>Few minutes ago the issue came up in irc://irc.freenode.org/#opensolaris
>why isaexec(1) does not use the isaexec(3C) function.
>Attached is a patch
>("solaris_usr_lib_isaexec_uses_isaexec_3c_function.diff.txt") which
>fixes the problem.
>
>* Benefits of the patch:
>- /usr/lib/isaexec will be smaller and slightly faster
>- Replacing isaexec(3C) using LD_PRELOAD&co. will now affect
>/usr/lib/isaexec , too (=consistency) ...
>
>* Known problems:
>- Patch is more or less untested
>- isaexec(1) is slightly more verbose (one additional error message) -
>the question is what's better: Reduction in binary size or more verbose
>error messages. If the second option is selected then I propose to add a
>comment to the isaexec(1) source code why it is kept seperate from
>isaexec(3C).Well, I can give you one guess: the isaexec(1) utility was created before the isaexec(3C) library interface. One change that would be required to this code is that you call basename() and not create a secondary copy "my_basename()". (basename lives in libc and not libgen) And the code would need to be C-style clean... Casper _______________________________________________ opensolaris-discuss mailing list opensolaris-discuss@opensolaris.org
