| Commit message (Collapse) | Author | Age | Files | Lines |
|---|
| | |
|
| | |
|
| | |
|
| | |
|
| | |
|
| |
|
|
|
| |
This makes it so that the values for MEM_ALIGNBYTES and BYTEORDER
are taken from config.sh, instead of always getting a default.
|
| |
|
|
|
| |
This means that changes to static extensions like Win32CORE actually
result in them being rebuilt.
|
| |
|
|
|
|
|
|
|
| |
Due to a bug in the CRT (msvcr71.dll), these 2 functions are not defined
in any lib Perl can use (static link CRTs dont apply, Perl only uses DLL
CRTs), but they are available as intrinsics. This solves a link error
about missing symbol __rotl64 in hv.obj, from usage in hv_func.h, on 32 bit
USE_64_BIT_INT VC 2003 builds. _rotr64 is included for completeness. This
fix is filed as [perl #120925].
|
| |
|
|
|
|
|
| |
Most of these are actually generated, so the maintenance complexity reduction
is not as impressive as the diffstat suggests.
(Incorporating a fix from Merijn)
|
| |
|
|
|
|
|
|
|
|
|
| |
The last Perl release that built with -Dusesfio was v5.8.0, and even that
failed many regression tests. Every subsequent release fails to build, and
in the decade that has passed we have had no bug reports about this. So it's
safe to delete all the code. The Configure related code will be purged in a
subsequent commit.
2 references to sfio intentionally remain in fakesdio.h and nostdio.h, as
these appear to be for using its stdio API-compatibility layer.
|
| | |
|
| | |
|
| |
|
|
|
|
| |
This was preventing several things from working properly,
like Errno_pm.PL, which tried to use $Config{cppstdin},
which was not set properly.
|
| |
|
|
|
|
|
|
|
|
|
| |
The dependency on $(CONFIGPM) is for git_version.h to exist early enough
to be copied to the CORE dir. In real life, git_version.h still is built
by a Desktop build since WinCE doesn't yet support being built with a
system Perl. This commit is a step in getting WinCE to build without
needing a Desktop build first and for better integration with Castaway's
Cross changes in the future. The changes in this commit were tested with
other unpublished changes that tried to do a WinCE build with a system
Perl.
|
| |
|
|
|
|
|
| |
Group the nmake line onto 1 line similar to the dmake version, for 1 shell
call out vs 3. "&&" makes sure the build stops if miniperl returns non-0
instead of ignoring the code, cd'ing back to dir win32, and continuing
the build. Part of RT #120727.
|
| |\ |
|
| | | |
|
| | | |
|
| | | |
|
| | | |
|
| | |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| | |
-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
|
| | | |
|
| | |
| |
| |
| |
| |
| |
| |
| |
| | |
Macro LIBC was used since day 1 of WinCE port (commit e1caacb4fd ), but
never before defined. On Desktop Win32 Perl, LIBC is msvcrt.lib, so fix
corelibc/msvcrt to be LIBC and not part of CELIBS. Then use LIBC in a
sanity check. Sanity checks will give an error message, vs running the
C compiler and geting cryptic messages about unknown .hs and random
missing symbols.
|
| | |
| |
| |
| | |
Thanks to Daniel Dragan for pointing out the omission.
|
| | | |
|
| | | |
|
| | |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| | |
This patch is in RT as [perl #120091] but also fixes [perl #118059].
Because the MS C lib, doesn't support sockets natively, Perl uses
open_osfhandle, to wrap a socket into CRT fd. Sockets handles must be
closed with closesocket, not CloseHandle (which CRT close calls).
Undefined behavior will happen otherwise according to MS. Swap the now
closed socket handle in the CRT to INVALID_HANDLE_VALUE. The CRT will
not call CloseHandle on INVALID_HANDLE_VALUE and returns success if it
sees INVALID_HANDLE_VALUE. CloseHandle will never be called on a socket
handle with this patch.
In #118059, a race condition was reported, where accept() failed with
ENOTSOCK, when a psuedofork was done, and connection from the child fake
perl process to parent fake perl process was done. The race condition's
effects occur inside the user mode code in mswsock.dll!_WSPAccept in the
parent perl, which winds up having a kernel handle of an unknown type
in it that is invalid. The invalid handle is passed to kernel calls, which
fail, because the handle is invalid, and the error is translated to
ENOTSOCK. The exact mechanism of the bug and 100% reproducabilty of the
race were never established, since mswsock.dll is closed source.
Another benefit of this patch is making it easier to use C debuggers on
a Win32 Perl because of less debugger-only bad handle exceptions
(NtGlobalFlag FLG_ENABLE_HANDLE_EXCEPTIONS/0xC0000008 STATUS_INVALID_HANDLE).
This commit reverts parts of commit 9b1f18150a
"Get rid of PERL_MSVCRT_READFIX" and parts of commit 46e77f1118
"Don't use the PERL_MSVCRT_READFIX when using VC++ 7.x onwards." and
contains additional changes not found in those 2 commits. The method for
selecting the definition of struct ioinfo isn't perfect. It will break if
VC > 6 is changed to use the older msvcrt.dll. For some versions of the
CRT, like 2005/8.0, it is impossible to know the definition of ioinfo
struct at C compile time, since the struct increased in size a number of
times with higher build numbers of v8.0 CRT. SxS and security updates make
that same perl binary will run with different v8.0 CRTs at different times.
For the case when ioinfo can not be hard coded, introduce
WIN32_DYN_IOINFO_SIZE. With WIN32_DYN_IOINFO_SIZE, the size of the ioinfo
struct is determined on Perl process startup using _mize. Since VC 2013
is a brand new product at the time of this patch, even though its struct
is known, and 2008 through 2012 have been stable so far, don't assume at
this time 2013's ioinfo will be stable. Therefore, VC 2003 and older
(including Mingw's v6.0), 2008 to 2012, are hard coded. 2013 is a candidate
one day to be hard coded. VC 2005 can never be hard coded.
All non-WIN32_DYN_IOINFO_SIZE compilers will work with
WIN32_DYN_IOINFO_SIZE. Non-WIN32_DYN_IOINFO_SIZE mode is simply more
efficient.
For future compatibility concerns, 3 forms of protection are offered.
If __pioinfo isn't exported anymore, the Perl build will break. In
WIN32_DYN_IOINFO_SIZE mode, if __pioinfo isn't heap memory anymore or the
start of a memory block, the runtime panic will happen. In with and without
WIN32_DYN_IOINFO_SIZE, only on a DEBUGGING build, the handle returned by
CRT's _get_osfhandle, which is the authentic handle in the CRT, is compared
to the handle found by accessing the ioinfo struct directly. If they don't
match, the handle swapping code is broken and the assert fails.
|
| |/
|
|
|
| |
All Win32 Visual C compilers except for VC 6/_MSC_VER == 1200 support LTCG.
LTCG generates slightly smaller and slightly faster binaries.
|
| |
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
Two tests (t/io/fs.t and cpan/HTTP-Tiny/t/110_mirror.t) fail on my system,
currently in Daylight Saving Time, both due to times written/read by
utime()/stat() being off by one hour...
Not sure what the issue is yet, but I've reproduced it in this C program:
#include <stdio.h>
#include <sys/stat.h>
#include <sys/utime.h>
void main(void) {
struct _utimbuf ut;
struct _stat st;
time_t t;
t = 760060800;
printf("Setting: %ld\n", t);
ut.actime = t;
ut.modtime = t;
if (_utime("test.c", &ut) == -1) {
perror("_utime failed\n");
return;
}
if (_stat("test.c", &st) != 0) {
perror("_stat failed\n");
return;
}
printf(" atime: %ld\n", st.st_atime);
printf(" mtime: %ld\n", st.st_mtime);
}
which outputs
Setting: 760060800
atime: 760057200
mtime: 760057200
with VC++ 2013, instead of
Setting: 760060800
atime: 760060800
mtime: 760060800
like Visual C++ 6.0 through 2012 all do.
|
| |
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
- CCTYPE has always been empty string in Makefile.ce, makedef.pl does not
complain about this, but it is a bug, so fix it
- create a profile for a EVC4 arm build
- put -GS- and -GL flags into compiler version specific defs so VC6 era
EVC4 will work
- don't delete ..\lib\Config.pm, it was made by win32/Makefile, for a
Desktop Perl, deleting this breaks miniperl when it loads its pure perl
modules for cross MakeMaker module building, and requires rebuilding the
Desktop Perl so miniperl works again. CE Perl has its Config.pm in the
xlib\$(MACHINE) dir
|
| |
|
|
|
|
|
|
|
| |
Comments from the committer:
win32/mini/*.obj (and hence win32/mini/ since it contains nothing other than
*.obj) should be .gitignored by the top-level .gitignore file anyway, but
git clean -dfX doesn't remove them without this patch. (git clean -dfx does,
but some developers want to keep untracked files other than build products
explicitly listed by .gitignore files.)
|
| |
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
On a WinCE build. On the 2nd nmake run, using Makefile.ce, eventually calls
the Extensions target which calls make_ext.pl. What happens is nmake for CE
for each module is called on the Desktop per module makefile from the
earlier Desktop build. Since the Desktop Perl already was built
sucessfully, all rules/deps are met in the Desktop per module makefile, and
nothing happens during the module building phase for a CE build. Previously
I used external file management tools to delete the per module Makefiles
before running Makefile.ce.
*make_ext.pl
- implement deleting and rebuilding the per module makefile on a Cross
build
- use constants for constant folding, there are opportunities for other
variables to be converted to constants in the future
- fix a bug from commit baff067e71 where unlink() on a file with an open
handle ($mfh) didn't delete the file from disk and a new per module
makefile would be not be built by make_ext.pl later since the per module
makefile was still on disk. This was observed on Win32. Also harden the
unlink code with a new _unlink sub that is fatal if the file is still on
disk after unlink supposedly deleted it.
- var $header and the quotemeta is because of an issue in Perl #119793
*Makefile.ce
- bring the debugging symbol generation flags and optimization flags
to be closer to a Dekstop VC Perl build
- ICWD is obsolete as of commit f6b3c354c9 , remove it
- MINIMOD is obsolete as of commit 7b4d95f74b , remove it
- make a poisoned config.h so if there is a XS building mixup between
a desktop and CE perl, the poisoned config.h for CE will stop the build
gracefully
- $(MINIPERL) has never been defined in Makefile.ce from day 1 (10 years)
replace with $(HPERL) everywhere, this was causing things to not run
silently since $(MINIPERL) was empty string. Use HPERL instead of
MINIPERL to allow flexibility to use the full perl binary if necessery
one day
- better cleaning on root makefile clean target
*win32/win32.h
*win32/win32iop.h
- silence alot of redefinition warnings which gave pages of warnings on
each WinCE compliand
*mg.c
- win32_get_errno is only on WIN32 build not WINCE
a "nmake -f Makefile.ce all" will now build the CE interp and all modules
in 1 shot with no user intervention
|
| | |
|
| | |
|
| | |
|
| |
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
Commit ea95436966 made changes to how errno.h constants are handled for
VC++ 2010 and above, which have added new values in the range 100-140.
Some versions of gcc-4.8.0 and above are now catching up and provide some
of these new values too (e.g. binaries from http://mingw-w64.sourceforge.net/
but not currently those from http://www.mingw.org/), but they don't provide
all of them. EADDRINUSE is provided, so convert_errno_to_wsa_error() gets
included, but the compilation fails because the following #defines which
VC++ 2010 and above provide are missing:
EBADMSG
EIDRM
ENODATA
ENOLINK
ENOMSG
ENOSR
ENOSTR
ENOTRECOVERABLE
EOTHER
ETIME
ETXTBSY
Simply ignore (#ifdef away) these constants for those compilers that don't
provide them.
|
| |
|
|
|
| |
This was added in 5.5/5.6 as a backwards-compatibility measure
when taint was extended to happen in more places.
|
| | |
|
| | |
|
| |
|
|
|
|
|
| |
This simplifies some of the logic necessary for coping with its various
problems.
Suggested by Nicholas Clark.
|
| |
|
|
|
|
|
|
|
|
| |
The win32/config.[gv]c files are generally kept up to date these days (and
we have tests to check that) so no changes are required in them.
The win32/config_H.[gv]c files are regenerated as per instructions in the
win32/Makefile and win32/makefile.mk, being careful to restore a couple of
things otherwise lost from the config_H.gc file. The files are now in sync
with the top-level master configuration file, config_h.SH.
|
| | |
|
| | |
|
| |
|
|
|
|
|
| |
Also change the Windows makefiles to disable this option if it is set with
this compiler selected. (Use !UNDEF (for nmake) and != (for dmake) to be
sure of disabling the option if it is defined on the command-line rather
than by editing the makefile.)
|
| |
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
For some reason (I didn't figure out why) cpan/Win32/Win32.xs did not
compile as of the previous commit, but in any case we cannot just blindly
include winsock2.h because (a) it is too late if winsock.h has already
been included and (b) early WinCE doesn't have it.
The definition of _WINSOCKAPI_ was also wrongly not updated to _WINSOCK2API_
but was mistaken in its comment "Don't drag in everything" anyway: it's
actually just an old-fashioned "include guard", not a means to include a
minimal API from the file. (The mistaken comment came from the original
ext/Errno/Errno_pm.PL before this round of $!-related changes. That file
also had a redundant "#include <winsock.h>" left in it which should have
been removed along with earlier work that removed the other one, so tidy
that up too.)
Steal a hopefully more correct incantation from win32/include/sys/socket.h
instead.
Now that we allow for the possibility of winsock.h being used again, we
must take care not to make use of WSAECANCELLED if it is not defined.
(That is the only WSAExxx constant which we use that was new in winsock2.h.)
|
| | |
|
| | |
|
| |
|
|
|
|
|
| |
The new errno2.h only ensures that those new Exxx constants which are
required by convert_wsa_error_to_errno() have definitions; the other new
Exxx constants will remain undefined on compilers other than VC10+, so we
must take care not to use them unless we know they are defined.
|
| |
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
There is no need for it to worry about ensuring that standard errno.h values
like EINTR are defined.
There is no point in it defining dummy constants like EINVALIDPROCTABLE which
do not exist in errno.h at all. They would only be given the corresponding
WSAExxx values anyway, so the win32sck.c function might just as well return
them directly.
Other constants like ESOCKTNOSUPPORT which do not exist in errno.h on Windows
but presumably exist elsewhere since POSIX tries to exports them are still
worth defining (where possible -- it isn't possible for EHOSTDOWN since there
is no corresponding WSAEHOSTDOWN); add comments for these for clarity.
Also comment which errno.h constants are new in VC10, and mention three of
the four that win32/include/sys/socket.h used to redefine because they are
used in the perl core (ENOTSOCK, EAFNOSUPPORT and ECONNABORTED) -- the other
one (ECONNRESET) does not appear to be used in the perl core as far as I can
tell.
|
| |
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
VC++ 2010 and above define a "POSIX supplement" of errno values ranging from
EADDRINUSE (100) to EWOULDBLOCK (140), but sys_nerr is still 43 and strerror()
returns "Unknown error" for them. We already avoid using strerror() if errno
> sys_nerr, but we treat such values as Windows error codes (i.e. <winerror.h>
values) and look up the corresponding system messages for them. There is no
better plan for these POSIX supplement errno values, but they must be
converted from <errno.h> values to <winerror.h> values first otherwise we
will look up the wrong system message. In practice, we only expect to find
errno > sys_nerr in the case of Windows sockets errors (we used to assign
WSAGetLastError() to errno), so we simply convert Exxx values to WSAExxx
values where possible, and use a default <winerror.h> value otherwise (namely,
ERROR_INVALID_FUNCTION).
This change fixes code such as this:
perl -le "$!=107; print $!"
which now outputs the expected "No connection could be made because the
target machine actively refused it." rather than "The program stopped
because an alternate diskette was not inserted." (in VC10+ builds).
Also: Fix the spelling of ECANCELED.
|
