RE: [openorb-devel] OpenORB compiler seems extremely buggy

[Rumpf M. <Mic...@en...>]

Hi,

The OpenORB compiler suite is known to be sub-optimal ;)

That's the reason why we do not fix compiler related bugs anymore. We
decided to put any efforts (and if available any resource) into a
rewrite of the compilers. Jesper Pedersen has started with the task but
the suite hasn't been finished yet. Due to a general lack of time under
the OpenORB project developers there is currently no resource dedicated
to the compiler suite exclusively. We know that we have many
deficiencies here but for 95% of the users the compilers are working
sufficiently.
However, for OpenORB 1.4.1 and above fixing the compilers should be the
most important task!

It would be great if you could provide us with your test suite so that
we can take it for validating our new compilers.

Have you ever thought about putting the suite on SourceForge for
example: http://corba-idl-test-suite.sf.net ;) ?

Thanks,
   Michael

> -----Original Message-----
> From: ope...@li...=20
> [mailto:ope...@li...] On Behalf=20
> Of Kunal Ashar
> Sent: Monday, April 26, 2004 12:24 PM
> To: ope...@li...
> Subject: [openorb-devel] OpenORB compiler seems extremely buggy
>=20
>=20
> We've been evaluating OpenORB as a potential candidate for a=20
> production ORB. As a first step, we've been putting OpenORB's=20
> compiler through its paces. We've evaluated both OpenORB=20
> 1.3.x and 1.4.x.
> =20
> We ran OpenORB's compiler through 350+ IDLs, of varying=20
> complexities and functionalities. All the IDLs tested OpenORB=20
> compiler's CORBA 2.3 compliance, and ability to handle=20
> complex, production scenarios.
> =20
> OpenORB failed to compile about 15% of the IDLs, (55 of them)=20
> all of them perfectly valid ones.
> =20
> We then ran OpenORB's compiler through 150+ regression IDLs=20
> (those with some errors or non-compliant components in them).=20
> Incredibly, OpenORB actually successfully compiled about 10%=20
> of them!! (20 of them).
> =20
> Turns out that OpenORB is certainly not CORBA compliant as claimed.
> =20
> 15% and 10% are quite high for an ORB that's being evaluated=20
> as a production candidate (other ORBs in the fray are ORBacus=20
> for Java, JacORB, VisiBroker for Java and JavaORB). This is a=20
> pretty serious shortcoming in OpenORB. We haven't even begun=20
> testing OpenORB's run-time capabilities yet.
> =20
> We've given up trying to fix OpenORB. We began fixing it when=20
> we found that OpenORB couldn't handle #if and #elif. We tried=20
> to continue when we found that OpenORB couldn't handle=20
> shorthand typedefs. But 75 issues are far too many.
> =20
> I know I've asked this question before, but is there a=20
> dedicated developer working on OpenORB's compiler? Or, is=20
> there a possibility of purchasing some sort of support for=20
> OpenORB? All my queries to Exolab have either bounced back or=20
> remained unanswered. If there is someone who is fixing/can=20
> fix at least some of the issues, I can provide the IDL test suite.
> =20
> Any help or response will be highly appreciated.
> Thanks,
> Regards,
> Kunal
>=20
> --=20
> ___________________________________________________________
> Sign-up for Ads Free at Mail.com
> http://promo.mail.com/adsfreejump.htm
>=20
>=20
>=20
> -------------------------------------------------------
> This SF.net email is sponsored by: The Robotic Monkeys at ThinkGeek
> For a limited time only, get FREE Ground shipping on all orders of $35
> or more. Hurry up and shop folks, this offer expires April 30th!
> http://www.thinkgeek.com/freeshipping/?cpg=3D12297
> _______________________________________________
> openorb-devel mailing list
> ope...@li...
> https://lists.sourceforge.net/lists/listinfo/openorb-devel
>=20