Keeping Jacks Up-to-date

Keeping Jacks Up-to-date

[Ranjit Mathew]

-----BEGIN PGP SIGNED MESSAGE-----
Hash: SHA1

Hi,

  Does anyone have any objections to moving Jacks to JLS 3rd
Edition?

This would mean:

  1. Updating jacks.html to note the same with a caveat that work is
     still in progress so some tests might be out-of-date.

  2. Updating references to various JLS sections. For example the
     current tests for "jls/definite-assignment/statements/while-
     statements" refer to section 16.2.9, whereas the new JLS has
     it in 16.2.10.

  3. Removing tests that are no longer wrong. For example,
     auto-boxing in JLS 3 invalidates some of the testcases.

  4. Writing tests for the new features, most notably, generics.
     (I see that this is already happening.)

I presume that patches to Jacks ought to be sent to mauve-patches.

Thanks,
Ranjit.

- --
Ranjit Mathew       Email: rmathew AT gmail DOT com

Bangalore, INDIA.     Web: http://ranjitmathew.hostingzero.com/




-----BEGIN PGP SIGNATURE-----
Version: GnuPG v1.4.2 (GNU/Linux)
Comment: Using GnuPG with Thunderbird - http://enigmail.mozdev.org

iD8DBQFDK4dCYb1hx2wRS48RAv/3AJ4hVH7qVyAog1jrKM1k4fQyxCe3WQCeJEBA
AhA7Vpq6x1ASL75MjXQrCwk=
=YclN
-----END PGP SIGNATURE-----

Re: Keeping Jacks Up-to-date

[Tom Tromey]

>>>>> "Ranjit" == Ranjit Mathew <rmathew@gmail.com> writes:

Ranjit> Does anyone have any objections to moving Jacks to JLS 3rd
Ranjit> Edition?

Nope.

Ranjit>   3. Removing tests that are no longer wrong. For example,
Ranjit>      auto-boxing in JLS 3 invalidates some of the testcases.

How about making them conditional so that people can continue to test
1.4 compilers?  I think this is probably worthwhile for the next year
or so.

Tom

Re: Keeping Jacks Up-to-date

[Ranjit Mathew]

-----BEGIN PGP SIGNED MESSAGE-----
Hash: SHA1

Tom Tromey wrote:
>>>>>>"Ranjit" == Ranjit Mathew <rmathew@gmail.com> writes:
> 
> 
> Ranjit> Does anyone have any objections to moving Jacks to JLS 3rd
> Ranjit> Edition?
> 
> Nope.

Thanks.


> Ranjit>   3. Removing tests that are no longer wrong. For example,
> Ranjit>      auto-boxing in JLS 3 invalidates some of the testcases.
> 
> How about making them conditional so that people can continue to test
> 1.4 compilers?  I think this is probably worthwhile for the next year
> or so.

Yes, that'll be better, though I'm not sure of the Tcl incantations
necessary to achieve that.

[I notice that there are separate constraints called "jsr201" and
"staticImport" - AFAIK, JSR201 included static import (along with
auto-boxing, enhanced for, etc.). And I also notice that mainline
GCJ's gcj_setup should switch off for these constraints.]

Ranjit.

- --
Ranjit Mathew       Email: rmathew AT gmail DOT com

Bangalore, INDIA.     Web: http://ranjitmathew.hostingzero.com/




-----BEGIN PGP SIGNATURE-----
Version: GnuPG v1.4.2 (GNU/Linux)
Comment: Using GnuPG with Thunderbird - http://enigmail.mozdev.org

iD8DBQFDLEwxYb1hx2wRS48RAm7TAJ9mlXZ6+V/yOc1h7GczRh4dPSl8WgCbBBny
5Q5rM/1BS98kma1fN4GR1UA=
=vnfp
-----END PGP SIGNATURE-----

Re: Keeping Jacks Up-to-date

[Tom Tromey]

>> How about making them conditional so that people can continue to test
>> 1.4 compilers?  I think this is probably worthwhile for the next year
>> or so.

Ranjit> Yes, that'll be better, though I'm not sure of the Tcl incantations
Ranjit> necessary to achieve that.

Actually I have a patch to add '!boxing' constraints where needed,
that I forgot to check in.  Whoops.

Ranjit> [I notice that there are separate constraints called "jsr201" and
Ranjit> "staticImport" - AFAIK, JSR201 included static import (along with
Ranjit> auto-boxing, enhanced for, etc.). And I also notice that mainline
Ranjit> GCJ's gcj_setup should switch off for these constraints.]

Yeah.  The 201 additions were not planned much.
Feel free to fix this stuff if you like.

Tom

Re: Keeping Jacks Up-to-date

[Ranjit Mathew]

On 19 Sep 2005 08:43:28 -0600, Tom Tromey <tromey@redhat.com> wrote:
> 
> Ranjit> [I notice that there are separate constraints called "jsr201" and
> Ranjit> "staticImport" - AFAIK, JSR201 included static import (along with
> Ranjit> auto-boxing, enhanced for, etc.). And I also notice that mainline
> Ranjit> GCJ's gcj_setup should switch off for these constraints.]
> 
> Yeah.  The 201 additions were not planned much.
> Feel free to fix this stuff if you like.

Judging by your recent patch, and just to be clear, you'd prefer
the constraints to be "boxing", "staticImport", "generics", etc.
rather than a blanket "jsr201", right?

Thanks,
Ranjit.

-- 
Ranjit Mathew      Email: rmathew AT gmail DOT com

Bangalore, INDIA.    Web: http://ranjitmathew.hostingzero.com/

Re: Keeping Jacks Up-to-date

[Tom Tromey]

>> Yeah.  The 201 additions were not planned much.
>> Feel free to fix this stuff if you like.

Ranjit> Judging by your recent patch, and just to be clear, you'd prefer
Ranjit> the constraints to be "boxing", "staticImport", "generics", etc.
Ranjit> rather than a blanket "jsr201", right?

Actually I expect we won't gain much by having separate constraints,
most compilers just have a single 1.4/1.5 switch (gcjx has separate
switches but it isn't totally clear that this makes sense in all
circumstances -- I imagine I'll end up getting rid of those).

Tom

Re: Keeping Jacks Up-to-date

[Ranjit Mathew]

On 20 Sep 2005 08:37:07 -0600, Tom Tromey <tromey@redhat.com> wrote:
> >> Yeah.  The 201 additions were not planned much.
> >> Feel free to fix this stuff if you like.
> 
> Ranjit> Judging by your recent patch, and just to be clear, you'd prefer
> Ranjit> the constraints to be "boxing", "staticImport", "generics", etc.
> Ranjit> rather than a blanket "jsr201", right?
> 
> Actually I expect we won't gain much by having separate constraints,
> most compilers just have a single 1.4/1.5 switch (gcjx has separate
> switches but it isn't totally clear that this makes sense in all
> circumstances -- I imagine I'll end up getting rid of those).

So we should just have a single "jls3" (or whatever) constraint
to cover all these?

Ranjit.

-- 
Ranjit Mathew      Email: rmathew AT gmail DOT com

Bangalore, INDIA.    Web: http://ranjitmathew.hostingzero.com/

Re: Keeping Jacks Up-to-date

[Tom Tromey]

Ranjit> So we should just have a single "jls3" (or whatever) constraint
Ranjit> to cover all these?

Yes, I think so.

Tom