segfault on 32bit

segfault on 32bit

[Marco Atzeri]

[-- Attachment #1: Type: text/plain, Size: 773 bytes --]

while trying to build the 32bit version on python,
I hit this never seen before issue:

checking for %zd printf() format support... make: *** No targets 
specified and no makefile found.  Stop.
*** ERROR: make failed
       4 [main] sh 18479 D:\cygwin32\bin\sh.exe: *** fatal error in 
forked process - WFSO timed out after longjmp
       4 [main] sh 18479 D:\cygwin32\bin\sh.exe: *** fatal error in 
forked proces
s - WFSO timed out after longjmp
   72365 [main] sh 18479 cygwin_exception::open_stackdumpfile: Dumping 
stack trac
e to sh.exe.stackdump
   72365 [main] sh 18479 cygwin_exception::open_stackdumpfile: Dumping 
stack trac
e to sh.exe.stackdump


$ uname -svr
CYGWIN_NT-10.0-WOW 3.1.4(0.340/5/3) 2020-02-19 08:45

what is a WFSO ?
Attached stackdump.

Marco



[-- Attachment #2: sh.exe.stackdump --]
[-- Type: text/plain, Size: 120 bytes --]

Stack trace:
Frame     Function  Args
006DC438  61084F28 (00000148, 00000000, 00000000, 00000018)
End of stack trace

Re: segfault on 32bit

[Csaba Ráduly]

On 07/04/2020 19:33, Marco Atzeri via Cygwin wrote:
> while trying to build the 32bit version on python,
> I hit this never seen before issue:
> 
> checking for %zd printf() format support... make: *** No targets specified and 
> no makefile found.  Stop.
> *** ERROR: make failed
>        4 [main] sh 18479 D:\cygwin32\bin\sh.exe: *** fatal error in forked 
> process - WFSO timed out after longjmp
>        4 [main] sh 18479 D:\cygwin32\bin\sh.exe: *** fatal error in forked proces
> s - WFSO timed out after longjmp
>    72365 [main] sh 18479 cygwin_exception::open_stackdumpfile: Dumping stack trac
> e to sh.exe.stackdump
>    72365 [main] sh 18479 cygwin_exception::open_stackdumpfile: Dumping stack trac
> e to sh.exe.stackdump
> 
> 
> $ uname -svr
> CYGWIN_NT-10.0-WOW 3.1.4(0.340/5/3) 2020-02-19 08:45
> 
> what is a WFSO ?

WaitForSingleObject, presumably.

Csaba
-- 
You can get very substantial performance improvements
by not doing the right thing. - Scott Meyers, An Effective C++11/14 Sampler
So if you're looking for a completely portable, 100% standards-conformat way
to get the wrong information: this is what you want. - Scott Meyers (C++TDaWYK)

Dropping 32 bit Cygwin? (was Re: segfault on 32bit)

[Corinna Vinschen]

[-- Attachment #1: Type: text/plain, Size: 1860 bytes --]

On Apr  7 22:27, Csaba Ráduly via Cygwin wrote:
> On 07/04/2020 19:33, Marco Atzeri via Cygwin wrote:
> > while trying to build the 32bit version on python,
> > I hit this never seen before issue:
> > 
> > checking for %zd printf() format support... make: *** No targets
> > specified and no makefile found.  Stop.
> > *** ERROR: make failed
> >        4 [main] sh 18479 D:\cygwin32\bin\sh.exe: *** fatal error in
> > forked process - WFSO timed out after longjmp
> >        4 [main] sh 18479 D:\cygwin32\bin\sh.exe: *** fatal error in forked proces
> > s - WFSO timed out after longjmp
> >    72365 [main] sh 18479 cygwin_exception::open_stackdumpfile: Dumping stack trac
> > e to sh.exe.stackdump
> >    72365 [main] sh 18479 cygwin_exception::open_stackdumpfile: Dumping stack trac
> > e to sh.exe.stackdump
> > 
> > 
> > $ uname -svr
> > CYGWIN_NT-10.0-WOW 3.1.4(0.340/5/3) 2020-02-19 08:45
> > 
> > what is a WFSO ?
> 
> WaitForSingleObject, presumably.

Right.  This looks like a fork problem.  It's 32 bit, so we have to
expect fork errors more often.

Given how much the distro has grown, and given that the number of 32 bit
systems is constantly shrinking, we should really start to think about
dropping 32 bit support.

Most users these days use 64 bit systems.  Some use 32 bit Cygwin on 64
bit systems, which doesn't really make sense, except, maybe, to build 32
bit packages.  These users can easily switch to 64 bit Cygwin.

The access statistics indicate that real 32 bit systems only make less
than 2% of installed systems.  It's a lot of hassle to keep 32 bit
Cygwin running for only a minor installation base.

I'm not saying that we switch off 32 bit support right now, but maybe
2020 should be the last year of 32 bit support.


Corinna

-- 
Corinna Vinschen
Cygwin Maintainer

[-- Attachment #2: signature.asc --]
[-- Type: application/pgp-signature, Size: 833 bytes --]

Re: Dropping 32 bit Cygwin? (was Re: segfault on 32bit)

[Corinna Vinschen]

[-- Attachment #1: Type: text/plain, Size: 1041 bytes --]

On Apr  8 10:34, Corinna Vinschen wrote:
> Right.  This looks like a fork problem.  It's 32 bit, so we have to
> expect fork errors more often.
> 
> Given how much the distro has grown, and given that the number of 32 bit
> systems is constantly shrinking, we should really start to think about
> dropping 32 bit support.
> 
> Most users these days use 64 bit systems.  Some use 32 bit Cygwin on 64
> bit systems, which doesn't really make sense, except, maybe, to build 32
> bit packages.  These users can easily switch to 64 bit Cygwin.
> 
> The access statistics indicate that real 32 bit systems only make less
> than 2% of installed systems.  It's a lot of hassle to keep 32 bit
> Cygwin running for only a minor installation base.
> 
> I'm not saying that we switch off 32 bit support right now, but maybe
> 2020 should be the last year of 32 bit support.

Oh, btw., time_t on 32 bit Cygwin is only 32 bit wide, so the end is
near: 2038-01-19 03:14:07 UTC.


Corinna

-- 
Corinna Vinschen
Cygwin Maintainer

[-- Attachment #2: signature.asc --]
[-- Type: application/pgp-signature, Size: 833 bytes --]

Re: Dropping 32 bit Cygwin? (was Re: segfault on 32bit)

[Sharuzzaman Ahmat Raslan]

Hi Corinna,

If I'm forced to abandon Windows, the next logical option for me is to use
ReactOS, in order to use Cygwin for work.

While it is usable as it is, right now it only has 32 bit, and 64 bit is
still a long way to go.

This is one primary reason that I believe Cygwin should keep the 32bit
version.

Also, as I personally experienced once, the 64bit version does not work
properly and I'm not sure why. It could be the antivirus that was installed
by my company. But, I can quickly continue my work after I install a 32bit
version, and the issue is not there.

So, having 32bit as a fallback is really helpful in that situation.

Thank you.




On Wed, Apr 8, 2020 at 4:45 PM Corinna Vinschen <corinna-cygwin@cygwin.com>
wrote:

> On Apr  8 10:34, Corinna Vinschen wrote:
> > Right.  This looks like a fork problem.  It's 32 bit, so we have to
> > expect fork errors more often.
> >
> > Given how much the distro has grown, and given that the number of 32 bit
> > systems is constantly shrinking, we should really start to think about
> > dropping 32 bit support.
> >
> > Most users these days use 64 bit systems.  Some use 32 bit Cygwin on 64
> > bit systems, which doesn't really make sense, except, maybe, to build 32
> > bit packages.  These users can easily switch to 64 bit Cygwin.
> >
> > The access statistics indicate that real 32 bit systems only make less
> > than 2% of installed systems.  It's a lot of hassle to keep 32 bit
> > Cygwin running for only a minor installation base.
> >
> > I'm not saying that we switch off 32 bit support right now, but maybe
> > 2020 should be the last year of 32 bit support.
>
> Oh, btw., time_t on 32 bit Cygwin is only 32 bit wide, so the end is
> near: 2038-01-19 03:14:07 UTC.
>
>
> Corinna
>
> --
> Corinna Vinschen
> Cygwin Maintainer
> --
> Problem reports:      https://cygwin.com/problems.html
> FAQ:                  https://cygwin.com/faq/
> Documentation:        https://cygwin.com/docs.html
> Unsubscribe info:     https://cygwin.com/ml/#unsubscribe-simple
>


-- 
Sharuzzaman Ahmat Raslan