Hi Peter, Andy, Thanks a lot for your responses. I will investigate this further based on your feedback / suggestions. Will get back to the list in case I learn something more about this problem.
Thank you once again. And I wish everyone a Very Happy New Year. ~ Urjit On 12/29/07, Andy Polyakov <[EMAIL PROTECTED]> wrote: > > Peter, > > > Sorry, I finally caught on, I build with xlc, not gcc on AIX. Which at > > least explains why I've never hit this on AIX. > > > > I do however build with gcc on PPC Linux on SLES-8 which uses gcc > > 3.(something). So yes, an "as" bug seems most likely since this doesn't > > cause problems on Linux. > > > > All I can suggest is that you get the sources for the as that came with > > SLES-8 (www.suse.de will have those) and try to build it on AIX. > > Well, question then is if GNU assembler has full and working support for > AIX 64-bit XCOFF executable format? Is it command line compatible with > AIX as, so that it can act as drop-in replacement? I'm not maintaining > that I know answer to these questions (I do not!), I'm only saying that > if you want to back-up this suggestion, then you should be prepared to > follow up such questions ;-) > > Urjit, > > Do note that assembler is called *all the time*! I mean when compiling > .c file gcc simply generates an .s file and then invokes the assembler, > naturally with -u flag. In other words there is far more evidence that > as command does not hang when -u flag is provided. Or in yet other words > your focus is likely to be wrong: it's unlikely to be -u flag per se, > but rather when gcc attempts to process .s file [without having > generated it itself]. > > > As as is internally called by gcc, how can I tell gcc to call this > command > > without -u? Does it even make sense (as calling as without -u may result > in > > build failures in other components)? > > "Exactly" to remark in last question. You should not fight the -u flag, > but something else. It can be some subtle bug in gcc compiler driver or > AIX libc... > > > Has anyone seen this before? > > I recall similar problem being reported on openssh-unix-dev list. A > while ago... > > > Do you guys think there is some problem with the assembler? > > Do you think I require a patch for the assembler? If yes, could you > please > > point me to a place from where I can get the patch for the assembler? > > I just want to point that I can't possibly answer such questions, so > don't expect me to. Cheers. A. >
