bug-bison
[Top][All Lists]
Advanced

[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

Re: compiling bison-3.0 on AIX 5.3


From: Ken Moffat
Subject: Re: compiling bison-3.0 on AIX 5.3
Date: Wed, 16 Oct 2013 19:57:02 +0100
User-agent: Mutt/1.5.21 (2010-09-15)

On Wed, Oct 16, 2013 at 03:06:35PM +0200, Akim Demaille wrote:
> Hi Ken, Hi Michael,
> 
> Le 21 sept. 2013 à 06:03, Ken Moffat <address@hidden> a écrit :
> 
> > On Sat, Sep 21, 2013 at 04:36:38AM +0100, Ken Moffat wrote:
> >> Killed it after more than 200 minutes CPU - memory use hadn't
> >> changed, it was still using as much CPU as it could get, nothing
> >> more in the output.
> >> 
> > Unfortunately, the new build didn't boot - looks as if ext4 isn't
> > supported for '/' by the yaboot bootloader - so I started to back it
> > up.  Discovered that testsuite.dir/at-groups/24-stderr had grown to
> > 2.4G.  The first 19 lines are text, the twentieth line is "extremely
> > long" (perhaps the rest of the file) and (eventually) renders as
> > inverse question-marks (i.e. invalid UTF-8) when I try to browse
> > it in vim.  I've copied the first 1K with 'dd', I'll attach that.
> 
> This bug _might_ have been fixed!  I suspect that
> 
> commit 4c4191cec2b30632a31993a6e46fb241ab9c3434
> Author: Andreas Schwab <address@hidden>
> Date:   Sat Sep 28 00:00:34 2013 +0200
> 
>     location: fix EOF check
>     
>     * location.c (location_caret): Use int, not char, for values from
>     getc.
> 
> explains the problem you are facing.  Could you please give
> a try to this tarball?
> 
>   http://www.lrde.epita.fr/~akim/download/bison-3.0.20-e4678.tar.gz
> 
> or
> 
>   http://www.lrde.epita.fr/~akim/download/bison-3.0.20-e4678.tar.xz
> 
> Thanks a lot!

 For me (linux ppc) it allows the testsuite to complete.

 There is now one failure:
ERROR: 409 tests were run,
1 failed unexpectedly.
23 tests were skipped.

 which is

383: Exception safety api.value.type=variant         FAILED
(c++.at:912)

 I looked for that commit in git, but I can't see it.  Google found
it as a coloured diff, I couldn't see how to get a raw one, so I
applied the one-line change by hand to the 3.0 release.

 There, the testsuite completed with the same error so the commit
has not introduced a new failure.  Things are moving forward.

 I'm attaching a tarball of everything from tests/testsuite.dir/383
in 3.0.20-e4678 except for the 'input' binary.

ĸen
-- 
das eine Mal als Tragödie, dieses Mal als Farce

Attachment: 383.tar.gz
Description: application/tar-gz


reply via email to

[Prev in Thread] Current Thread [Next in Thread]