GCC 3.3 status 2004-01-04

John Tytgat John.Tytgat at aaug.net
Mon Jan 5 16:31:23 PST 2004

In message <3FF89EE4.1040509 at dsvr.net>
          Nick Burrett <nick at dsvr.net> wrote:

> John Tytgat wrote:
> > In message <3FF852A8.8010201 at dsvr.net>
> >           Nick Burrett <nick at dsvr.net> wrote:
> > 
> > 
> >>Does anybody have any thoughts or suggestions of other things that need 
> >>to be cleared up or fixed ?
> > 
> > 
> > I have two problems when test driving the current release:
> > 2. When I use Acorn linker instead (and explicitly list the relevant
> >    libraries in its command line), I get an undefined symbol:
> You need to use the -rescan feature of drlink on the Acorn linker 
> command line.

Ok indeed that works.  I should have known that.  Will someone look at
the first reported problem (i.e. the problem when using gcc as linker) ?

Something else : I noticed that the assembler can fail with the following
error :

gcc -c -IOSLib: -ICVSInclude: -mpoke-function-name -mthrowback -pedantic -Wall -Wno-long-long -DJTEXCEPT_WITH_MSGTRANS -o o.drawutils cpp.drawutils
Serious error: As can't write o.drawutils: No such file or directory  at line 0
gccpkg:bin.arm-riscos-aof.3_3_0.as: Aborted
AMU: *** exit (1) ***

When the CSD is set to the directory holding the cpp/o directories, it
works.  I think this is because of reset of Prefix$Dir in 
source/unix/unix.c(_exit) during the first *sub*process exec'ed in gcc
but I'm not sure.

And 'as' is the second (of subsequent) subprocess, it doesn't see the same
CSD anymore.

I changed this in UnixLib in a recent CVS commit but I'm not able to build
gcc nor test it if this really solves the problem.  So can either someone
reproduce the problem with the old code and verify if the latest fix in
UnixLib fixes this problem and/or make a new GCC 3.3.3 test release so
that I can further check it myself please ?

John Tytgat, in his comfy chair at home                                 BASS
John.Tytgat at aaug.net                             ARM powered, RISC OS driven

More information about the gcc mailing list