[gs-bugs] [Bug 691463] various errors and warnings building with DEC Tru64's CC

bugzilla-daemon at ghostscript.com bugzilla-daemon at ghostscript.com
Fri Jul 30 21:41:55 UTC 2010


http://bugs.ghostscript.com/show_bug.cgi?id=691463

--- Comment #10 from Hin-Tak Leung <hintak at ghostscript.com> 2010-07-30 21:41:53 UTC ---
(In reply to comment #7)
> Linux needs
> them and DEC seems to need them removed. Do to the broken test of patch #14 my
> uploads of patch #15 to #17 got also rolled back and so they are not tested
> now.

It is more Marcos's trying put put cflags into part of the compiler command 
- if you do
    ./autogen.sh option1="A B" option2="C D"

and inside ./autogen.sh it says 
    runthis "$@"
depends on what shell you have, it could be this getting passed to the
sub-shell as:
   runthis "option1="A B" option2="C D""
or this:
   runthis "option1=A B option2=C D"
or this:
   runthis option1=A B option2=C D

And the sub-shell can strip quotes as part of its input or not; so the actual
arguments being passed to runthis could be:
  runthis/option1=A/B option2=C/D  (3 arguments)
  runthis/option1=A B option2=C D  (1 argument)
  runthis/option1=A/B/option2=C/D  (4 arguments)

depending on how many level of quotes are stripped and at what stage. Relying
on quotes to protect arguments with spaces just isn't a good idea; and CFLAGS
should be set in CFLAGS rather than in CC in any case.

In fact I think doing:
   option1="A B" option2="C D" ./autogen.sh
might be prefered, as they get passed spaces in those variables intact as
environment variables rather than shell varibles. (but that syntax is
bash/sh-ism, and would break when the user's shell is csh, for example).

-- 
Configure bugmail: http://bugs.ghostscript.com/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are the QA contact for the bug.


More information about the gs-bugs mailing list