Re: branch-1-5 UnixWare fixes

[Ralf Wildenhues]

Hi Tim,

* Tim Rice wrote on Mon, Sep 26, 2005 at 08:41:21PM CEST:
> On Mon, 26 Sep 2005, Ralf Wildenhues wrote:
> : * Tim Rice wrote on Sun, Sep 25, 2005 at 10:05:21PM CEST:

> : > I re-ran the tests using ksh88 and all tests passed.
> : > Must be a ksh bug.
> : > Version M-11/16/88h               passes quote test
> : > Version M-12/28/93e-SCO           fails quote test
> : 
> : Please post the failing output of
> :   make check TESTS=quote.test VERBOSE=x
> : 
> : Thank you.
> 
> -----------< begin "make check TESTS=quote.test VERBOSE=x" >------------
> address@hidden 14% gmake check TESTS=quote.test VERBOSE=x

> === Running quote.test
> == compile mode

> = trying: \\ quoting
> = failed: mkdir .libs
>  cc -c "-DVAR=\\test\\" foo.c  -KPIC -DPIC -o .libs/foo.o
>  cc -c "-DVAR=\\test\\" foo.c -o foo.o >/dev/null 2>&1
> = trying: \" quoting
> = failed: mkdir .libs
>  cc -c "-DVAR=\"test\"" foo.c  -KPIC -DPIC -o .libs/foo.o
>  cc -c "-DVAR=\"test\"" foo.c -o foo.o >/dev/null 2>&1
> = trying: \` quoting
> = failed: mkdir .libs
>  cc -c "-DVAR=\`test\`" foo.c  -KPIC -DPIC -o .libs/foo.o
>  cc -c "-DVAR=\`test\`" foo.c -o foo.o >/dev/null 2>&1
> = trying: \$ quoting
> = failed: mkdir .libs
>  cc -c "-DVAR=\$test\$" foo.c  -KPIC -DPIC -o .libs/foo.o
>  cc -c "-DVAR=\$test\$" foo.c -o foo.o >/dev/null 2>&1
*snip*
> FAIL: quote.test

OK.  This is mostly harmless, and we should eventually adjust
quote.test.  I tried to fix this in CVS HEAD with a patch on
2004-12-28 but had to throw that back out on 2005-06-05 because
that broke on other systems' shells.

Thanks for posting. 

> : *snip*
> : > : Content-Description: branch-1-5-uw.patch
*snip*
> : > : >  
> : > : > -sysv4*uw2* | unixware7*)
> : > : > +unixware7*)
> : > : 
> : > : Now, this macro doesn't have a match for sysv4*uw2* any more.  Is this
> : > : intentional?
> : > 
> : > Quite intentional. Look at the case above, it already had sysv4*uw2* so
> : > the one I removed would never have been used anyway. Adding the "pc)"
> : > case to the $host_vendor part did what the other case was supposed to do.
> : 
> : Erm, it had sysv4.2uw2* but not sysv4*uw2*.  If both should be treated
> : similarly here, then you should replace the former with the latter (in
> : the line where sysv4 is also matched)
> 
> Remind me never to take a proof reading job. I guess my brain was focusing
> in the uw2 part.
> All UnixWare 2.x versions (and 1.x) are sysv4.2, so the sysv4*uw2 entry
> becomes redundant.

Well, I _was_ cheating a bit:
http://lists.gnu.org/archive/html/bug-libtool/2005-09/msg00013.html
http://lists.gnu.org/archive/html/libtool-patches/2005-09/msg00036.html
(be sure to look through both threads -- I installed a bogus patch first,
stumbling over this exact same issue ;-)

> : > It could be cleanded up further by having "sysv5* | unixware7*)".
> : > (UnixWare 7 is sysv5)
> : 
> : Hmm, then both of those should be treated similarly, I guess?
> 
> Yes.

OK.

Cheers,
Ralf