[mercury-users] mercury-10.04.1 hard_coded/char_signed test

[Peter Wang]

On 2010-09-11, Keri Harris <keri at gentoo.org> wrote:
> On Friday 10 September 2010 20:15:37 you wrote:
> > 
> > If this is still failing for you, could you please let us know further
> > details of the environment in which you are running it, grade, gcc
> > version, glibc version etc.
> > 
> 
> Initially I ended up not being able to reproduce this. After some 
> investigation I found it's reproducible after the following sequence of 
> events:
> 
> 
> 1) install mercury-10.04 from source tarball
>   ( ./configure && make && make install )
> 
> 2) install mercury-10.04.1 using mercury-10.04 to bootstrap with
>   ( ./configure && make && make install )

Ok, this makes sense if the buggy mercury-10.04 compiler is used to
regenerate char.c -- the buggy char.det_from_int will end up in the
standard library when you install it.  char.c shouldn't need to be
regenerated though, so maybe it's a different (but similar) situation.

Peter
--------------------------------------------------------------------------
mercury-users mailing list
Post messages to:       mercury-users at csse.unimelb.edu.au
Administrative Queries: owner-mercury-users at csse.unimelb.edu.au
Subscriptions:          mercury-users-request at csse.unimelb.edu.au
--------------------------------------------------------------------------