This is the mail archive of the
mailing list for the GCC project.
Re: PR java/9157: ICE on malformed ?: condition
- From: Tom Tromey <tromey at redhat dot com>
- To: Ranjit Mathew <rmathew at gmail dot com>
- Cc: GCJ Patches <java-patches at gcc dot gnu dot org>, GCC Patches <gcc-patches at gcc dot gnu dot org>
- Date: 31 Jan 2005 12:16:46 -0700
- Subject: Re: PR java/9157: ICE on malformed ?: condition
- References: <41FE4381.firstname.lastname@example.org>
- Reply-to: tromey at redhat dot com
>>>>> "Ranjit" == Ranjit Mathew <email@example.com> writes:
Ranjit> BTW, how can I add a testcase for this PR - it used
Ranjit> to ICE before and should now fail with an error - there
Ranjit> is no "transition" w.r.t. to the failure status...
This is the meaning of PR 1222 (our oldest PR).
As Andrew points out, this may actually be misguided, as error
messages do change. Though I suppose we could just update the
expected output when we intentionally change a message.
It shouldn't be very hard to implement this. E.g., we could look for
the "shouldfail" (or whatever is appropriate) key in the xfail file,
and then assume the ".out" file holds the required error message (or
perhaps treat it as a regexp or something).
Another thing is, even in the presence of "shouldfail", an ICE should
be a FAIL, not a PASS. I don't know if we currently do this (I know
we don't for jacks, but that is a limitation in the jacks framework).