RE: libfloss on HP NonStop

[Schmitz, Joachim]

Phew, finally and after many iterations (and email back and forth):
'./configure' is done, 'make' only with the '-ik' options (see below), 'make 
-ki check' just hangs (?!?) after a failed attempt to link test-sigaction (see 
below too).

Attached the patches I had to apply. Not all are clean and nice, some are 
hacks, some just meant to get the darn thing compiled...
It certainly needs more work, and I appreciate any feedback and pointer where 
to go next.

Sill a few remaining issues
set-mode-acl.o and copy-acl.o have unresolved reference to facl. This should be 
solvable with the new and improved ACL support for NonStop in gnulib, we 
discussed recently, and several patches in the attachment are either taken from 
there or just quick hacks to appease the compiler, because I know a better 
version is out there...

test-alignof.c doesn't compile, error message:
  CHECK (char)
  ^
"/usr/local/Floss/testdir/gltests/test-alignof.c", line 40: error(197):
          invalid size for bit field
and similar to the other CHECK(type) stuff in the same file. No idea what that 
error message is trying to tell me, nor what that macro CHECK is actually 
supposed to do.

regex.o has unresolved reference to libintl_gettext, when linking 
test-regex-quote, I believe because it does not link libintl.a (it does indeed 
not, but is that symblen in there?)? Looks like a fix for configure is needed?

test-sigaction.c does not compile, because SA_RESETHAND and SA_RESTART don't 
exist here. Would #define'ing them to 0 help? To something else? To what? 
Where? What else?

As mentioned there is more work  (and internal investigation) to be done, esp. 
for the FILE related stuff. Help and pointers welcome...

For the next step, I'd appreciate a new testdir from the newest gnulib, if 
that's possible?

Bye, Jojo
-----Original Message-----
From: Bruno Haible [mailto:address@hidden 
Sent: Saturday, October 02, 2010 6:06 PM
To: Schmitz, Joachim
Cc: Paul Eggert; Paolo Bonzini; 'bug-gnulib'
Subject: Re: libfloss on HP NonStop

Joachim Schmitz wrote:
> The -lfloss is being taken care of by the cc wrapper script, so 
> nothing you need to be worried about.

Does everyone use this wrapper script? What if someone runs configure with 
CC=c89?

> there are more of these workaround/additions now. They either provide 
> feature NonStop doesn't have otherwise (e.g. adjtime, fsync, the pty 
> family or alloca) or workaround implementation specifics that are just 
> unusual on other platforms (e.g. Block sizes for read, write) or 
> provide some extra features (e.g. the exec family, here we spread the load
> across CPUs) and also covers the slight differences across the different OS 
> releases.     

Can you tell us more about the workarounds here? alloca, fsync, truncate, and 
openpty/forkpty all are already supported by gnulib.

Maybe the simplest start on all this is to evaluate how far gnulib already 
covers the needs on your platform, and which gnulib tests fail.
I've prepared a testdir from today's gnulib, with the proposed changes, for you 
at  http://www.haible.de/bruno/gnu/testdir.tar.gz

What compilation errors, link errors, test failures does it show with
  $ ./configure
  $ make
  $ make check
?

Bruno

testdir.diff
Description: Binary data