| Commit message (Collapse) | Author | Age | Files | Lines |
|
|
|
|
| |
p4raw-link: @25793 on //depot/perl: b0a2e8e64dee1b1bd00055b3c5b69e2c3c888f55
p4raw-id: //depot/perl@25794
|
|
|
|
|
| |
p4raw-link: @25733 on //depot/perl: bdf33aa7435b6e4ab927ed6430f3f240f211d879
p4raw-id: //depot/perl@25734
|
|
|
|
|
| |
Message-ID: <lrk6glej0e.fsf@caliper.activestate.com>
p4raw-id: //depot/perl@25728
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
The nmake makefile (win32/Makefile) currently adds -nologo, $(STRPOOL)
and -W3 to the ccflags in CFG_VARS, but the dmake makefile
(win32/makefile.mk) doesn't. This means that extensions built with
perls built with VC++ and dmake spit out damn M$ logo messages and use
a lower warning level.
The reason for the difference is probably that the dmake makefile
supports multiple compilers. Therefore, put the extra ccflags into a
new macro, and set it appropriately for each compiler. Add the new
macro to the nmake makefile too for orthogonality.
p4raw-id: //depot/perl@25694
|
|
|
|
|
| |
Message-ID: <lrd5ml7i8s.fsf@caliper.activestate.com>
p4raw-id: //depot/perl@25688
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
Introduce new CCTYPE's MSVC70FREE and MSVC70 for the free and the full
versions of MS VC++ 7.x respectively. Perl can now be built with the
free VC++ toolkit simply by running "nmake CCTYPE=MSVC70FREE" with no
hacking of the win32/Makefile required. Update README.win32 with
these changes.
Also add the CCTYPE MSVC to win32/Makefile. This is already supported
in win32/makefile.mk and is useful for avoiding the use of -DELAYLOAD
for MS VC++ 3.x through 5.x. Update README.win32 for this too.
p4raw-id: //depot/perl@25672
|
|
|
| |
p4raw-id: //depot/perl@25573
|
|
|
|
|
| |
p4raw-link: @25387 on //depot/perl: 7dd121aefd69a3ba2201a35698deaeda3a8a77ad
p4raw-id: //depot/perl@25488
|
|
|
|
|
|
| |
From: "Jan Dubois" <jand@ActiveState.com>
Message-Id: <200508222237.j7MMbFPA027764@smtp3.ActiveState.com>
p4raw-id: //depot/perl@25321
|
|
|
|
|
|
|
|
|
|
|
| |
It makes life easier for third-party extensions if perl provides a
decent set of libs to link against. cl and gcc builds already have
this (except that gcc was missing one lib), but bcc32 builds only
have one or two libs supplied. In particular, the ODBC libs
required for building libwin32 were not specified. One of these
doesn't exist in the Borland lib dirs by default, so generate it
from the DLL if it is missing.
p4raw-id: //depot/perl@25289
|
|
|
|
|
| |
Thanks to John E. Malmberg <wb8tyw@qsl.net> for pointing this out.
p4raw-id: //depot/perl@25288
|
|
|
| |
p4raw-id: //depot/perl@25286
|
|
|
|
|
|
| |
From: "Piotr Fusik" <pfusik@op.pl>
Message-ID: <001401c595bd$dccb5d80$0bd34dd5@piec>
p4raw-id: //depot/perl@25261
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
This was disabled by change 22689 because it was severely broken
at the time. See the thread starting here:
http://www.xray.mpe.mpg.de/mailing-lists/perl5-porters/2004-04/msg00370.html
The PERL_MALLOC_WRAP code in handy.h has since changed along the
lines of what was discussed in that thread (in particular, by change
25034) and I find that everything is now fine with PERL_MALLOC_WRAP
re-enabled.
p4raw-id: //depot/perl@25226
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
Trawling through all these config files, I think I've found the
reason why bcc32 builds with USE_LARGE_FILES were failing (e.g.
(see http://www.xray.mpe.mpg.de/mailing-lists/perl5-porters/2005-06/msg00809.html):
lseeksize and lseektype were set to 8 and __int64, but Borland's
<io.h> suggests that they should be 4 and long respectively.
Changing them accordingly makes all tests pass. And I see that
win32/win32.c has special handling for __BORLANDC__ + USE_LARGE_FILES,
presumably to cope with this difference?
p4raw-id: //depot/perl@25216
|
|
|
|
|
|
|
|
|
|
|
| |
uselargefiles, usemultiplicity, useperlio and useithreads are now
off in the canned win32/config_H.* files, so change these to match.
Also scrap the misleading ~USE_PERLIO~ and ~USE_SITECUST~ tags
which don't actually work since no such CFG_VARS are passed in to
config_sh.PL. (The useperlio and usesitecustomize entries actually
get substituted by the more general "if (/^([\w_]+)=(.*)$/)" case.)
p4raw-id: //depot/perl@25215
|
|
|
|
|
|
| |
No existing entries changed, just adding some missing ones based
on Porting/config.sh
p4raw-id: //depot/perl@25214
|
|
|
| |
p4raw-id: //depot/perl@25211
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
These are used in their unedited form to build miniperl, so should
represent a minimal configuration
The plan is to use miniperl's "-V" output in makedef.pl, so it is
important that it doesn't include any of the optional things in the
makefiles that the user might have chosen to disable
(The actual config.h used to build perl itself is, of course, an
edited version of these canned configs produced by using miniperl
to run config_h.PL so these changes won't affect perl itself)
p4raw-id: //depot/perl@25208
|
|
|
|
|
|
| |
No actual changes, just reordering the contents of the files so that
diff works better.
p4raw-id: //depot/perl@25206
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
They were broken because change 24806 forgot to add -DWITH_STATIC
to the recipe for compiling perllib.c which meant that perllibst.h
was not included and hence although they built OK, the static
extensions could not actually be used!
In fact, since perllibst.h is always created (it is more or less
empty when static extensions are not being used), there is no need
for -DWITH_STATIC at all. So rather than adding it to Makefile,
just drop it from makefile.mk and perllib.c.
Also add the missing cleanup of perllibst.h to Makefile.
p4raw-id: //depot/perl@25150
|
|
|
|
|
|
|
|
| |
Message-ID: <42CC3CE9.5050606@divsol.com>
(reverted all dual-lived modules since they must work with older
perls too so must wait for a new Devel::PPPort)
p4raw-id: //depot/perl@25101
|
|
|
| |
p4raw-id: //depot/perl@25068
|
|
|
| |
p4raw-id: //depot/perl@25040
|
|
|
|
|
|
| |
See the thread starting here for arguments:
http://www.xray.mpe.mpg.de/mailing-lists/perl5-porters/2005-06/msg00994.html
p4raw-id: //depot/perl@25033
|
|
|
|
|
| |
Message-ID: <Mahogany-0.66.0-1232-20050627-220727.00@rbnet.it>
p4raw-id: //depot/perl@24999
|
|
|
|
|
| |
Doesn't fix bcc32 + USE_LARGE_FILES, though :-(
p4raw-id: //depot/perl@24977
|
|
|
| |
p4raw-id: //depot/perl@24975
|
|
|
|
|
|
|
|
| |
That fact that it wasn't const already was highlighted by a warning
from pp_open_dir() generated by change 24743. Rather than undo the
const change in pp_open_dir(), this seems to make more sense. Hope I
haven't broken Netware or WinCE.
p4raw-id: //depot/perl@24974
|
|
|
|
|
|
|
|
| |
It still whines about 'in' and 'b' being assigned values that are
never used on lines 348 and 505 respectively (as side effects of the
calls to c2l and l2c respectively), but I'm not sure how to shut
them up.
p4raw-id: //depot/perl@24947
|
|
|
|
|
| |
("Suspicious pointer conversion")
p4raw-id: //depot/perl@24946
|
|
|
|
|
| |
("Superfluous & with function")
p4raw-id: //depot/perl@24937
|
|
|
| |
p4raw-id: //depot/perl@24926
|
|
|
|
|
|
| |
The try.c test program added to Configure by change 24898 outputs
"Yes" using each of VC++, GCC and BCC32, so here goes...
p4raw-id: //depot/perl@24905
|
|
|
|
|
|
| |
This fixes failures in ext/B/t/deparse.t and lib/warnings.t (and
renders change 24870 obsolete, hence it is reverted here).
p4raw-id: //depot/perl@24883
|
|
|
|
|
|
| |
From: "Jan Dubois" <jand@ActiveState.com>
Message-Id: <200506161843.j5GIhKhL024472@smtp3.ActiveState.com>
p4raw-id: //depot/perl@24880
|
|
|
| |
p4raw-id: //depot/perl@24866
|
|
|
| |
p4raw-id: //depot/perl@24863
|
|
|
| |
p4raw-id: //depot/perl@24862
|
|
|
|
|
|
|
|
|
|
|
|
| |
This pattern match seems a little saner, and stops " for Win32"
getting picked up from the following output:
Borland C++ 5.5.1 for Win32 Copyright (c) 1993, 2000 Borland
Maybe it breaks older versions, but at least the latest now works
correctly. If anyone knows what older versions' messages look like
then speak up!
p4raw-id: //depot/perl@24855
|
|
|
| |
p4raw-id: //depot/perl@24852
|
|
|
|
|
| |
(IO/Socket still needs to go, though)
p4raw-id: //depot/perl@24837
|
|
|
|
|
|
|
|
|
| |
This allows extra build options, e.g. -DNO_HASH_SEED, to be
specified on the command line, which is otherwise not possible
under nmake using the existing BUILDOPT macro
Suggested by Jan Dubois <jand@ActiveState.com>
p4raw-id: //depot/perl@24830
|
|
|
|
|
| |
Message-Id: <1118691701.6333.32.camel@mdk10>
p4raw-id: //depot/perl@24825
|
|
|
|
|
| |
Message-Id: <1118599750.2732.2.camel@mdk10>
p4raw-id: //depot/perl@24806
|
|
|
| |
p4raw-id: //depot/perl@24792
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
the relevant makefiles and tables-of-contents to reflect this.
p4raw-id: //depot/perl@24680
p4raw-branched: from //depot/maint-5.8/perl@24679 'branch in'
pod/perl587delta.pod
p4raw-integrated: from //depot/maint-5.8/perl@24679 'copy in'
pod/perlhist.pod (@24141..) 'merge in' win32/pod.mak (@23153..)
p4raw-integrated: from //depot/maint-5.8/perl@23573 'edit in' pod.lst
(@23153..) Makefile.SH (@23406..) pod/perl.pod (@23476..)
'ignore' pod/perltoc.pod (@23408..) win32/Makefile
win32/makefile.mk (@23480..) vms/descrip_mms.template
(@23560..) 'merge in' MANIFEST (@23461..)
|
|
|
|
|
|
|
| |
Message-ID: <20050531153327.85145.qmail@unknown-209-73-178-237.yahoo.com>
(with a couple of minor tweaks)
p4raw-id: //depot/perl@24679
|
|
|
| |
p4raw-id: //depot/perl@24615
|
|
|
|
|
|
| |
(None are actually defined, but it's nice to have them there anyway)
p4raw-link: @24508 on //depot/perl: 0dbb1585a715e56312e579a5f0e7f82241b38352
p4raw-id: //depot/perl@24545
|