On Mon, 2008-07-21 at 14:00 +0200, Walter Franzini wrote: > The problem is not related to compilation, but to the test suite. It > seems to me that the package is build by an user that belong to the > undefined group 1000 and this make aegis fails numerous tests. Aegis doesn't like unknown users or unknown groups. That has probably happened is the user build bot is logged in as has the default group from the fourth column of the /etc/passwd file as 1000, but that group does not appear in the /etc/group file. A second, less likely, possibility is a file with its group set to a number which does not appear in the /etc/group file. This was once a bigger issue that it is not, because Aegis once stored the _string_ of the GID in the meta-data. We could add a simple program to the start of each test file which checks for thinks Aegis really can't cope with -- like unlisted GIDs, UIDs < 100, etc -- and have the tests fail more gracefully in broken environments. In the test_prelude, actually, which would be good reason to update all of the test scripts to use it. Regards Peter Miller <millerp@canb.auug.org.au> /\/\* http://miller.emu.id.au/pmiller/ PGP public key ID: 1024D/D0EDB64D fingerprint = AD0A C5DF C426 4F03 5D53 2BDB 18D8 A4E2 D0ED B64D See http://www.keyserver.net or any PGP keyserver for public key. "If the programmer can simulate a construct faster than a compiler can implement the construct itself, then the compiler writer has blown it badly." -- Guy Steele
Attachment:
signature.asc
Description: This is a digitally signed message part