my first guess would be to compile with -O disabled

On Mon, 8 Mar 2010 16:06:13 +0100 =?KOI8-R?B?z8zYx8Egy9LZ1sHOz9fTy8HR?= wrote:
> Do you have a patch?
> Does the pipe problem go away if you compile with -std=gnu99 as used
> in 
> http://svn.genunix.org/repos/on/branches/ksh93/gisburn/scripts/buildksh93.sh?

> On Thu, Mar 4, 2010 at 2:22 PM, Dr. Werner Fink <[email protected]> wrote:
> > On Thu, Mar 04, 2010 at 01:54:06PM +0100, Werner Fink wrote:
> >> On Wed, Mar 03, 2010 at 01:20:09AM -0500, Glenn Fowler wrote:
> >> > the AT&T Software Technology ast 2010-03-01 source and binary release
> >> > has been posted to the download site
> >> >         http://www.research.att.com/sw/download/
> >> > the notes and changes link has details on the release
> >>
> >>
> >> Hmmm ... this version shows memory leaks (again) and has
> >> problems with pipes:
> >>
> >> + /usr/src/packages/BUILD/ksh93/arch/linux.i386/bin/ksh shtests
> >> test basic.sh begins at 2010-03-04+11:38:14
> >>         basic.sh[436]: early termination not causing broken pipe
> >>         basic.sh[477]: pipefail causing /bin/true to wait for other end of 
> >> pipe
> >> test basic.sh failed at 2010-03-04+11:38:42 with exit code 2 [ 95 tests 2 
> >> errors ]
> >> test signal.sh begins at 2010-03-04+11:38:42
> >>         signal.sh[55]: pipe with --pipefail PIPE trap hangs
> >>         signal.sh[296]: kill -PIPE $$ failed, required termination by 
> >> signal 'KILL'
> >> test signal.sh failed at 2010-03-04+11:39:57 with exit code 2 [ 26 tests 2 
> >> errors ]
> >> test shcomp-basic.ksh begins at 2010-03-04+11:39:57
> >>         shcomp-basic.ksh[436]: early termination not causing broken pipe
> >>         shcomp-basic.ksh[477]: pipefail causing /bin/true to wait for 
> >> other end of pipe
> >> test shcomp-basic.ksh failed at 2010-03-04+11:40:25 with exit code 2 [ 95 
> >> tests 2 errors ]
> >> test shcomp-signal.ksh begins at 2010-03-04+11:40:25
> >>         shcomp-signal.ksh[55]: pipe with --pipefail PIPE trap hangs
> >>         shcomp-signal.ksh[296]: kill -PIPE $$ failed, required termination 
> >> by signal 'KILL'
> >> test shcomp-signal.ksh failed at 2010-03-04+11:41:41 with exit code 2 [ 26 
> >> tests 2 errors ]
> >> + /usr/src/packages/BUILD/ksh93/arch/linux.i386/bin/ksh 
> >> /usr/src/packages/SOURCES/leak1.sh
> >> [leak1.sh: leak count at 2086]
> >> + /usr/src/packages/BUILD/ksh93/arch/linux.i386/bin/ksh 
> >> /usr/src/packages/SOURCES/leak2.sh
> >> [leak2.sh: leak count at 2548]
> >>
> >> after running those testing scripts two ksh processes remains within
> >> an endless loop:
> >>
> >> 399       8441   99  0.1  3844 1364 1092  20   0 R 1   9:32 
> >> /usr/src/packages/BUILD/ksh93/arch/linux.i386/bin/ksh
> >> 399       7964   95  0.1  3844 1368 1092  20   0 R 0  11:19 
> >> /usr/src/packages/BUILD/ksh93/arch/linux.i386/bin/ksh
> >>
> >> an attached strace shows:
> >>
> >> write(1, "o\nhello\nhello\nhello\nhello\nhello\n"..., 8192) = -1 EPIPE 
> >> (Broken pipe)
> >> --- SIGPIPE (Broken pipe) @ 0 (0) ---
> >> write(1, "o\nhello\nhello\nhello\nhello\nhello\n"..., 8192) = -1 EPIPE 
> >> (Broken pipe)
> >> --- SIGPIPE (Broken pipe) @ 0 (0) ---
> >>
> >> for both processes. Beside this you'll find the scripts leak1.sh and 
> >> leak2.sh
> >> in the attachments.
> >>
> >> Sometimes a further line is shown:
> >>
> >>         shcomp-signal.ksh[296]: kill -VTALRM $$ failed, required 
> >> termination by signal 'KILL'
> >
> >
> > With an older compiler I do not see the problems with the pipes
> > but the memory leaks remain.
> >
> >       Werner
> >
> >
> > --
> >  "Having a smoking section in a restaurant is like having
> >          a peeing section in a swimming pool." -- Edward Burr
> > _______________________________________________
> > ast-developers mailing list
> > [email protected]
> > https://mailman.research.att.com/mailman/listinfo/ast-developers
> >

> -- 
>       ,   _                                    _   ,
>      { \/`o;====-    Olga Kryzhanovska   -====;o`\/ }
> .----'-/`-/     [email protected]   \-`\-'----.
>  `'-..-| /     Solaris/BSD//C/C++ programmer   \ |-..-'`
>       /\/\                                     /\/\
>       `--`                                      `--`
> _______________________________________________
> ast-users mailing list
> [email protected]
> https://mailman.research.att.com/mailman/listinfo/ast-users

_______________________________________________
ast-users mailing list
[email protected]
https://mailman.research.att.com/mailman/listinfo/ast-users

Reply via email to