On 28 June 2012 17:05, Glenn Fowler <g...@research.att.com> wrote:
>
> On Thu, 28 Jun 2012 12:48:07 +0200 Cedric Blancher wrote:
>> On 27 June 2012 22:15, Cedric Blancher <cedric.blanc...@googlemail.com> 
>> wrote:
>> > On 27 June 2012 22:07, Glenn Fowler <g...@research.att.com> wrote:
>> >>
>> >> On Wed, 27 Jun 2012 21:25:47 +0200 Cedric Blancher wrote:
>> >>> On 27 June 2012 20:12, Glenn Fowler <g...@research.att.com> wrote:
>> >>> >
>> >>> > On Wed, 27 Jun 2012 12:51:50 +0200 Cedric Blancher wrote:
>> >>> >> On 27 June 2012 04:25, Glenn Fowler <g...@research.att.com> wrote:
>> >>> >> >
>> >>> >> > the AT&T Software Technology ast beta 2012-06-26 source and binary 
>> >>> >> > release
>> >>> >> > has been posted to the download site
>> >>> >> >        http://www.research.att.com/sw/download/beta/
>> >>> >> > the package names and md5 checksums are
>> >>> >> >            INIT  904949c5a95c2c75ed543290802c1401
>> >>> >> >        ast-base  b5cc252849e2bc0732fd393fdc47b96c
>> >>> >> >        ast-open  5d7e1a86a578496c775310caff3bf625
>> >>> >> >         ast-gpl  91404fb7954a953772f1a64c99c4b1ee
>> >>> >> >         ast-ksh  94652c804ec2f7d3fda5e8f749145482
>> >>> >> >       ast-jmake  6ccf071941eac512d05530808b4fd42c
>> >>> >> >      ast-sudoku  f425de48d3376ce1cad81fa54a7a091a
>> >>> >> > the md5 sums should match the ones listed on the download page
>> >>> >> > if not then don't download
>> >>> >> >
>> >>> >> > the git source repository will be updated tomorrow
>> >>> >> >
>> >>> >> > if this beta checks out ok in the next 2 days it will become an 
>> >>> >> > official release
>> >>> >> >
>> >>> >> > thanks for all the detailed reports and tests on the ast-user and 
>> >>> >> > ast-developers lists
>> >>> >> >
>> >>> >> > _______________________________________________
>> >>> >> > ast-users mailing list
>> >>> >> > ast-us...@research.att.com
>> >>> >> > https://mailman.research.att.com/mailman/listinfo/ast-users
>> >>> >
>> >>> >> The xargs builtin is no longer working with the print builtin on 
>> >>> >> Fedora Linux:
>> >>> >> ksh -c 'builtin xargs ; ls -1 | xargs print ; print done'
>> >>> >> xargs: print: command not found [No such file or directory]
>> >>> >> xargs: out of space [No such file or directory]
>> >>> >> done
>> >>> >
>> >>> > what is the output of
>> >>> >
>> >>> >        ksh -c 'builtin xargs; print ${.sh.version}; type xargs; xargs 
>> >>> > --?-version'
>> >>> >
>> >>
>> >>> ksh -c 'builtin xargs; print ${.sh.version}; type xargs; xargs 
>> >>> --?-version'
>> >>> Version AJM 93u+ 2012-06-26
>> >>> xargs is a shell builtin
>> >>>   version         xargs (AT&T Research) 2012-04-11
>> >>
>> >> thanks
>> >> I'm trying to narrow this down to a reproducable case
>> >> can you run in a directory that only contains the file "foo"
>> >
>> > mkdir x
>> > cd x
>> > mkdir foo
>> > ksh -c 'builtin print ; builtin xargs ; ls -1 | xargs print ; true'
>> > xargs: print: command not found [No such file or directory]
>> > xargs: out of space [No such file or directory]
>> >
>> >>
>> >> and I should have asked before:
>> >> did you run bin/package make with different CC or CCFLAGS?
>> >> what is the output of
>> >>        probe -l C make cc | grep CC.VERSION.STRING
>> >
>> > ./arch/linux.i386-64/bin/probe -l C make cc | grep CC.VERSION.STRING
>> > probe: cannot generate probe key
>> >
>> > gcc --version
>> > gcc 4.6.2
>> > Copyright (C) 2011 Free Software Foundation, Inc.
>> > This is free software; see the source for copying conditions.  There is NO
>> > warranty; not even for MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.
>> >
>> > Build with bash ./bin/package PACKAGE_OPTIONS=map-libc.
>
>> Any clues what is going wrong?
>
> I have not been able to reproduce, so no, I don't know what's going on
> try a build without PACKAGE_OPTIONS just to rule out map-libc
> if you do it like this you can use the same source tree and not clobber your 
> linux.i386-64 build
> --
> hosttype=linux.i386-64-nomap
> cd $PACKAGEROOT
> mkdir arch/$hosttype
> bin/package make HOSTTYPE=$hosttype
> bin/package use HOSTTYPE=$hosttype
> # the last command puts you in an interactive $hosttype ksh with PATH 
> pointing to $hosttype #
> ksh -c 'builtin xargs ; ls -1 | xargs print ; print done'
> --
>
> I'm also sending you off-list a linux.i386-64 ksh built with gcc-4.6.3 and no 
> special options
> this works on a fedora core 15 linux.i386-64

We've figured the problem out: The builtin xargs did not change but I
mistook a Solaris box for a Fedora one because the admins here keep
changing and reusing the hostnames for unknown reasons.
On Solaris the builtin xargs will not fail because it has a
/usr/bin/print, which, per ldd, uses libshell.so.1 and seems to be
using the ksh builtin print.

Mystery solved.

The real bug however is that sh_run() used by b_xargs() can't use
shell builtins.

David, is there a reason for that?

Ced

PS: CC: ast-developers@research.att.com since this has become a developer issue
-- 
Cedric Blancher <cedric.blanc...@googlemail.com>
Institute Pasteur
_______________________________________________
ast-developers mailing list
ast-developers@research.att.com
https://mailman.research.att.com/mailman/listinfo/ast-developers

Reply via email to