Thread: CVS HEAD
Hi all, Anyone having trouble running HEAD under cygwin? Specifically, anyone seeing the postmaster SEGV on startup (again, this is with the cvs sources). Seems that (1.5.9-1, at least) cygwin SEGVs after dup()'ing an FD over 130 or so times, as is done by count_usable_fds. [5-liner program confirms] Can I get a confirmation another cygwin install? Cheers, Claudio --- Certain disclaimers and policies apply to all email sent from Memetrics. For the full text of these disclaimers and policies see <a href="http://www.memetrics.com/emailpolicy.html">http://www.memetrics.com/em ailpolicy.html</a>
Claudio,
On Tue, Apr 13, 2004 at 03:25:07PM +1000, Claudio Natoli wrote:
> Anyone having trouble running HEAD under cygwin?
I haven't tried PostgreSQL CVS HEAD, but see below...
> Specifically, anyone seeing the postmaster SEGV on startup (again,
> this is with the cvs sources).
Yes, see below...
> Seems that (1.5.9-1, at least) cygwin SEGVs after dup()'ing an FD over
> 130 or so times, as is done by count_usable_fds. [5-liner program
> confirms]
>
> Can I get a confirmation another cygwin install?
I'm seeing PostgreSQL 7.4.2's postmaster SEGV on startup under 1.5.9-1
too. AFAICT, the stackdump indicates the problem is in the Cygwin DLL.
Unfortunately, due to the following problem:
http://cygwin.com/ml/cygwin/2004-03/msg01368.html
and starting a new job, I have not been able to debug this problem
further. These problems are also preventing me from releasing Cygwin
PostgreSQL 7.4.2 too. :,(
Jason
--
PGP/GPG Key: http://www.tishler.net/jason/pubkey.asc or key servers
Fingerprint: 7A73 1405 7F2B E669 C19D 8784 1AFD E4CC ECF4 8EF6
> I'm seeing PostgreSQL 7.4.2's postmaster SEGV on startup under 1.5.9-1
> too. AFAICT, the stackdump indicates the problem is in the
> Cygwin DLL.
Well, that confirms things. 7.4.2 includes a repeated call to dup(0) to get
an estimate for the max number of fds, whilst 7.4.1 does not.
Code like that below should not SEGV, IMNSHO. Want me to push this over to
the cygwin list?
#include <errno.h>
int main(void)
{
int i;
for (i = 0; ; i++)
{
printf("i = %d\n",i);
if (dup(0) < 0)
{
printf("dup failed\n");
return -1;
}
}
return 0;
}
> and starting a new job,
Congrats.
Cheers,
Claudio
---
Certain disclaimers and policies apply to all email sent from Memetrics.
For the full text of these disclaimers and policies see
<a
href="http://www.memetrics.com/emailpolicy.html">http://www.memetrics.com/em
ailpolicy.html</a>
FYI, similar problem reported, 21 mar 2004, in "How many fds can *you* open?" http://cygwin.com/ml/cygwin/2004-03/msg00984.html But cfg terminated the thread with rudeness. :( --- Claudio Natoli <> wrote: > > > > I'm seeing PostgreSQL 7.4.2's postmaster SEGV on > startup under 1.5.9-1 > > too. AFAICT, the stackdump indicates the problem > is in the > > Cygwin DLL. > > Well, that confirms things. 7.4.2 includes a > repeated call to dup(0) to get > an estimate for the max number of fds, whilst 7.4.1 > does not. > > Code like that below should not SEGV, IMNSHO. Want > me to push this over to > the cygwin list? > > > #include <errno.h> > > int main(void) > { > int i; > for (i = 0; ; i++) > { > printf("i = %d\n",i); > if (dup(0) < 0) > { > printf("dup failed\n"); > return -1; > } > } > return 0; > } > __________________________________ Do you Yahoo!? Yahoo! Tax Center - File online by April 15th http://taxes.yahoo.com/filing.html
Claudio, On Wed, Apr 14, 2004 at 10:55:42AM +1000, Claudio Natoli wrote: > > I'm seeing PostgreSQL 7.4.2's postmaster SEGV on startup under > > 1.5.9-1 too. AFAICT, the stackdump indicates the problem is in the > > Cygwin DLL. > > Well, that confirms things. 7.4.2 includes a repeated call to dup(0) > to get an estimate for the max number of fds, whilst 7.4.1 does not. > > Code like that below should not SEGV, IMNSHO. Want me to push this > over to the cygwin list? > > [snip] Yes, I would really appreciate it. And, thanks for tracking this down. > > and starting a new job, > > Congrats. Thanks. Unfortunately, one of my compromises was significantly less time for open source. :,( Jason -- PGP/GPG Key: http://www.tishler.net/jason/pubkey.asc or key servers Fingerprint: 7A73 1405 7F2B E669 C19D 8784 1AFD E4CC ECF4 8EF6
Claudio,
On Wed, Apr 14, 2004 at 07:46:52AM -0400, Jason Tishler wrote:
> On Wed, Apr 14, 2004 at 10:55:42AM +1000, Claudio Natoli wrote:
> > > I'm seeing PostgreSQL 7.4.2's postmaster SEGV on startup under
> > > 1.5.9-1 too. AFAICT, the stackdump indicates the problem is in
> > > the Cygwin DLL.
> >
> > Well, that confirms things. 7.4.2 includes a repeated call to dup(0)
> > to get an estimate for the max number of fds, whilst 7.4.1 does not.
> >
> > Code like that below should not SEGV, IMNSHO. Want me to push this
> > over to the cygwin list?
> >
> > [snip]
>
> Yes, I would really appreciate it. And, thanks for tracking this down.
The following seems to indicate that the problem has been resolved in
a recent snapshot:
http://cygwin.com/ml/cygwin/2004-04/msg00560.html
Would you be willing to try out the latest one and report back?
Thanks,
Jason
--
PGP/GPG Key: http://www.tishler.net/jason/pubkey.asc or key servers
Fingerprint: 7A73 1405 7F2B E669 C19D 8784 1AFD E4CC ECF4 8EF6
> The following seems to indicate that the problem has been resolved in > a recent snapshot: > > http://cygwin.com/ml/cygwin/2004-04/msg00560.html > > Would you be willing to try out the latest one and report back? [try 2] Sure thing. Just tried with the Apr 16 snapshot, and CVS HEAD looks ok. Cheers, Claudio --- Certain disclaimers and policies apply to all email sent from Memetrics. For the full text of these disclaimers and policies see <a href="http://www.memetrics.com/emailpolicy.html">http://www.memetrics.com/em ailpolicy.html</a>
Claudio, On Sat, Apr 17, 2004 at 03:28:20PM +1000, Claudio Natoli wrote: > > The following seems to indicate that the problem has been resolved > > in a recent snapshot: > > > > http://cygwin.com/ml/cygwin/2004-04/msg00560.html > > > > Would you be willing to try out the latest one and report back? > > Sure thing. Just tried with the Apr 16 snapshot, and CVS HEAD looks > ok. I just tried it last night too and PostgreSQL 7.4.2 (built against cygserver) passed all regression tests. Unfortunately, the Apr 16 snapshot breaks mutt. Sigh... :,( Jason -- PGP/GPG Key: http://www.tishler.net/jason/pubkey.asc or key servers Fingerprint: 7A73 1405 7F2B E669 C19D 8784 1AFD E4CC ECF4 8EF6