Re: Checking for specific error messages (was: Re: svn trunk r16292: FAIL (x86_64-unknown-linux-gnu shared ra_dav bdb))

"Peter N. Lundblad" <peter@famlundblad.se> writes:
> While looking into this, I checked a lot of other invocations of
> run_and_verify_svn, where an error is expected. Many (most?) places just
> use SVNAnyOutput, and a few uses a specific error message. The former way
> has the benefit that it that it doesn't break if an error message is
> changed (or different in some RA/FS implementation), while it obviously
> doesn't really check that the acutal error was the expected. Does anyone
> have a good advise on how to handle this in general? Maybe we should have
> a way to test for specific error codes, not just specific strings. That
> might be useful for scripts as well (i.e. an option to always output error
> codes with error messages.)

I think it's fine to test for specific error strings -- the
SVNAnyOutput stuff is just laziness, or sometimes, it reflects the
fact that that particular error isn't what the test is mainly about,
so the test author decided not to care about the error string.

We have automated test runs so we can catch it when an error string
changes, it's no big deal IMHO.

-Karl

-- 
www.collab.net  <>  CollabNet  |  Distributed Development On Demand
---------------------------------------------------------------------
To unsubscribe, e-mail: dev-unsubscribe@subversion.tigris.org
For additional commands, e-mail: dev-help@subversion.tigris.org