From mboxrd@z Thu Jan 1 00:00:00 1970 From: Cyril Hrubis Date: Fri, 13 Aug 2021 16:27:11 +0200 Subject: [LTP] [PATCH 1/3] syscalls/creat08: Convert to new API In-Reply-To: <97c36f43-f567-f384-0c55-4282ed1cd448@suse.cz> References: <20210806154557.19551-1-mdoucha@suse.cz> <97c36f43-f567-f384-0c55-4282ed1cd448@suse.cz> Message-ID: List-Id: MIME-Version: 1.0 Content-Type: text/plain; charset="us-ascii" Content-Transfer-Encoding: 7bit To: ltp@lists.linux.it Hi! > > Btw, it looks that this is one of the test failing on tubmbleweed, it > > looks like the "bin" group is no longer present on the system. > > > > Can't we just pick two non-zero numbers for the groupids instead of > > trying to resolve whatever groups are supposed to be on the system? > > > > We may as well add an library API to request one or two group ids to the > > test library so that they are hardcoded only in a single place. > > > > See: > > https://progress.opensuse.org/issues/96644 > > https://openqa.opensuse.org/tests/1872454 > > The test passes on regular Tumbleweed, the ticket you've linked is about > JeOS. Yes, this needs to be fixed but I'd leave it to a separate > patchset. The rewrite isn't breaking anything that isn't already broken. > But you can drop the CVE test patch and I'll resubmit it with the new > patchset. Fair enough. > Latest Tumbleweed tests: > https://openqa.opensuse.org/tests/1873051#step/open10/8 > https://openqa.opensuse.org/tests/1873529#step/open10/8 > > We can't pick just any group numbers. The test user (nobody) must be a > member of the first group but not the other. Using nobody/nogroup is > probably safe so for the second group we can just scan all existing > groups and pick for example max(gid)+1. I guess that simply git for nobody/nogroup + 1 should work well then. -- Cyril Hrubis chrubis@suse.cz