LSB checkers updated
"Wichmann, Mats D" <[email protected]>
| Newsgroups | gmane.linux.lsb.test-suite |
|---|---|
| Message-ID | <A06801158AE07847B27A52C1A074BC1D08282340@fmsmsx404.amr.corp.intel.com> |
I did some cleanup work on the checkers the past few days, closing bugs 735 (accept standard help arguments - by adding common argument handling to all the checkers and updating all the manpages, as well as creating one for lsbcmdchk which was missing a manpage - an additional bug which was opened and closed), and 984 (pkgchk not logging header check errors). This did advance the state of play, but there are still several issues that need work: Runtime-platform checkers (libchk, cmdchk): * All plannned cmdchk work is done * Libchk is still failing on 18 virtual table entries over two classes on six of the seven platforms; the seventh is behaving differently and we don't know why. In any case, an answer is needed for the 18 fails. Last time it was suggested to just remove the tests, but it turns out from the way the data is generated out of the database for libchk to test, it's not trivial to just remove. Also we need to explore the difference on the 7th platform (ppc64). Application checkers (appchk, pkgchk, archk): * archk is not needed for certification today. Needs to understand LSB source-only interfaces and not flag them as errors. * appchk has only one current open issue, that's a couple of new failures on ia64 that were exposed when a previous fix was applied. * pkgchk is now doing what it's supposed to, but as a result is flagging lots of failures. We need to dig deeper and decide which of these are real failures and which should not be reported as such.