Re: [Gcl-devel] Re: GCL produces wrong EOL on mingw

[Vadim V. Zhytnikov]

Mike Thomas ?????:
> Hi everyone.
>
> I am not getting these file name errors with CVS Maxima and stable CVS 
> ANSI GCL.  If someone can show me how to reproduce it on W2000 or XP I 
> can have a look.
>
> Are they related to the CLtL1 build or a particular OS?  Is anyone 
> getting them now?

To observe the problem it is necessary to replace

#+:gcl(abs-name (file-namestring abs-dir))
#-:gcl(abs-name (ignore-errors (file-namestring abs-dir)))

in defsystem.lisp

by

(abs-name (ignore-errors (file-namestring abs-dir)))

I just rebuild latest GCL CVS (ANSI) and latest Maxima CVS
on top of it.  With

#+:gcl(abs-name (file-namestring abs-dir))
#-:gcl(abs-name (ignore-errors (file-namestring abs-dir)))

I have clean Maxima build. With

(abs-name (ignore-errors (file-namestring abs-dir)))

I still get a problem:

;    - Compiling module "prerequisites"
;      - Compiling source file 
"C:/msys/1.0/home/vadim/maxima/src/mormac.lisp"
Compiling C:/msys/1.0/home/vadim/maxima/src/mormac.lisp.
End of Pass 1.
End of Pass 2.
OPTIMIZE levels: Safety=0 (No runtime error checking), Space=0, Speed=3
Finished compiling binary-gcl/mormac.o.

;      - Loading binary file "binary-gcl/mormac.o"
Loading binary-gcl/mormac.o
start address -T 10314000 Finished loading binary-gcl/mormac.o

;      - Compiling source file 
"C:/msys/1.0/home/vadim/maxima/src/compat.lisp"
Compiling C:/msys/1.0/home/vadim/maxima/src/compat.lisp.
Error in LET [or a callee]: Cannot create the file 
binary-gcl/C:/msys/1.0/home/vadim/maxima/src/j/compat.data.

Fast links are on: do (use-fast-links nil) for debugging
Broken at CONDITIONS::CLCS-OPEN.  Type :H for Help.
 1 (Continue) Retry opening file 
#p"binary-gcl/C:/msys/1.0/home/vadim/maxima/src/j/compat.data".
 2 (Abort) Return to top level.
dbl:>>

A few observations from my experience with the problem.
It 100% reproducible and compilation always stop at the same file
provided background GCL remains unchange.  It is enough to
rebuild GCL with any tiny modification and error jumps to
other file.


> The only serious bug I can report at present is the ANSI test crash 
> which I have tried to isolate a couple of times including this morning 
> without success.  I'm happy to ignore it for 2.6.2 unless someone can 
> find a quick fix, especially as the HEAD branch exhibits different, less 
> severe, problems in the test.
>
> As to whether it is related to CT/LF issues , who would know.  I would 
> expect it to be one off instead of two.
>
> On Monday I intend checking in the changes made during the installer 
> saga, putting out a test package and then crossing my fingers for a 
> merciful end to 2.6.2.
>
> Cheers
>
> Mike Thomas.
>
>
> Camm Maguire wrote:
>
> > Greetings, and thanks for the pointer!  Mike, is this still extant?
> > Can you link this to anything else at all, especially a call to GBC?
> > Since the overlap always appears to be 2 chars wide, could this be
> > elated to the cr/lf problems we've been seeing?
> >
> > Take care,
> >
> > "Vadim V. Zhytnikov" <address@hidden> writes:
> >
> >  
> >
> > > Camm Maguire ?????:
> > >   
> > >
> > > > Hi Vadim!  Can you give an example or pointer to a discussion on the
> > > > lists?
> > > >
> > > >     
> > >
> > > Here you are!
> > >
> > > http://www.math.utexas.edu/pipermail/maxima/2003/006442.html
> > > http://www.math.utexas.edu/pipermail/maxima/2004/006552.html
> > >
> > >   
> > >
> > > > Take care,
> > > > "Vadim V. Zhytnikov" <address@hidden> writes:
> > > >
> > > >     


--
     Vadim V. Zhytnikov

      <address@hidden>
     <address@hidden>