Re: [gcc] FYI, libffi FAILs with cygwin snapshot 20100205, 20100207 & 20100210...

[Marco Atzeri <marco_atzeri at yahoo dot it>]

--- Gio 11/2/10, Corinna Vinschen  ha scritto:

> On Feb 11 17:04, Corinna Vinschen
> wrote:
> > On Feb 11 15:39, Marco Atzeri wrote:
> > > --- Gio 11/2/10, Corinna Vinschen ha scritto:
> > > > On Feb 11 07:40, Christian Joensson
> > > > wrote:
> > > > > >> Test Run By chj on Wed Feb 10
> 11:39:41 2010
> > > > > > 115,118c115,118
> > > > > > < 7 8. 9 1 9. 3: 8 17. 12
> > > > > > < res: 8 17. 12
> > > > > > < 7 8. 9 1 9. 3: 8 17. 12
> > > > > > < res: 8 17. 12
> > > > > > ---
> > > > > >> 7 8 9 1 9 3: 8 17 12
> > > > > >> res: 8 17 12
> > > > > >> 7 8 9 1 9 3: 8 17 12
> > > > > >> res: 8 17 12
> > > > > >
> > > > > >
> > > > > > Note the crept in "." (dot) which
> is symptomatic
> > > > for the situation...
> > > > > > if this rings a bell in anyone's
> ear?
> > > > > 
> > > > > well, maybe this never shows up on
> cygwin developers'
> > > > list.. but
> > > > > 
> > > > > 20100204 works... 20100205 doesn't...
> > > > 
> > > > That's the same observation Marco made in
> > > > http://cygwin.com/ml/cygwin/2010-02/msg00257.html
> > > > and it points to a problem in the new,
> multibyte-aware
> > > > regex imported
> > > > from FreeBSD.
> > > 
> > > Autom4te::Channels::msg('obsolete.'
> > >? ? ? ? ? ? ?
> ? ? ? ? ? ? ? ?
> ? ? ^
> > > also in my case a extra dot is the likely
> problem.
> > 
> > Ok, but when and where did it creep in?? What's
> the original search
> > pattern, what's the string to search?? And at
> which point did the dot
> > show up?

unfortunately I don't speak perl and aclocal is a really
massive script so I have no clue how catch the failing pattern.


> 
> For a start, I just applied a patch to the new regex
> functions.? I used
> the test file from Henry Spencer together with the test
> application
> from glibc, and the results were hair-raising.? I did
> practically
> nothing else today but to debug this and to minimize the
> problems with
> this regex implementation, especially in the multibyte
> cases.? AFAICS
> there are no really bad problems left.? Makes me
> wonder if this code was
> really tested at all in FreeBSD :(
> 

Thanks,
waiting for the next snapshot to test it

> 
> Corinna

Marco





--
Problem reports:       http://cygwin.com/problems.html
FAQ:                   http://cygwin.com/faq/
Documentation:         http://cygwin.com/docs.html
Unsubscribe info:      http://cygwin.com/ml/#unsubscribe-simple