Don Cragun wrote: > >Glenn Fowler wrote: > >> > >> there is an inconsistency in the implementation of > >> getconf SSIZE_MAX > >> vs other symbols like > >> getconf SIZE_MAX > >> getconf INT_MAX > >> getconf UINT_MAX > >> > >> the inconsistency is that > >> getconf SIZE_MAX > >> returns the actual largest value for size_t, whereas > >> getconf SSIZE_MAX > >> returns the _POSIX_SSIZE_MAX minmax value, as does > >> getconf _POSIX_SSIZE_MAX > >> > >> this shows that getconf is capable of distinguishing between > >> actual and minmax values: > >> getconf AIO_LISTIO_MAX => 4096 > >> getconf _POSIX_AIO_LISTIO_MAX => 2 (the minmax value) > >> > >> man did not reveal a special case for SSIZE_MAX > >> > >> for consistency "getconf SSIZE_MAX" should return the > >> largest value for ssize_t > >> > >> any solaris rationale why it doesn't? > > > >CC:'ing Don Cragun&&April for help... > > > >Don/April: Is it possible that this is a bug in Solaris /usr/bin/getconf > >? > > Roland, > Yes. This is a bug in /usr/bin/getconf. I'll file a bug > report this afternoon.
Thanks! (please add a refernce to the CR #6437624 (http://bugs.opensolaris.org/bugdatabase/view_bug.do?bug_id=6437624) in the bug report). Another problem: The ksh93 testsuite has a special test script (usr/src/lib/libshell/common/tests/sun_solaris_getconf.sh) to make sure Solaris /usr/bin/getconf and ksh93's "getconf" builtin are compatible, however currenly the test will fail for "SSIZE_MAX" because there is a difference in the returned values. Can I get an "exception" from you (I guess you have the authority for this since you're Sun's standards guru... :-) ) to modify "sun_solaris_getconf.sh" and filter the SSIZE_MAX test for now, please (otherwise "sun_solaris_getconf.sh" fails and blows-up the whole test run... ;-( ) ? ---- Bye, Roland -- __ . . __ (o.\ \/ /.o) roland.mainz at nrubsig.org \__\/\/__/ MPEG specialist, C&&JAVA&&Sun&&Unix programmer /O /==\ O\ TEL +49 641 7950090 (;O/ \/ \O;)
