]
xs_boot_epilog (Xp): change ax from U32 to I32
we need to match the declaration of ax
Jarkko Hietaniemi [Sun, 2 Aug 2015 12:39:32 +0000 (08:39 -0400)]
Remove uselongdoubles false claim.
Missed this spot in
568793b6.
Jarkko Hietaniemi [Sun, 2 Aug 2015 02:50:39 +0000 (22:50 -0400)]
Check for POSIX locales once.
Makes loc_tools users safe under minitest.
Currently op/quotemeta.t and op/taint.t are failing there.
Karl Williamson [Fri, 15 May 2015 19:49:31 +0000 (13:49 -0600)]
utf8.h: Add dummy param for when macros placed in API
These macros are not in the public API, but they might be someday. We
may want to check for valide UTF-8 at some point. Add a parameter so
that is possible then without having to change the API.
This also changes to use the short name of one macro
Karl Williamson [Fri, 15 May 2015 19:32:23 +0000 (13:32 -0600)]
t/op/tr.t: Fix comment
This was confusing, and a likely interpretation is wrong: [i-j] doesn't
match the gap characters in EBCDIC
Karl Williamson [Fri, 15 May 2015 15:16:18 +0000 (09:16 -0600)]
utf8.h: Fix typo in macro name definition
The trailing underscore was unintended.
Karl Williamson [Fri, 15 May 2015 15:05:14 +0000 (09:05 -0600)]
utf8.h, utfebcdic.h: Add comments; white-space only
Karl Williamson [Thu, 14 May 2015 01:25:22 +0000 (19:25 -0600)]
t/opbasic/cmp.t: Add test for NUL
NUL can present problems for string handling; better test it
Karl Williamson [Wed, 13 May 2015 23:53:55 +0000 (17:53 -0600)]
pp_pack.c: Add comment
Karl Williamson [Wed, 13 May 2015 23:53:26 +0000 (17:53 -0600)]
pp.c: Fix typo in comment
Karl Williamson [Wed, 13 May 2015 23:46:00 +0000 (17:46 -0600)]
regen/mk_PL_charclass.pl: Suppress extra null array element
We don't output a trailing comma after the final element in these C
arrays, and thus prevent the C compiler from generating a useless null
element
Karl Williamson [Wed, 13 May 2015 23:27:48 +0000 (17:27 -0600)]
utf8.h: Add UTF8_SKIP as a synonym for UTF8SKIP
Most of the other names in utf8.h have an underscore; this allows
someone to keep things consistent in their code.
Karl Williamson [Wed, 13 May 2015 23:25:28 +0000 (17:25 -0600)]
utfebcdic.h: Comments only
Karl Williamson [Sun, 10 May 2015 00:21:55 +0000 (18:21 -0600)]
utf8.h: White-space only
Align some macro definitions vertically to make it easier to read
Karl Williamson [Sat, 9 May 2015 19:28:11 +0000 (13:28 -0600)]
handy.h: White-space only
Stops lines from exceeding 80 cols, and aligns columns vertical for
easier reading
Karl Williamson [Sat, 9 May 2015 03:23:34 +0000 (21:23 -0600)]
op.c: Fix typo in comment
Karl Williamson [Thu, 7 May 2015 17:11:41 +0000 (11:11 -0600)]
perlapi: Use F<> around file names
Karl Williamson [Thu, 7 May 2015 16:28:43 +0000 (10:28 -0600)]
perlapi: Use C<> instead of I<> for parameter names, etc
The majority of perlapi uses C<> to specify these things, but a few
things used I<> instead. Standardize to C<>.
Karl Williamson [Thu, 9 Apr 2015 18:41:53 +0000 (12:41 -0600)]
re/re_tests: Rmv comment
This was a comment I added for debugging, and forgot to remove
Karl Williamson [Tue, 7 Apr 2015 04:35:02 +0000 (22:35 -0600)]
toke.c: White-space, comment, added braces only
This makes && and || begin lines instead of end them for easier
following of the logic, as suggested in Perl Best Practices. A couple
of { } pairs were added to make the code within an 'if' stand out from
the 'if' test itself.
Steve Hay [Fri, 31 Jul 2015 07:50:26 +0000 (08:50 +0100)]
Upgrade Filter-Util-Call from version 1.54 to 1.55
Steve Hay [Fri, 31 Jul 2015 07:20:18 +0000 (08:20 +0100)]
Upgrade Encode from version 2.75 to 2.76
H.Merijn Brand [Thu, 30 Jul 2015 10:38:52 +0000 (12:38 +0200)]
Add Ivan Pozdeev to AUTHORS
H.Merijn Brand [Thu, 30 Jul 2015 09:08:38 +0000 (11:08 +0200)]
extra guard against spaces in paths (spotted by Ivan Pozdeev)
H.Merijn Brand [Thu, 30 Jul 2015 08:48:51 +0000 (10:48 +0200)]
Consistently use -o a.out (spotted by Ivan Pozdeev)
H.Merijn Brand [Thu, 30 Jul 2015 08:39:07 +0000 (10:39 +0200)]
There is no default dflt for usedtrace (spotted by Ivan Pozdeev)
Tony Cook [Thu, 30 Jul 2015 02:01:17 +0000 (12:01 +1000)]
perldelta for
3480fbaaaea8
Daniel Dragan [Wed, 29 Jul 2015 13:54:51 +0000 (09:54 -0400)]
Cwd.pm, dont repeatedly access magic %ENV vars
%ENV is magic/tied, a lexical is faster than calling out to magic APIs
with gets and sets (or checks for them). Less glob and hash lookups too.
Return the lexical instead of the magic var so there isn't a search for
magic from pp_leavesub's sv_mortalcopy. Even though env mg vtable doesn't
have a getter or GMAGIC, SvVSTRING_mg executes mg_find for random, get
and set magic inside Perl_sv_setsv_flags.
Chris 'BinGOs' Williams [Wed, 29 Jul 2015 13:42:45 +0000 (14:42 +0100)]
Update perlfaq to CPAN version 5.021010
[DELTA]
5.021010
20152015-07-28 18:30:48Z
* Fix typo (Douglas L. Jenkins)
* Fix Time::Seconds example (Niko Tyni)
* Basic README.pod (Shlomi Fish)
Jarkko Hietaniemi [Tue, 28 Jul 2015 23:11:38 +0000 (19:11 -0400)]
Use U64 (a type) instead of U64TYPE (a define).
Builds on top of
16d89be8.
Karl Williamson [Wed, 29 Jul 2015 04:17:01 +0000 (22:17 -0600)]
Allow perl to work again with all Unicode releases
Compiling perl to work with early Unicode releases is a goal, but has
been broken since 5.18. This sequence of commits gets this working
again, actually better than before.
Karl Williamson [Wed, 29 Jul 2015 03:24:40 +0000 (21:24 -0600)]
perldelta entry for mktables changes
Karl Williamson [Wed, 29 Jul 2015 03:19:52 +0000 (21:19 -0600)]
lib/unicore/README.perl: Update
This makes some changes and adds instructions for compiling older
Unicode versions.
Karl Williamson [Wed, 29 Jul 2015 02:52:57 +0000 (20:52 -0600)]
mktables: Minimize use of version numbers for decisions
I found these places where file existence can be used instead of knowing
what version something happened in. Sometimes those numbers are wrong,
and one of these was. If it can be avoided, better not to use version
numbers
Karl Williamson [Wed, 29 Jul 2015 02:34:22 +0000 (20:34 -0600)]
mktables: improve LineBreak table for early Unicodes
It turns out that the generated map files look better (even if
functionally equivalent) if the default mapping is the one to the
above-Unicode code points. This was the only one that had it different.
Karl Williamson [Tue, 28 Jul 2015 21:58:47 +0000 (15:58 -0600)]
mktables: Make sure \p{Space} works in all Unicodes
This isn't defined by Unicode until a later version, but Perl wants it
in all versions.
Karl Williamson [Tue, 28 Jul 2015 21:21:16 +0000 (15:21 -0600)]
mktables: Fix up Name_Alias in early Unicodes
perl needs the Name_Alias property accessible in all releases in order
for charnames to work properly. However the property was not created
until Unicode version 5.0. Previously, the property was made available
to all Unicode versions, which is contrary to the policy of exposing
properties to public use only when Unicode so exposes them. Thus the
behavior is as close as possible to Unicode-specified. This commit
creates an internal-only property for the perl core, and removes the
general access on early Unicode releases.
Karl Williamson [Tue, 28 Jul 2015 04:45:31 +0000 (22:45 -0600)]
mktables: Add handling of WB and SB for early Unicodes
This allows \b{wb} and \b{sb} to work on all Unicode releases. The huge
number of differences in charclass_invlists.h is only because the names
of the SB and WB tables change, and the code automatically
re-alphabetizes things.
Karl Williamson [Mon, 27 Jul 2015 14:12:02 +0000 (08:12 -0600)]
mktables: Fix GCB to work on early Unicodes
The GCB property was not properly being generated in early Unicode
releases. The huge commit diff is due solely to the fact that the name
changes of this property so it is sure to not be accessible outside the
perl core, and the property tables are automatically resorted
alphabetically.
Karl Williamson [Mon, 27 Jul 2015 04:57:51 +0000 (22:57 -0600)]
mktables: Convert Hangul_Syllable_Type to use new early infrastructure
This allows us to remove some special handling.
Karl Williamson [Mon, 27 Jul 2015 04:47:56 +0000 (22:47 -0600)]
mktables: Convert PropValueAliases.txt to use new early infrastructure
This file is crucial to compiling perl these days. This commit converts
to use the new infrastructure for dealing with compiling Unicode
releases prior to when this file was made available.
Karl Williamson [Mon, 27 Jul 2015 04:40:39 +0000 (22:40 -0600)]
mktables: Convert PropertyAliases.txt to use new early infrastructure
This file is crucial to compiling perl these days. This commit converts
to use the new infrastructure for dealing with compiling Unicode
releases prior to when this file was made available.
Karl Williamson [Mon, 27 Jul 2015 04:35:02 +0000 (22:35 -0600)]
mktables: Add code for easier handling of early Unicode versions
This adds infrastructure to the constructor of the Input_file class to
allow an alternative to be specified when compiling a Unicode release
that is earlier than the file first became available.
This is only used when the property is used by core perl and has to work
in all releases. For example the qr/\X/ construct should always work,
but relies on a property that isn't specified before Unicode 4.1. This
allows for easier specification of how to handle this type of case.
Karl Williamson [Mon, 27 Jul 2015 04:18:02 +0000 (22:18 -0600)]
mktables: Use Input_file class for always skipped files
Until this commit there were two mechanisms available to specify files
in the Unicode Character Database are not used by mktables. Now there
is one. The global that contained such files is deleted, and instead
all such files are specified by an Input_file class object. This has
the advantage of just one method, and the constructor already has
parameters to specify when a file first appeared, and when it was
removed. This allows automatic generation of the pod, listing just the
appropriate files for the version being compiled. It also allows for
the automatic check of all files to see that they are DOS 8.3 filesystem
compatible. And it allows for some code simplification.
Unicode specifies some .html files in the UCD. These are always skipped
(so far, and likely forever), and were in the global. Now they are in
the constructor, which means that the code that looks for potential
files that aren't being handled has to be changed to also look for .html
files as well.
Karl Williamson [Mon, 27 Jul 2015 04:25:57 +0000 (22:25 -0600)]
mktables: For 8.3 filesystems, the suffix matters
Two files can have the same file name, but be different if they have
different suffixes. Until this commit, mktables thought they were the
same, because it ignored the suffix when calculating this. Some files
are version strings like "3.1" which look like a floating point number.
These are converted to like "3_1" first so that the .1 doesn't look like
a suffix.
Karl Williamson [Sun, 26 Jul 2015 16:33:21 +0000 (10:33 -0600)]
mktables: Use new infrastructure for optional files
This follows up the previous commit by actually using the new
infrastructure it created. The optional Unihan files are switched to
use the new capabilities. This means that the globals they previously
used are no longer necessary, and are ripped out here.
Karl Williamson [Sun, 26 Jul 2015 16:14:11 +0000 (10:14 -0600)]
mktables: Improve handling of optional files
Some files in the Unicode Character Database are optional for perl.
Currently these are all Unihan files for East Asian languages. This
commit adds infrastructure to the class for input file objects to more
automatically handle the case when the file is present vs not.
Currently, one has to hand edit mktables to cope with the file getting
plunked down in the directory to be compiled. This commit is a step
towards making the editing unnecessary.
Karl Williamson [Mon, 27 Jul 2015 03:30:07 +0000 (21:30 -0600)]
mktables: Add specifiable construction handler for Input_file class
This object is for the files in the Unicode Character Database. This
commit adds the capability for the constructor new() to be passed a code
reference to call after everything else is constructed, but before
control is returned.
Karl Williamson [Sun, 26 Jul 2015 04:35:36 +0000 (22:35 -0600)]
mktables: Unicode has withdrawn some files in the UCD
This adds the capability to mktables to seamlessly handle the case where
Unicode has removed a file from the Unicode Character Database that
existed in prior versions. The constructor for the object tells what
release no longer has it. The file is looked for only when compiling a
Unicode version in the window that had it.
Karl Williamson [Sun, 26 Jul 2015 03:45:46 +0000 (21:45 -0600)]
mktables: Revamp constructor and run method for input objects
Each .txt file that Unicode furnishes as part of the Unicode Character
Database has an object created for it, and there is a run() method to
actually look at the input file and process it. This commit moves some
code from the run() method into the constructor to pave the way for
future commits (hence some of the awkward constructs and indentation in
this one; these minimize the commitdiffs in later ones).
Karl Williamson [Sun, 26 Jul 2015 01:19:05 +0000 (19:19 -0600)]
mktables: Allow multiple eof handlers for input files
The Input_file class is hereby extended so there can be more than one
subroutine stacked to call when an EOF is encountered. This new
capability will be used in later commits. There is no current need for
any outside class to look at this, so that capability is removed.
Karl Williamson [Sun, 26 Jul 2015 01:11:40 +0000 (19:11 -0600)]
mktables: Suppress output files that shouldn't be written
mktables takes files as input as furnished by Unicode, munges things,
and outputs files containing tables that are more suitable for use by
the rest of perl. A placeholder table is one that isn't supposed to be
output, but whose values are needed for internal calculations. Prior to
this commit, those tables were being output.
Karl Williamson [Sun, 26 Jul 2015 01:04:03 +0000 (19:04 -0600)]
mktables: Rmv extraneous sub return value
The return value of this function isn't looked at.
Karl Williamson [Sun, 26 Jul 2015 00:57:13 +0000 (18:57 -0600)]
mktables: Handle default for Gc property better
Unicode changed the default name they specify in their data from the
abbreviation to the long form in version 7
Karl Williamson [Sun, 26 Jul 2015 00:52:37 +0000 (18:52 -0600)]
mktables: Handle \b{gcb}, \X better in Unicode 1
The Hangul syllables in version 1 of Unicode were withdrawn and replaced
with the ones we know in version 2. Should perl ever be compiled on
version 1, this commit causes the old syllables to work better than they
do now, though imperfectly. This is used only in the \X and \b{gcb}
constructs.
Karl Williamson [Sat, 25 Jul 2015 04:43:55 +0000 (22:43 -0600)]
mktables: Two files came along later than mktables thinks
NormTest.txt LBTest.txt were really in later releases. This fixes that
and moves each constructor to be in the respective release's group of
files.
Karl Williamson [Fri, 24 Jul 2015 14:44:25 +0000 (08:44 -0600)]
mktables: Use variables for repeated text
This text is actually used to group values in the generated pod. By
using a variable, we make sure it is exact in every usage, so the values
are sure to get grouped properly.
Karl Williamson [Fri, 24 Jul 2015 03:04:03 +0000 (21:04 -0600)]
mktables: Rmv wrong code point in list
This was a typo, but it hasn't done any harm as it is an unassigned code
point, which gets weeded out. Still it has no business being here.
Karl Williamson [Sun, 5 Jul 2015 04:23:41 +0000 (22:23 -0600)]
mktables: Easier debugging
This turns off some inconsequential annoying warnings during certain
debugging situations
Karl Williamson [Sat, 11 Jul 2015 04:09:18 +0000 (22:09 -0600)]
mktables: Remove unused assignment
Karl Williamson [Mon, 27 Jul 2015 04:02:19 +0000 (22:02 -0600)]
mktables: Reindent after prev. commit
That commit added a nesting block, so indent an extra level, and reflow
the text to fit in 79 columns.
Karl Williamson [Sat, 4 Jul 2015 18:01:18 +0000 (12:01 -0600)]
Properly handle the Unicode kIICore property
This property is not included in the standard Perl distribution, but it
is normative in the Unicode Unihan database and perl can be compiled to
include it. This property is currently unique in that it operates much
like how Perl defines string truthfulness: non-empty values are
considered true. That is \p{kIICore} matches all characters which have
a non-empty value for this property, plus the actual values have
meaning that may need to be examined in some circumstances. These can
be retrieved via Unicode::UCD::prop_invmap().
Unicode 7.0 changed this property without my noticing, and went a very
different direction with it than I anticipated. And the perl
interpreter would loop when trying to deal with it under some
circumstances.
This property is true for all 'core' Chinese/Japanese/Korean characters
that every implementation of CJK things should strive to handle, i.e., the
minimally acceptable set, though the values now specify a precedence as
their first letter, A, B, or C (I suppose this means one could implement
just the A level ones first). The remaining letters in each value
encode the standards which were used as the source for the character.
In previous versions of the Standard, every non-null value was the
string "2.1".
Karl Williamson [Sat, 4 Jul 2015 17:43:16 +0000 (11:43 -0600)]
perluniprops: Don't list longer name as shorter
perluniprops has cross references to shorter synonyms for properties.
In the section for the properties accessible by Unicode::UCD (not \p{}),
sometimes the shorter name wasn't actually shorter. This fixes that.
Karl Williamson [Sun, 5 Jul 2015 04:28:05 +0000 (22:28 -0600)]
Unicode::UCD: Handle inverted input
No current input comes inverted, but it could some time in the future,
and we wouldn't know. In one case, it's easy to handle, so do so; in
another, die with a message so won't sneak past. At that point, if and
when it happens, time could be spent figuring out the best way to handle
the situation.
Karl Williamson [Wed, 1 Jul 2015 20:13:21 +0000 (14:13 -0600)]
lib/Unicode/UCD.t: White-space only
Re-indent after the previous commit
Karl Williamson [Sat, 27 Jun 2015 21:43:01 +0000 (15:43 -0600)]
lib/Unicode/UCD.t: Fix to work on older Unicodes
This commit causes this test to pass tests even when run on old
Unicodes, back to 3.0, where it becomes just too much. Most tests
aren't structured to pass on older Unicodes, but it's somewhat important
that this one does.
Karl Williamson [Sat, 27 Jun 2015 21:33:43 +0000 (15:33 -0600)]
Unicode::UCD: Add pod text about old Unicodes
Karl Williamson [Sat, 27 Jun 2015 21:32:49 +0000 (15:32 -0600)]
Unicode::UCD: Handle old Unicode files
The formats of some older Unicode releases can be different than
previously expected.
Karl Williamson [Sat, 27 Jun 2015 21:31:31 +0000 (15:31 -0600)]
Unicode::UCD: Handle old Unicode Blocks file format
Karl Williamson [Sat, 27 Jun 2015 21:26:55 +0000 (15:26 -0600)]
XS-APItest/t/handy.t: Handle early Unicodes
This changes, for properties that aren't defined in all Unicode
versions, to use synonyms that are defined in all. It also better
checks for empty property tables, and knows that the format of some
returns can be different than previously constrained to be.
Karl Williamson [Sat, 27 Jun 2015 21:19:52 +0000 (15:19 -0600)]
regen/mk_invlists.pl: Handle early Unicodes CF
In very early Unicode releases, the case folding table can be in a
different format.
Karl Williamson [Sat, 27 Jun 2015 20:32:15 +0000 (14:32 -0600)]
regen/mk_invlists.pl: White-space only
Reindent after the previous commit introduced an outer block
Karl Williamson [Sat, 27 Jun 2015 20:30:30 +0000 (14:30 -0600)]
regen/mk_invlists.pl: Properly handle empty properties
This failed to adequately handle empty properties; something that wasn't
seen until compiling older Unicode releases.
Karl Williamson [Sat, 27 Jun 2015 20:14:31 +0000 (14:14 -0600)]
regen/mk_PL_charclass.pl: Use names known in all Unicodes
This just changes, for properties that aren't defined in all Unicode
versions, to use synonyms that are defined in all
Karl Williamson [Sat, 27 Jun 2015 20:11:04 +0000 (14:11 -0600)]
regen/regcharclass.pl: Work on early Unicodes
This just changes, for properties that aren't defined in all Unicode
versions, to use synonyms that are defined in all
Karl Williamson [Sat, 27 Jun 2015 19:46:14 +0000 (13:46 -0600)]
mktables: Add new alias status for internal-only
This makes it easier to control whether a synonym for a property value
is shown to and is accessible by users.
Karl Williamson [Fri, 26 Jun 2015 20:17:06 +0000 (14:17 -0600)]
mktables: Get fold generation to work on early Unicodes
CaseFolding.txt isn't present in early Unicodes. For these lowercasing
is used instead. But the format of the lowercasing table isn't quite
the same, so add code to handle the differences.
Karl Williamson [Fri, 26 Jun 2015 19:20:18 +0000 (13:20 -0600)]
mktables: Skip Hangul autogeneration on early Unicodes
Unicode Version 1 used completely different Hangul Syllables, so the
algorithm for the later versions doesn't apply
Karl Williamson [Fri, 26 Jun 2015 18:27:41 +0000 (12:27 -0600)]
mktables: White-space, comment, debugging changes only
This adds comments, fixes up some white space and adds some things to
trace (when enabled)
Karl Williamson [Fri, 26 Jun 2015 18:23:13 +0000 (12:23 -0600)]
mktables: Make surrogates work in all Unicodes
Surrogates aren't in all Unicode releases. This creates a table that is
accessible in all releases, but is empty in those that don't have
surrogates.
Karl Williamson [Fri, 26 Jun 2015 18:14:12 +0000 (12:14 -0600)]
mktables: Fixup Unicode 5.0 propvalue aliases for Blocks
These used an anomalous form in this release only.
Karl Williamson [Fri, 26 Jun 2015 18:12:16 +0000 (12:12 -0600)]
mktables: Add true/false aliases in early Unicodes
Before Unicode 5.0, the aliases for True and False were just assumed,
and not specifically mentioned. This commit adds them for those early
releases.
Karl Williamson [Fri, 26 Jun 2015 17:59:30 +0000 (11:59 -0600)]
utf8_heavy.pl: White-space only
Reindent after previous commit added an outer block
Karl Williamson [Fri, 26 Jun 2015 17:50:18 +0000 (11:50 -0600)]
mktables: Allow strictly named map tables
There are several types of tables generated by mktables. Most are
binary (match) tables, but another class is mapping tables. The names
for these may be loosely matched, but until this commit only the match
tables could have strict matching applied.
Strict matching is used for certain table names where loose could be
ambiguous, and for all names that aren't to be used by anything except
the perl core.
Karl Williamson [Fri, 26 Jun 2015 17:27:49 +0000 (11:27 -0600)]
mktables: Move debugging line out of loop
It should only be printed once, when it does get printed (as it is
usually commented out).