Re: soft lock in progress -> R port fix

[Chris Kuethe <ckuethe@ualberta.ca>]

On Sun, 24 Aug 2003, Christian Weisgerber wrote:

> Chris Kuethe <ckuethe@ualberta.ca> wrote:
>
> > would an update to math/R be accepted?
>
> Potentially yes.  Fixing breakage is what the next week is all
> about.

ok, i'll roll up something today. it won't be perfect, but it's a
start.

> > it has been suggested that configure should not be patched, rather,
> > configure.ac should be patched and configure be regenerated.
>
> Depends, but yes, non-trivial changes should preferably go into the
> source file.

indeed. this one is just barely non-trivial. i guess i'll fix the .ac
file and regenerate; It's The Right Thing To Do.(tm)

> > http://maths.newcastle.edu.au/~rking/R/devel/03b/0124.html
>
> Looks promising.  Hmm.
>
> DYLIB_EXT=".so.${PACKAGE_VERSION_MAJOR}.${PACKAGE_VERSION_MINOR}"
>
> Not sure the author understands the meaning of library versions...

elsewhere it is suggested that he just leave it at version 0.0 since
they're application plugins, not system libraries.

i'm also getting the symbol size mismatches; i patched ld.so to tell
me if it was a weak symbol or not, and what the size difference is...

/home/ckuethe/R/R-1.7.1/bin/R.bin: /home/ckuethe/R/R-1.7.1/library/ts/libs/ts.so : WARNING: symbol(FortEntries) size mismatch (req 180, size 60) relink your program
/home/ckuethe/R/R-1.7.1/bin/R.bin: /home/ckuethe/R/R-1.7.1/library/ts/libs/ts.so : WARNING: symbol(CallEntries) size mismatch (req 24, size 264) relink your program
/home/ckuethe/R/R-1.7.1/bin/R.bin: /home/ckuethe/R/R-1.7.1/library/ts/libs/ts.so : WARNING: symbol(CEntries) size mismatch (req 180, size 240) relink your program

> > there is a suggestion that there may be a problem with the ieee math
> > library on alphas for which there is another workaround that i am
> > unable to test.
> >
> > http://maths.newcastle.edu.au/~rking/R/devel/03b/0152.html
>
> Actually I understand that to mean that he still had to install a
> modified libm on his system.

that's my take on it too. and i don't have any alphas to test with.

-- 
Chris Kuethe, GCIA CISSP: Secure Systems Specialist - U of A CNS
      office: 157 General Services Bldg.    +1.780.492.8135
              chris.kuethe@[pyxis.cns.]ualberta.ca

     GDB has a 'break' feature; why doesn't it have 'fix' too?