This is the mail archive of the
mailing list for the Mauve project.
Re: can Mauve help me test a PersonalJava implementation?
- From: Steven Schwarz <schwarz at solekai dot com>
- To: Tom Tromey <tromey at redhat dot com>
- Cc: mauve-discuss at sources dot redhat dot com,Steven Schwarz <schwarz at solekai dot com>
- Date: Mon, 02 May 2005 09:44:48 -0700 (PDT)
- Subject: Re: can Mauve help me test a PersonalJava implementation?
- Reply-to: Steven Schwarz <schwarz at solekai dot com>
Thanks for the reply.
I have already produced a version of the SimpleTestHarness that will
run on a 1.1.8 VM. Then there are 39 1.1-tagged test classes that I
have identified as needing repair actually to run on a 1.1.8 VM. I
have so far made repairs for 2 and have in mind doing a good number of
the 39. But there may be some test classes that I will simply not
choose to utilize, those that make use of the AWT Robot class being
definite candidates for that list. It might just be better to change
their tagging in my opinion.
When I am done, if there is interest in having this work committed,
I'll certainly be happy to make it available.
With regard to my question about history, I know about the -D option,
but my question was whether the current repository has revisions that
go back that far (4 years is an eternity in many development projects,
and many have to make compromises that "lose history" for various
reasons over such a long span).
On 30-Apr-2005 Tom Tromey wrote:
>>>>>> "Steven" == Steven Schwarz <email@example.com> writes:
> Steven> Of these, I only had success with JDK 1.4.2. All of the others failed
> Steven> to compile the SimpleTestHarness, as well as some of the classes
> Steven> involved in various individual tests.
> The intent was always that this would work properly -- that you could
> compile the 1.x subset of Mauve with a 1.x compiler and runtime, and
> then try it.
> However, this has obviously bit-rotted, probably because nobody really
> uses anything < 1.4 any more.
> Patches to fix this are welcome ... :-)
> But of course we can't promise it won't break again unless this is
> done with some regularity.
> Steven> * With respect to this vendor's claim, it appears to be made in
> Steven> marketing literature dated 2001. Is it possible to go back to the
> Steven> sources that would have been current in 2001 in the current CVS
> Steven> repository?
> Sure, see the '-D' option to cvs.