Malcolm Beattie [Tue, 25 Nov 1997 17:59:53 +0000 (17:59 +0000)]
Fix minor thinkos in hv.c and pp_ctl.c. This is 5.004_55.
p4raw-id: //depot/perl@301
Malcolm Beattie [Tue, 25 Nov 1997 16:29:36 +0000 (16:29 +0000)]
Add t/avhv.t to MANIFEST and bump patchlevel.h to 55.
p4raw-id: //depot/perl@300
Ilya Zakharevich [Tue, 25 Nov 1997 15:59:16 +0000 (15:59 +0000)]
Move malloc_mutex initialisation/destruction:
Subject: patch to 5.004_54 for pthreads with Perl's malloc
p4raw-id: //depot/perl@299
Owen Taylor [Fri, 21 Nov 1997 22:28:17 +0000 (17:28 -0500)]
Make hv_ functions cope better with 'm'-magic:
Subject: [5.004_54] Another neglected patch
p4raw-id: //depot/perl@298
Malcolm Beattie [Tue, 25 Nov 1997 15:47:36 +0000 (15:47 +0000)]
Fix typo in Thread.xs.
p4raw-id: //depot/perl@297
Malcolm Beattie [Tue, 25 Nov 1997 15:42:07 +0000 (15:42 +0000)]
Integrate from ansi branch to mainline.
p4raw-id: //depot/perl@296
Jarkko Hietaniemi [Tue, 25 Nov 1997 00:49:52 +0000 (16:49 -0800)]
AIX patch for DynaLoader/dl_aix.xs and hints/aix.sh:
Subject: Re: _54 on AIX
p4raw-id: //depot/perl@295
Jarkko Hietaniemi [Tue, 25 Nov 1997 00:49:52 +0000 (16:49 -0800)]
AIX patch for hints/aix.sh:
Subject: Re: _54 on AIX
p4raw-id: //depot/perl@294
Malcolm Beattie [Tue, 25 Nov 1997 14:17:05 +0000 (14:17 +0000)]
Fix scalar dereference of threadsv variables (e.g. $$_).
p4raw-id: //depot/perl@291
Jarkko Hietaniemi [Thu, 20 Nov 1997 06:10:51 +0000 (22:10 -0800)]
AIX patch (including Configure support for {sched,pthread}_yield,
pthread initial detach state, renaming perl_thread to perl_os_thread
and struct thread to struct perl_thread):
Subject: Re: _54 on AIX
p4raw-id: //depot/perl@290
Malcolm Beattie [Tue, 25 Nov 1997 12:33:02 +0000 (12:33 +0000)]
Rename perl_thread to perl_os_thread.
p4raw-id: //depot/perl@289
Jarkko Hietaniemi [Wed, 19 Nov 1997 08:07:10 +0000 (00:07 -0800)]
Remove bincompat3 support:
Subject: Re: ANNOUNCE: perl5.004_54 is available
p4raw-id: //depot/perl@288
Ilya Zakharevich [Fri, 21 Nov 1997 15:02:09 +0000 (10:02 -0500)]
Emacs/tags update:
Subject: Emacs/tags update for 5.004_54
p4raw-id: //depot/perl@287
Nick Ing-Simmons [Sun, 23 Nov 1997 23:03:56 +0000 (23:03 +0000)]
Add $$_ test
p4raw-id: //depot/ansiperl@286
Nick Ing-Simmons [Sat, 22 Nov 1997 21:29:30 +0000 (21:29 +0000)]
Duplicate perl_threadsv
p4raw-id: //depot/ansiperl@283
Nick Ing-Simmons [Sat, 22 Nov 1997 21:18:11 +0000 (21:18 +0000)]
Munge pseudo-Configure stuff to add -thread to archname as
Malcolm seems to think that is way to test for threads.
Update @INC stuffing hackery to have traditional @INC
search order archlib, privlib, sitearch, site.
p4raw-id: //depot/ansiperl@282
Nick Ing-Simmons [Sat, 22 Nov 1997 19:28:21 +0000 (19:28 +0000)]
Builds and passes all but english.t on win32 VC++
p4raw-id: //depot/ansiperl@281
Nick Ing-Simmons [Sat, 22 Nov 1997 18:10:50 +0000 (18:10 +0000)]
ansiperl builds with Borland C++ again
p4raw-id: //depot/ansiperl@280
Nick Ing-Simmons [Sat, 22 Nov 1997 16:42:51 +0000 (16:42 +0000)]
Resolve ansiperl against mainline
p4raw-id: //depot/ansiperl@279
Nick Ing-Simmons [Sat, 22 Nov 1997 16:30:27 +0000 (16:30 +0000)]
Resolve ansiperl against win32
p4raw-id: //depot/ansiperl@278
Gurusamy Sarathy [Sat, 22 Nov 1997 09:48:02 +0000 (09:48 +0000)]
- shift() inside BEGIN|END|INIT now shifts @ARGV instead of @_
- added a test for the above
- fixed up perly.c.diff and vms/perl_c.vms for above and added the
ansification hunks
p4raw-id: //depot/win32/perl@277
Gurusamy Sarathy [Sat, 22 Nov 1997 07:24:01 +0000 (07:24 +0000)]
[win32] Generic change in win32 branch: don't just turn on CRIPPLED_CC
when USE_THREADS. GCC for instance, can do without macros that use
globals. Instead, selectively re#define only those macros
that use globals to their functional equivalents. Tests 100% on
Solaris/gcc (after `chmod +x t/op/nothread.t t/lib/thread.t` (hint,hint)).
p4raw-id: //depot/win32/perl@276
Gurusamy Sarathy [Sat, 22 Nov 1997 05:27:04 +0000 (05:27 +0000)]
[win32] Integrate mainline.
p4raw-id: //depot/win32/perl@275
Malcolm Beattie [Fri, 21 Nov 1997 18:28:22 +0000 (18:28 +0000)]
$_ is now per-thread (rather a lot of changes). Only tested under
*-linux-thread at the moment.
p4raw-id: //depot/perl@274
Paul Marquess [Thu, 20 Nov 1997 23:12:09 +0000 (23:12 +0000)]
Filter patch to toke.c:
Subject: Tiny core patch for source filters
p4raw-id: //depot/perl@273
Nick Ing-Simmons [Fri, 21 Nov 1997 00:54:43 +0000 (00:54 +0000)]
Basic integrate of lastest perl into ansiperl
p4raw-id: //depot/ansiperl@272
Malcolm Beattie [Thu, 20 Nov 1997 12:12:00 +0000 (12:12 +0000)]
Initial stab at IRIX configuration support for threading. Manually
applied parts of following patches:
Subject: Perl 5.004_54 on IRIX
Date: Wed, 19 Nov 1997 18:37:14 +0200 (EET)
From: Jarkko Hietaniemi <jhi@iki.fi>
Subject: Re: Perl 5.004_54 on IRIX
Date: 19 Nov 1997 17:10:17 -0800
From: Scott Henry <scotth@sgi.com>
p4raw-id: //depot/perl@271
Malcolm Beattie [Wed, 19 Nov 1997 17:45:37 +0000 (17:45 +0000)]
The new jumbo regexp stuff did SSPUSHINT on a char* instead of
SSPUSHPTR causing Alpha to core dump in pat.t. While fixing it,
also fixed two instances of referring to SVs after destruction.
p4raw-id: //depot/perl@270
Malcolm Beattie [Wed, 19 Nov 1997 15:33:23 +0000 (15:33 +0000)]
avhv_keys under Digital UNIX made avhv.t fail because *keysp was
changed by mg_get(*keysp) (!). Introducing a new local variable
fixed it but I don't know if it's a compiler problem or some
other corruption happening elsewhere.
p4raw-id: //depot/perl@269
Jarkko Hietaniemi [Tue, 18 Nov 1997 19:08:19 +0000 (21:08 +0200)]
Let Configure sort out get{host,net}byaddr* prototypes:
Subject: [PATCH] 5.004_54: little something for
get{hos,ne}tbyaddr protos (Configure, config_h.SH, pp_sys.c)
p4raw-id: //depot/perl@268
Ilya Zakharevich [Sun, 16 Nov 1997 00:29:39 +0000 (19:29 -0500)]
Jumbo regexp patch applied (with minor fix-up tweaks):
Subject: Version 7 of Jumbo RE patch available
p4raw-id: //depot/perl@267
Joshua Pritikin [Sat, 15 Nov 1997 19:18:30 +0000 (14:18 -0500)]
Separate avhv_foo() key handling into avhv_keys(). Slightly tweaked
version of patch:
Subject: tie fake hash patch for 5.004_54
p4raw-id: //depot/perl@266
Malcolm Beattie [Tue, 18 Nov 1997 16:51:04 +0000 (16:51 +0000)]
Bring MANIFEST up to date. Add new thread tests.
p4raw-id: //depot/perl@265
Malcolm Beattie [Tue, 18 Nov 1997 16:41:27 +0000 (16:41 +0000)]
magic_setisa enhanced to update %FIELDS automatically when @ISA
is assigned to. Added tests to t/op/array.t. magic_setisa now
warns about including non-existent packages in @ISA when -w is on.
p4raw-id: //depot/perl@264
Malcolm Beattie [Tue, 18 Nov 1997 16:38:57 +0000 (16:38 +0000)]
Fix typo in win32 -> mainline integration.
p4raw-id: //depot/perl@263
Malcolm Beattie [Tue, 18 Nov 1997 11:56:09 +0000 (11:56 +0000)]
Integrate win32 branch back into mainline.
p4raw-id: //depot/perl@262
Gurusamy Sarathy [Tue, 18 Nov 1997 00:14:02 +0000 (00:14 +0000)]
[Win32] Export our own FD_SET() et al to complete sockets-as-handles pretense.
p4raw-id: //depot/win32/perl@261
Nick Ing-Simmons [Sun, 16 Nov 1997 23:16:16 +0000 (23:16 +0000)]
Generic file changes for MYMALLOC
p4raw-id: //depot/ansiperl@260
Nick Ing-Simmons [Sun, 16 Nov 1997 23:14:36 +0000 (23:14 +0000)]
MYMALLOC for Win32:
1. Initialize malloc_mutex before it is used (all platforms!)
2. Adjust #ifdef muddle to allow MYMALLOC and win32_ to coexist
3. Tweak win32/config*.* to define MYMALLOC
4. Provide sbrk() in terms of VirtualAlloc().
Also fixup -MT (perl95) build to handle Perl_current_thread
via call to DLL (as though an extension).
p4raw-id: //depot/ansiperl@259
Nick Ing-Simmons [Sat, 15 Nov 1997 20:42:28 +0000 (20:42 +0000)]
Implement dTHR via __declspec(thread) - part 2
p4raw-id: //depot/ansiperl@258
Nick Ing-Simmons [Sat, 15 Nov 1997 19:52:53 +0000 (19:52 +0000)]
Use __declspec(thread) var rather tha TslAlloc & co.
p4raw-id: //depot/ansiperl@257
Gurusamy Sarathy [Sat, 15 Nov 1997 02:58:09 +0000 (02:58 +0000)]
Add #include guard in Thread.xs so it will build even under
no USE_THREADS (for win32). This was missed because of edit
w/o checkout perforce kludge.
p4raw-id: //depot/win32/perl@256
Nick Ing-Simmons [Sat, 15 Nov 1997 00:33:46 +0000 (00:33 +0000)]
Integrate mainline (5.004_54?) into ansiperl
p4raw-id: //depot/ansiperl@255
Nick Ing-Simmons [Sat, 15 Nov 1997 00:25:26 +0000 (00:25 +0000)]
Interate win32 into ansiperl
p4raw-id: //depot/ansiperl@254
Gurusamy Sarathy [Fri, 14 Nov 1997 22:04:58 +0000 (22:04 +0000)]
Integrate mainline changes into win32 branch. Now would be a good time
to reverse integrate the win32 branch into mainline.
p4raw-id: //depot/win32/perl@253
Malcolm Beattie [Fri, 14 Nov 1997 15:07:19 +0000 (15:07 +0000)]
Two more delays added to test suite to help *-solaris-thread.
p4raw-id: //depot/perl@252
Malcolm Beattie [Fri, 14 Nov 1997 15:05:57 +0000 (15:05 +0000)]
Remove stale code from Thread.xs.
p4raw-id: //depot/perl@251
Malcolm Beattie [Fri, 14 Nov 1997 10:12:40 +0000 (10:12 +0000)]
Add delay to signal handling in t/op/magic.t. (Solaris with pthreads
doesn't run handlers for self-sent signals until kill has returned.)
p4raw-id: //depot/perl@250
Gurusamy Sarathy [Fri, 14 Nov 1997 05:14:44 +0000 (05:14 +0000)]
Fix various details in win32 makefiles and Config.pm setup.
- ldflags is set for both compilers now
- extensions list is now correct
- delete perl95.exe on distclean
- cf_time now gets updated (once)
- ccdlflags is set for Borland
- fix startperl so dprofpp works
p4raw-id: //depot/win32/perl@249
Malcolm Beattie [Thu, 13 Nov 1997 18:01:27 +0000 (18:01 +0000)]
Rewrite thread return code to distinguish between ordinary return
and die() and make join propagate the die. Add tiny method eval
which just does "return eval { shift->join; }". Add Thread::Specific
class for access to thread specific user data along with specific.t.
Rename Class to classname throughout Thread.xs for consistency.
Fix pp_specific to pp_threadsv in global.sym. Add support to
pp_entersub in pp_hot.c to lock stash for static locked methods.
p4raw-id: //depot/perl@248
Malcolm Beattie [Thu, 13 Nov 1997 14:13:30 +0000 (14:13 +0000)]
Change CONTEXT to PERL_CONTEXT throughout source (since the #define
to avoid the Digital UNIX clash no longer works). Changed the #ifdef
in pp_sys.c for whether getnet* function get protoyped (since the
default had a broken prototype for getnetbyaddr).
p4raw-id: //depot/perl@247
Nick Ing-Simmons [Thu, 13 Nov 1997 02:44:40 +0000 (02:44 +0000)]
Integrate Win32 branch
p4raw-id: //depot/ansiperl@246
Nick Ing-Simmons [Thu, 13 Nov 1997 00:47:54 +0000 (00:47 +0000)]
Integrate (-ay) win32 branch at its creation to
establish and ancestor as per perkforce technote #9
p4raw-id: //depot/ansiperl@245
Gurusamy Sarathy [Wed, 12 Nov 1997 22:26:39 +0000 (22:26 +0000)]
More cleanups of win32/win32*.[ch] files. win32/win32iop.h now
contains the all the declarations and macros for the win32io layer.
New std-ish functions are exported now. All win32-specific exported
functions begin with "win32_" consistently. win32 version of
init_os_extras() is now exported, so embedders can get the in-core
xsubs.
p4raw-id: //depot/win32/perl@244
Gurusamy Sarathy [Wed, 12 Nov 1997 07:41:52 +0000 (07:41 +0000)]
Really delete deleted files.
p4raw-id: //depot/win32/perl@243
Gurusamy Sarathy [Wed, 12 Nov 1997 07:40:54 +0000 (07:40 +0000)]
Egregious IOsubsystem code excised. Phew, what a relief! Two
files (win32/win32io.[ch]) completely removed, as are all traces
of them in makefiles and MANIFEST. RunPerl() retains the void* arg
for later. Various myfoo() things regularized to my_foo(). CPP not
required anymore to create a perl binary :)
p4raw-id: //depot/win32/perl@242
Gurusamy Sarathy [Wed, 12 Nov 1997 05:31:28 +0000 (05:31 +0000)]
Fix various win32 code blemishes:
- s/stolen/win32/g
- s/(CROAK|WARN)/lc($1)/eg
- remove deadcode from most places
p4raw-id: //depot/win32/perl@241
Gurusamy Sarathy [Wed, 12 Nov 1997 04:36:29 +0000 (04:36 +0000)]
Carry over changes in ansiperl branch. Win32 branch is now
the leading edge.
p4raw-id: //depot/win32/perl@240
Gurusamy Sarathy [Wed, 12 Nov 1997 03:39:57 +0000 (03:39 +0000)]
Add missing win32_closesocket() and export it (extension writers' complaint).
p4raw-id: //depot/win32/perl@239
Gurusamy Sarathy [Wed, 12 Nov 1997 03:25:17 +0000 (03:25 +0000)]
Clean up win32/win32sck.c (runtime load of Winsock now gone, it can be
done cleaner, if really needed (perhaps only for efficiency reasons?)).
Redundant EXTERN_C definitions and related warnings fixed.
p4raw-id: //depot/win32/perl@238
Nick Ing-Simmons [Wed, 12 Nov 1997 02:45:15 +0000 (02:45 +0000)]
Fixup Win32
- #undef start_env before re-#defining it
- change pp_specific pp_threadsv in global.sym
- re-build embed.h
- avoid HAVE_THREAD_INTERN - we don't and empty struct
is a pain. If we did have it it would contain cached
values of things we can only get at _IN_ the thread
so new_struct_thread is wrong place to call it.
- add new macro SET_THREAD_SELF - we must (in main thread)
define in win32thread.h, support in win32thread.c,
test and call in perl.c
p4raw-id: //depot/ansiperl@237
Nick Ing-Simmons [Wed, 12 Nov 1997 01:54:23 +0000 (01:54 +0000)]
Integrate mainline after it integrated us.
Accepted 'theirs' everywhere - so two branches should
now point to same files again.
Almost all of these were what was suggested, others were
whitespace diffs. A few dubious spots which we will now
go fix.
p4raw-id: //depot/ansiperl@236
Gurusamy Sarathy [Wed, 12 Nov 1997 01:22:26 +0000 (01:22 +0000)]
Minor tweaks to add a thread_intern struct that should ultimately
contain all the win32-specific statics.
Win32 branch now passes all tests with or w/o USE_THREADS.
p4raw-id: //depot/win32/perl@235
Gurusamy Sarathy [Tue, 11 Nov 1997 23:08:54 +0000 (23:08 +0000)]
Initial (untested) integration of mainline changes.
p4raw-id: //depot/win32/perl@234
Malcolm Beattie [Tue, 11 Nov 1997 18:07:30 +0000 (18:07 +0000)]
Typo in thread.h: ADD_THREAD_INTERN should be HAVE_THREAD_INTERN
p4raw-id: //depot/perl@233
Malcolm Beattie [Tue, 11 Nov 1997 17:49:12 +0000 (17:49 +0000)]
t/TEST (reverted to @229 version) should have been included in the
previous change (231) but my way of recovering it didn't work
properly. The change 231 comments about successful tests applies
to this t/TEST (i.e. as of this change).
p4raw-link: @229 on //depot/perl:
2faa37ccf8e46b865687f0ab4992b29a75eb79ea
p4raw-id: //depot/perl@232
Malcolm Beattie [Tue, 11 Nov 1997 17:46:59 +0000 (17:46 +0000)]
Fix up ansiperl integration. Back to passing all expected tests
with usethreads. Untested with non-threaded perl.
p4raw-id: //depot/perl@231
Malcolm Beattie [Tue, 11 Nov 1997 16:36:22 +0000 (16:36 +0000)]
Initial integration of ansi branch into mainline (untested).
p4raw-id: //depot/perl@230
Malcolm Beattie [Tue, 11 Nov 1997 15:20:43 +0000 (15:20 +0000)]
Change name of OP_SPECIFIC to OP_THREADSV. Fixed perl_get_sv when
getting per-thread magicals. Fixed thr->errsv initialisation.
p4raw-id: //depot/perl@229
Malcolm Beattie [Tue, 11 Nov 1997 12:48:26 +0000 (12:48 +0000)]
Fix up $@ (ERRSV now refers to GvSV(errgv) for non-threaded perl and
thr->errsv for threaded perl). Fix pp_tie and pp_dbmopen to use
GvCV(gv) instead of gv so AUTOLOAD stuff works. All tests now pass
again for non-threaded perl. Enhanced perl_get_sv to return
per-thread magicals where necessary for threaded perl.
p4raw-id: //depot/perl@228
Gurusamy Sarathy [Tue, 11 Nov 1997 02:11:23 +0000 (02:11 +0000)]
Slightly more refined lock() keyword recognition (using %INC).
p4raw-id: //depot/win32/perl@226
Gurusamy Sarathy [Tue, 11 Nov 1997 00:26:09 +0000 (00:26 +0000)]
"weak" lock keyword (hardcoded initial implementation) now works.
if not defined(&Thread::join) and defined(&__PACKAGE__::lock), 'lock'
is recognized as a sub, a regular keyword otherwise. Could be
generalized by storing a flag for every op in OP struct, and turning
the flag off when Thread.xs loads.
p4raw-id: //depot/win32/perl@225
Gurusamy Sarathy [Wed, 8 Oct 1997 02:18:23 +0000 (22:18 -0400)]
Merge a patch in preparation for "weak keywords":
Message-Id: <
199710080618.
CAA23899@aatma.engin.umich.edu>
Subject: [PATCH] global overrides for keywords
p4raw-id: //depot/win32/perl@224
Gurusamy Sarathy [Mon, 10 Nov 1997 22:41:31 +0000 (22:41 +0000)]
Remove runlevel. It was used to count how many runops() calls
we were in the process of executing, and longjmp() to the topmost
one (if not already there). We use a null top_env->je_prev
to distinguish that now.
p4raw-id: //depot/win32/perl@223
Gurusamy Sarathy [Mon, 10 Nov 1997 04:47:48 +0000 (04:47 +0000)]
Win32 branch now contains all non-ansification changes in ansiperl branch.
USE_THREADS case builds and passes all tests using both compilers.
Additional tweaks:
- fixup win32/makedef.pl to skip more symbols for non-thread build.
- sync win32/Makefile with win32/makefile.mk
>>>Non-thread build fails a lot of tests.<<<
p4raw-id: //depot/win32/perl@222
Gurusamy Sarathy [Mon, 10 Nov 1997 00:57:53 +0000 (00:57 +0000)]
Initial (untested) merge of all non-ansi changes on ansiperl branch
into win32 branch.
p4raw-id: //depot/win32/perl@221
Gurusamy Sarathy [Sun, 9 Nov 1997 22:44:41 +0000 (22:44 +0000)]
Integrate latest mainline into win32 branch.
p4raw-id: //depot/win32/perl@220
Nick Ing-Simmons [Sun, 9 Nov 1997 21:46:06 +0000 (21:46 +0000)]
Conditionalize english.t,
Enhance times() for NT,
(Failed) attempt to implement alarm(),
Fixed config.h dependancy in makefile.mk
p4raw-id: //depot/ansiperl@219
Nick Ing-Simmons [Sun, 9 Nov 1997 15:38:00 +0000 (15:38 +0000)]
Dick Hardt's patch for build on Alpha
p4raw-id: //depot/ansiperl@218
Nick Ing-Simmons [Sun, 9 Nov 1997 03:31:20 +0000 (03:31 +0000)]
MakeMaker not in vofig noise fix for dmake
p4raw-id: //depot/ansiperl@217
Nick Ing-Simmons [Sun, 9 Nov 1997 03:15:06 +0000 (03:15 +0000)]
Fix 'anydbm.t' - if the gv is passed 1st call to inherited
TIEHASH works, but 2nd call (after db is closed, attempt
to reopen) tries to AUTOLOAD TIEHASH rather than using
cached value.
p4raw-id: //depot/ansiperl@216
Nick Ing-Simmons [Sat, 8 Nov 1997 16:41:24 +0000 (16:41 +0000)]
Cleanup MakeMaker 'not in config' noise
p4raw-id: //depot/ansiperl@215
Nick Ing-Simmons [Sat, 8 Nov 1997 15:07:24 +0000 (15:07 +0000)]
Remove 'configure' leaving configure.gnu and Configure
Win32 ignores case and keeps trying to update
repository copy of 'configure' or 'Configure' with
the other.
p4raw-id: //depot/ansiperl@214
Nick Ing-Simmons [Sat, 8 Nov 1997 15:03:39 +0000 (15:03 +0000)]
Get threads working again on Win32
Root cause of fail was init_thread_intern() in
new_struct_thread() (which is called in parent thread)
clobbering dTHR of parent thread.
It is doubtfull if setting 'self' in new_struct_thread()
is 'right' but left in for now.
p4raw-id: //depot/ansiperl@213
Nick Ing-Simmons [Sat, 8 Nov 1997 00:34:03 +0000 (00:34 +0000)]
Add :base_thread to :default in Opcode.pm
This allows lib/safe.t to pass when threaded.
It is unclear if 'lock' should be safe as it allows
denial of service attack, but could not figure out
how to add just 'specific' (sic) to :default
without triggering 'already tagged' warning noise.
p4raw-id: //depot/ansiperl@212
Nick Ing-Simmons [Fri, 7 Nov 1997 23:59:31 +0000 (23:59 +0000)]
Merge changes as of 18:00 CST
p4raw-id: //depot/ansiperl@211
Nick Ing-Simmons [Fri, 7 Nov 1997 23:52:35 +0000 (23:52 +0000)]
Reverse integrate Malcolm's chanes into local
repository, then import result back into my view
of Malcolm's repository.
Builds and passes (most) tests with GNU C++/Solaris
and Borland C++, Win32.
p4raw-id: //depot/ansiperl@210
Malcolm Beattie [Fri, 7 Nov 1997 18:12:36 +0000 (18:12 +0000)]
Change pp_tie and pp_dbmopen to use perl_call_sv instead of a
DIY pp_entersub (in preparation for AUTOLOAD change). dbmopen
not tested. ofslen now maps to thr->Tofslen in thread.h. Added
missing #ifdef USE_THREADS around some DEBU_L statements in die().
Building without USE_THREADS fails quite a lot of tests. It looks
as though the move to per-thread magicals must be missing some
#ifdef USE_THREADS.
p4raw-id: //depot/perl@209
Nick Ing-Simmons [Fri, 7 Nov 1997 01:37:28 +0000 (01:37 +0000)]
Raw integrate of latest perl
p4raw-id: //depot/ansiperl@208
Malcolm Beattie [Thu, 6 Nov 1997 14:58:00 +0000 (14:58 +0000)]
Update README.threads and Todo.5.005.
p4raw-id: //depot/perl@207
Malcolm Beattie [Thu, 6 Nov 1997 14:37:37 +0000 (14:37 +0000)]
Remove #ifdef DEPRECATED stuff: newXSUB, pp_entersubr, FREE_TMPS().
p4raw-id: //depot/perl@206
Malcolm Beattie [Thu, 6 Nov 1997 14:31:38 +0000 (14:31 +0000)]
Per-thread magicals now stored in their own thr->magicals and keyed
more directly. cvcache and oursv become ordinary struct thread
fields instead of #defined thr->Tfoo ones. SvREFCNT_inc now checks
for 0 again. Main thread initialisation done by new function
init_main_thread instead of (now fixed) new_struct_thread.
p4raw-id: //depot/perl@205
Malcolm Beattie [Wed, 5 Nov 1997 17:18:18 +0000 (17:18 +0000)]
Per-thread magicals mostly working (and localisable). Now getting
intermittent occasional "Use of uninitialized value" warnings
which may be due to some op flag black magic I've broken.
p4raw-id: //depot/perl@204
Nick Ing-Simmons [Wed, 5 Nov 1997 01:04:10 +0000 (01:04 +0000)]
Builds C++ Borland, MSVC++ (Win32) and GCC++ (Solaris)
p4raw-id: //depot/ansiperl@203
Nick Ing-Simmons [Wed, 5 Nov 1997 00:50:27 +0000 (00:50 +0000)]
Compile(d) at least once with threads on win32
but did not work
p4raw-id: //depot/ansiperl@202
Nick Ing-Simmons [Wed, 5 Nov 1997 00:32:13 +0000 (00:32 +0000)]
Trivial integrate
p4raw-id: //depot/ansiperl@201
Malcolm Beattie [Tue, 4 Nov 1997 12:06:09 +0000 (12:06 +0000)]
Up patchlevel to 5.004_54 (I missed _53 for the last release).
p4raw-id: //depot/perl@200
Nick Ing-Simmons [Sat, 1 Nov 1997 00:18:52 +0000 (00:18 +0000)]
Integrate mainline @ 18:15 CST 31 Oct 1997
p4raw-id: //depot/ansiperl@199
Nick Ing-Simmons [Sat, 1 Nov 1997 00:08:33 +0000 (00:08 +0000)]
win32thread.* not in MANIFEST which has muddled moving
back and forth between depots.
p4raw-id: //depot/ansiperl@198
Nick Ing-Simmons [Sat, 1 Nov 1997 00:02:49 +0000 (00:02 +0000)]
Test changes
p4raw-id: //depot/ansiperl@197