-BEGIN PGP SIGNED MESSAGE-
Hash: SHA1
According to Eric Blake on 12/19/2007 2:22 PM:
> memmem isn't standardized, but even so, it has some bugs based on comparison
> with strstr.
>
> memmem(haystack,len,"",0) should return haystack, not NULL. This should be
-ahead for newlib, then we'll need to
delete cygwin's copy of memmem.cc.
- --
Don't work too hard, make some time for fun as well!
Eric Blake [EMAIL PROTECTED]
-BEGIN PGP SIGNATURE-
Version: GnuPG v1.4.5 (Cygwin)
Comment: Public key at home.comcast.net/~ericblak
of more symbols could not be
| found during program initialization. Usually this means an incorrect
| or out of date version of one or more DLLs is being erroniously found
| on the PATH.
| Killed
s/erroniously/erroneously/
- --
Don't work too hard, make some time for fun as well!
Eric Blak
__cleanup to occur _after_ _my_tls.call_signal_handler.
|
| I think you're right. The call to call_signal_handler was added years
| after the addition of the cleanup so it would make sense to move the
| cleanup handling after that.
|
| Want to submit a patch?
Sure.
2008-06-09 Eric B
Blake <[EMAIL PROTECTED]>
Fix usage of recently fixed Interlocked* functions.
* winbase.h (ilockincr, ilockdecr, ilockexch, ilockcmpexch): Add
volatile qualifier, to match Interlocked* functions.
- --
Don't work too hard, make some time for fun as well!
-BEGIN PGP SIGNED MESSAGE-
Hash: SHA1
|> 2008-07-12 Eric Blake <[EMAIL PROTECTED]>
|>
|> Fix usage of recently fixed Interlocked* functions.
|> * winbase.h (ilockincr, ilockdecr, ilockexch, ilockcmpexch): Add
|> volatile qualifier, to match Interlocked* func
-BEGIN PGP SIGNED MESSAGE-
Hash: SHA1
OK to apply, to match newlib and in preparation for POSIX 200x?
2008-09-19 Eric Blake <[EMAIL PROTECTED]>
* errno.cc (_sys_errlist): Add ECANCELED, ENOTRECOVERABLE,
EOWNERDEAD.
- --
Don't work too hard, make some time
key if a value called '@' exists? Do you
> have an idea for a simple solution?
"@" for the named value, and "%.val" for the unnamed default?
- --
Don't work too hard, make some time for fun as well!
Eric Blake e...@byu.net
-BEGIN PGP SIGNA
e could be a key and a value called @ so you
> have three @ items. :-P
If there is no key or value @, then use @ for the default for
compatibility. If there is either a key or a value named @, then use:
@ - named key
@%val - named value
%val - default value
- --
Don't work too hard,
om within Cygwin, so...?
And it's not standardized, which means portable code shouldn't use it.
--
Eric Blake
then NOW is the only time worth
doing it.
- --
Don't work too hard, make some time for fun as well!
Eric Blake e...@byu.net
-BEGIN PGP SIGNATURE-
Version: GnuPG v1.4.9 (Cygwin)
Comment: Public key at home.comcast.net/
st32_t): Define as int.
Are there any corresponding patches needed to ? I haven't
checked yet, but we should make absolutely sure that we are consistent
across all uses of these types.
- --
Don't work too hard, make some time for fun as well!
Eric Blake e...@byu.net
-BE
-BEGIN PGP SIGNED MESSAGE-
Hash: SHA1
I noticed a complaint about comparing signed and unsigned values, when
compiling with DEBUGGING enabled. net.cc also has a lot of trailing blanks.
2009-05-20 Eric Blake
* net.cc (gethostby_helper): Use correct signedness.
- --
Don&#
changes to mktemp.cc,
modulo a changed variable name, mirror newlib):
2009-07-03 Eric Blake
Add fpurge, mkstemps.
* cygwin.din (fpurge, mkstemps): New exports.
* include/cygwin/version.h (CYGWIN_VERSION_API_MINOR): Bump.
* mktemp.cc (_gettemp): Add parameter.
-BEGIN PGP SIGNED MESSAGE-
Hash: SHA1
According to Eric Blake on 7/3/2009 6:31 AM:
> With that vote of confidence, here's the patch (the changes to mktemp.cc,
> modulo a changed variable name, mirror newlib):
>
> 2009-07-03 Eric Blake
>
>
e arc4random approach to newlib?
- --
Don't work too hard, make some time for fun as well!
Eric Blake e...@byu.net
-BEGIN PGP SIGNATURE-
Version: GnuPG v1.4.9 (Cygwin)
Comment: Public key at home.comcast.net/~ericblake/eblake.gpg
Comment: Using GnuPG with
-BEGIN PGP SIGNED MESSAGE-
Hash: SHA1
POSIX requires dup2(1,1) to return 1 (if stdout is open), not 0. I wonder
how long that bug has been present? And the STC:
#include
int main() { return dup2 (1, 1); }
2009-07-21 Eric Blake
* dtable.cc (dup2): Correct return value for
-BEGIN PGP SIGNED MESSAGE-
Hash: SHA1
According to Christopher Faylor on 8/30/2009 6:55 PM:
> On Sun, Aug 30, 2009 at 06:32:59PM -0600, Eric Blake wrote:
>> According to Eric Blake on 8/24/2009 9:15 AM:
>>> While we're at it, fcntl and dup2 both have another minor
-BEGIN PGP SIGNED MESSAGE-
Hash: SHA1
According to Eric Blake on 9/3/2009 9:58 AM:
> faccessat has at least two, and probably three bugs.
Here's a fix for 1 (typo) and 3 (check for EINVAL in more places), but not
for 2 (euidaccess, and the followup request of lchmod).
2009-09-
l sum total over the past.
- --
Don't work too hard, make some time for fun as well!
Eric Blake e...@byu.net
-BEGIN PGP SIGNATURE-
Version: GnuPG v1.4.9 (Cygwin)
Comment: Public key at home.comcast.net/~ericblake/eblake.gpg
Comment: Using GnuPG with Mozilla - h
en processed yet? It's been more
than a week since I mailed them, and although the USPS isn't lightning
fast, a week is generally long enough for domestic mail.
- --
Don't work too hard, make some time for fun as well!
Eric Blake e...@byu.net
-BEGIN PGP SIGNATURE
cc (fcntl64): Obey POSIX rules.
>> * syscalls.cc (dup2): Likewise.
>
> Thanks for the patch. Go ahead and check this in.
>
> In particular, thanks for turning (100 * NOFILE_INCR) into a #define.
Now applied.
- --
Don't work too hard, make some time for fun as well!
Eri
-BEGIN PGP SIGNED MESSAGE-
Hash: SHA1
According to Eric Blake on 9/22/2009 3:02 PM:
> I've got a patch in testing for both of these issues.
Does this look okay to apply? The fix in path.cc affects more than just
link, hence I had to add a new option to keep mkdir("d/
x27;t encountered anything it broke, but
several things (like link("file","missing/")) that were fixed. But I'm
not completely positive how symlinks to multi-level directories will
behave, so I'm testing it further, and may need yet another tweak.
- --
Don't work
elf-contained,
and make cygwin more consistent with Linux. For example:
touch a
ln -s c b
link a b/
should fail because b/ is not an existing directory, but without this
patch, it succeeds and creates the regular file c as a link to a.
2009-09-24 Eric Blake
* syscalls.cc (link)
-BEGIN PGP SIGNED MESSAGE-
Hash: SHA1
According to Eric Blake on 9/3/2009 9:58 AM:
> Second, it is not performing the proper checks when AT_EACCESS is set and the
> applications' effective id differs from the real id (to fix this would
> require
> adding a parameter
t fchmodat(,AT_SYMLINK_NOFOLLOW) (aka BSD lchmod), so if/until we
implement lchmod, we should not mistakenly change the permissions on the
file the symlink is pointing to. I've also posted a newlib patch to
declare e[uid]access.
2009-09-24 Eric Blake
* syscalls.cc (fchownat): lchmod is not
mp; 0212-1990)" and have your XP installation disk ready.
Let's document this.
2009-09-24 Eric Blake
* setup2.sgml (setup-locale-problems): Document how to install
non-default charsets.
- --
Don't work too hard, make some time for fun as well!
Eric Blake
-BEGIN PGP SIGNED MESSAGE-
Hash: SHA1
According to Corinna Vinschen on 9/25/2009 4:06 AM:
>>> 2009-09-24 Eric Blake
>>>
>>> * setup2.sgml (setup-locale-problems): Document how to install
>>> non-default charsets.
>> Shoot.
>
>
- --
Don't work too hard, make some time for fun as well!
Eric Blake e...@byu.net
-BEGIN PGP SIGNATURE-
Version: GnuPG v1.4.9 (Cygwin)
Comment: Public key at home.comcast.net/~ericblake/eblake.gpg
Comment: Using GnuPG with Mozilla - http://enigmail.mozdev.org/
iEYEA
look now?
>> >- char new_buf[strlen (newpath) + 5];
>> >+ if (isdirsep (newpath[nlen - 1]) || has_dot_last_component (newpath,
>> >false))
>
> Same observation: Couldn't newpath be zero length?
Also fixed below; actually this one is fixed by using PC_NULL
ch to new-features.sgml and tweaking
unistd.h in newlib?
P.S Any reason that "dtable.h" and "cygheap.h" aren't self-contained?
2009-09-26 Eric Blake
* exec.cc (fexecve): New function.
* cygwin.din (execvpe, fexecve): Export new fexecve and existing
her
cygheap.h should include these files itself, so that clients don't have to
do it. But I'm not too worried - there aren't that many clients.
- --
Don't work too hard, make some time for fun as well!
Eric Blake e...@byu.net
-BEGIN PGP SIGNATURE-
Ver
-BEGIN PGP SIGNED MESSAGE-
Hash: SHA1
According to Christopher Faylor on 9/26/2009 8:57 AM:
>> But how does it look now?
>
> It looks good. Thanks. Please check in.
I missed one corner case in my testing; how about this followup?
2009-09-29 Eric Blake
*
-BEGIN PGP SIGNED MESSAGE-
Hash: SHA1
According to Corinna Vinschen on 9/29/2009 1:35 PM:
> On Sep 29 13:18, Eric Blake wrote:
>> I missed one corner case in my testing; how about this followup?
>>
>> 2009-09-29 Eric Blake
>>
>> * syscalls.cc (
a=blob;f=tests/test-rename.h
2009-09-30 Eric Blake
* path.cc (has_dot_last_component): Detect "a/.//".
- --
Don't work too hard, make some time for fun as well!
Eric Blake e...@byu.net
-BEGIN PGP SIGNATURE-
Version: GnuPG v1.4.9 (Cygwin)
Comment: Publi
7;t test for backslash path separator since that's a Win32
> path following Win32 rules. */
> const char *last_comp = strrchr (dir, '\0');
Looked like a decent rewrite to me, except why did you use strrchr instead
of strchr to find the end of the string?
- --
Don&#x
he latest patch, but it
was kind of kind of convenient being able to build under 1.5, as it meant
fewer cygwin 1.7 processes to stop before installing a just-built dll.
- --
Don't work too hard, make some time for fun as well!
Eric Blake e...@byu.net
-BEGIN PGP SIGNATURE-
V
-BEGIN PGP SIGNED MESSAGE-
Hash: SHA1
According to Eric Blake on 10/7/2009 9:04 PM:
> I haven't spent time trying to locate where the leak is happening, but
> process explorer confirms that this STC leaves a handle open to the file,
> preventing further re-creation of a n
vailable from Windows.
>
> Why not send these type of musings to the cygwin-developers list? It really
> is more appropriate for this type of discussion.
Sorry about the wrong list. At any rate, what about this patch?
2009-10-08 Eric Blake
* hires.h (hires_ms): Ch
two values:
tv_nsec / 100 (nsecs) - scaling down
tv_nsec * 10 (usecs) - scaling up
so that x will be in terms of 100ns ticks. The relations should be:
/ 100 = 10/NSPERSEC = 10/1000
* 10 = NSPERSEC/100 = 1000/1000000
since NSPERSEC falls in between nanoseconds a
-BEGIN PGP SIGNED MESSAGE-
Hash: SHA1
According to Eric Blake on 10/9/2009 6:45 AM:
> OK, here's the respin without the churn.
>
>> It looks like you either don't need the systime() call or it should
>> call systime_ns.
>
> Done. hires_us still uses s
e open.
That way, the clock_gettime() is as close to the assignment as possible,
rather than appearing to be early by the amount of time spent opening the
handle. But with that change backed out, the patch is now in.
- --
Don't work too hard, make some time for fun as well!
Eric Bla
tin_va_end() is a no-op
for x86?]
- --
Don't work too hard, make some time for fun as well!
Eric Blake e...@byu.net
-BEGIN PGP SIGNATURE-
Version: GnuPG v1.4.9 (Cygwin)
Comment: Public key at home.comcast.net/~ericblake/eblake.gpg
Com
ever ported to a system where va_end is more
>> than a no-op? [At least, I'm assuming that __builtin_va_end() is a no-op
>> for x86?]
>
> That's probably a good idea, given that POSIX requires the usage of
> va_end. PTC?
Sure, I'll do an audit.
- --
Don&#
-BEGIN PGP SIGNED MESSAGE-
Hash: SHA1
As promised:
2009-10-13 Eric Blake
* external.cc (cygwin_internal): Use va_end.
* fork.cc (child_copy): Likewise.
* libc/bsdlib.cc (warn, warnx, err, errx): Likewise.
* pinfo.cc (commune_request): Likewise
ill
default it to empty, which can be changed by either 'make DESTDIR=...' or
'env DESTDIR=... make -e'.
- --
Don't work too hard, make some time for fun as well!
Eric Blake e...@byu.net
-BEGIN PGP SIGNATURE-
Version: GnuPG v1.4.9 (Cygwin)
Comment: Publi
e thinks it can be fixed. Complaining on the
subscriber-only mingw is the wrong approach to get autoconf to even be
aware of his complaint or ideas for improvement (not that me replying on
the cygwin-patches list is any better ;)
- --
Don't work too hard, make some time for fun as well!
Eric Blake
orward to cygwin/sysinfo.h, rather than directly declaring the two functions?
- --
Don't work too hard, make some time for fun as well!
Eric Blake e...@byu.net
-BEGIN PGP SIGNATURE-
Version: GnuPG v1.4.9 (Cygwin)
Comment: Public key at home.comcast.net/~ericblake/eblak
ot;a=b", "", 0) == -1); // and here
assert (errno == EINVAL);
errno = 0;
assert (setenv (NULL, "", 0) == -1);
assert (errno == EINVAL); // and here
return 0;
}
2009-11-16 Eric Blake
* environ.cc (setenv): Detect invalid argument.
(uns
leep (which fails), then checking the current time
to see if enough elapsed time has occurred.
2009-11-18 Eric Blake
* signal.cc (nanosleep): Support 'infinite' sleep times.
(sleep): Avoid uninitialized memory.
- --
Don't work too hard, make some time for fun as wel
-BEGIN PGP SIGNED MESSAGE-
Hash: SHA1
According to Christopher Faylor on 11/18/2009 1:47 PM:
> On Wed, Nov 18, 2009 at 01:13:53PM -0700, Eric Blake wrote:
>> 2009-11-18 Eric Blake
>>
>> * signal.cc (nanosleep): Support 'infinite' sleep times.
>&
-BEGIN PGP SIGNED MESSAGE-
Hash: SHA1
According to Eric Blake on 11/20/2009 7:15 AM:
>>> * signal.cc (nanosleep): Support 'infinite' sleep times.
>>> (sleep): Avoid uninitialized memory.
>> Sorry but, while I agree with the basic idea, thi
y likely
that someone does 'sleep 50d', and the 100% CPU utilization bug in that
case is not a regression since cygwin 1.5 did the same.
- --
Don't work too hard, make some time for fun as well!
Eric Blake e...@byu.net
-BEGIN PGP SIGNATURE-
Version: Gnu
were to move to git, in which case the git-merge-changelog
program from gnulib almost always merges ChangeLog entries exactly how I
was expecting. Maybe I should go ahead and ITP that as a standalone program.
- --
Don't work too hard, make some time for fun as well!
Eric Blake e..
lusion of this patch.
>
> So, thanks for the patch and please check in. This will then go into 1.7.2.
Thanks; committed now.
- --
Don't work too hard, make some time for fun as well!
Eric Blake e...@byu.net
-BEGIN PGP SIGNATURE-
Version: GnuPG v1.4.9 (Cygwin)
Comment:
-BEGIN PGP SIGNED MESSAGE-
Hash: SHA1
According to Eric Blake on 12/17/2009 7:04 AM:
> According to Christopher Faylor on 12/16/2009 9:08 AM:
>>> How about the following, then? Same changelog.
>> It wonder if your while (!done) loop could be expressed as a for loop but
2009-12-26 Eric Blake
* include/pty.h (openpty, forkpty): Mark last two arguments const,
to match glibc 2.8.
* libc/bsdlib.cc (openpty, forkpty): Likewise.
- --
Don't work too hard, make some time for fun as well!
Eric Blake e...@byu.net
-BEGI
should have received a copy of the GNU General Public License
along with this program. If not, see <http://www.gnu.org/licenses/>. */
/* Written by Eric Blake , 2009,
and Bruno Haible , 2009. */
#include
#include
#include
#include
#include
#include
#include
#define ASSER
as open. So even though mq_open does not leak an fd to the
current process, it should pass O_CLOEXEC as part of its internal open()
call in order to avoid leaking the fd to unrelated child processes.
--
Don't work too hard, make some time for fun as well!
Eric Blake e...@byu
use http://cygwin.com/acronyms/#SHTDI
> and I have only so much time. I'm just trying to lay the groundwork
> here. I'm not at all opposed to patches adding the still missing
> functionality.
Fair enough.
--
Don't work too hard, make some time for fun as well!
Eric Blake e...@byu.net
signature.asc
Description: OpenPGP digital signature
nteractive "*P")
(and (not (eq major-mode 'fundamental-mode))
(copyright-update arg))
nil)
(add-hook 'before-save-hook 'my-copyright-update)
--
Eric Blake ebl...@redhat.com+1-801-349-2682
Libvirt virtualization library http://libvirt.org
signature.asc
Description: OpenPGP digital signature
101 - 162 of 162 matches
Mail list logo
