Re: [classpath-0.98 merge branch] update and status

Re: [classpath-0.98 merge branch] update and status

[Andrew John Hughes]

On 21:56 Sun 14 Sep     , Matthias Klose wrote:
> Merged trunk into the classpath merge branch.
> 
> Imported new classpath snapshot into the merge branch:
> 
>  - Reverted the Thread optimization (does require gcj local changes,
>    and I did run out of time).
> 
>  - Don't explicitely configure with --disable-gjdoc. This needs
>    configuration with --with-antlr-jar as well.
> 
>  - For ANTLR, the classpath configury wants a JAVA runtime (which
>    currently is not configured/set). I added a check for the binary
>    runantlr, which may be Debian/Ubuntu specific). Did not yet check,
>    what is needed to run the installed gjdoc.
> 

At present, the build fails for me because I don't have a 'runantlr'.
How do we handle other such dependencies like gjavah?

> The changelogs are appended.
> 
> The merge branch currently builds in maintainer and non-maintainer mode; the
> testsuite suceeds in both cases with no failures. Not yet tested with mauve.
> 
> There are still some todo's; not sure which ones are required before the merge
> to the trunk:
> 
>  - Apply the Thread optimization change
> 
>  - Apply the nio change, backout out for the 2008-08-11 classpath import
> 
>  - Tom Tromey had a comment that he wanted to add/fix something:
>    http://gcc.gnu.org/bugzilla/show_bug.cgi?id=36762
> 
>  - http://gcc.gnu.org/bugzilla/show_bug.cgi?id=36763
> 
>  - * scripts/makemake.tcl: Switch gnu.java.math to BC compilation.
>    concerns by Andrew Haley.
> 
>  - Build a gjdoc binary.
> 
> I'll be offline for at least one week, starting tomorrow on Monday.
> 
>   Matthias

> libjava/
> 
> 2008-09-14  Matthias Klose  <doko@ubuntu.com>
> 
> 	Import GNU Classpath (libgcj-import-20080914).
> 
> 	* Regenerate class and header files.
> 	* Regenerate auto* files.
> 	* configure.ac: Don't pass --disable-gjdoc to classpath.
> 	* sources.am: Regenerated.
> 	* HACKING: Mention to build gjdoc in maintainer builds.
> 	* gnu/classpath/Configuration.java: Update classpath version.
> 	* gcj/javaprims.h: Update.
> 
> libjava/classpath/
> 
> 2008-09-14  Matthias Klose  <doko@ubuntu.com>
> 
> 	* m4/ac_prog_javac_works.m4, m4/ac_prog_javac.m4, m4/acinclude.m4:
> 	Revert local changes.
> 	* m4/ac_prog_antlr.m4: Check for an runantlr binary.
> 	* tools/Makefile.am, lib/Makefile.am: Revert local changes (JCOMPILER).
> 	* tools/Makefile.am: Remove USE_JAVAC_FLAGS, pass ANTLR_JAR in
> 	GLIBJ_CLASSPATH.
> 
> 2008-09-14  Matthias Klose  <doko@ubuntu.com>
> 
> 	Revert:
> 	Daniel Frampton  <zyridium at zyridium.net>
> 
> 	* AUTHORS: Added.
> 	* java/lang/InheritableThreadLocal.java,
> 	* java/lang/Thread.java,
> 	* java/lang/ThreadLocal.java:
> 	Modified to use java.lang.ThreadLocalMap.
> 	* java/lang/ThreadLocalMap.java:
> 	New cheaper ThreadLocal-specific WeakHashMap.


-- 
Andrew :)

Support Free Java!
Contribute to GNU Classpath and the OpenJDK
http://www.gnu.org/software/classpath
http://openjdk.java.net
PGP Key: 94EFD9D8 (http://subkeys.pgp.net)
Fingerprint = F8EF F1EA 401E 2E60 15FA  7927 142C 2591 94EF D9D8

Re: [classpath-0.98 merge branch] update and status

[Matthias Klose]

Andrew John Hughes schrieb:
> On 21:56 Sun 14 Sep     , Matthias Klose wrote:
>> Merged trunk into the classpath merge branch.
>>
>> Imported new classpath snapshot into the merge branch:
>>
>>  - Reverted the Thread optimization (does require gcj local changes,
>>    and I did run out of time).
>>
>>  - Don't explicitely configure with --disable-gjdoc. This needs
>>    configuration with --with-antlr-jar as well.
>>
>>  - For ANTLR, the classpath configury wants a JAVA runtime (which
>>    currently is not configured/set). I added a check for the binary
>>    runantlr, which may be Debian/Ubuntu specific). Did not yet check,
>>    what is needed to run the installed gjdoc.
>>
> 
> At present, the build fails for me because I don't have a 'runantlr'.
> How do we handle other such dependencies like gjavah?

in maintainer mode, you are expected to have them. That's the case for gjavah
and ecj1. But we need an antlr binary for a normal build as well if we do
configure with --enable-gjdoc. Or we have to check for any "JAVA" interpreter
and disable gjdoc if we cannot find one.

  Matthias

Re: [classpath-0.98 merge branch] update and status

[Matthias Klose]

On 21:56 Sun 14 Sep     , Matthias Klose wrote:
> There are still some todo's; not sure which ones are required before the merge
> to the trunk:
>
>  - Apply the nio change, backout out for the 2008-08-11 classpath import
>
>  - Build a gjdoc binary.

these are fixed now.

>  - Tom Tromey had a comment that he wanted to add/fix something:
>    http://gcc.gnu.org/bugzilla/show_bug.cgi?id=36762
>
>  - * scripts/makemake.tcl: Switch gnu.java.math to BC compilation.
>    concerns by Andrew Haley.

these were not required for the merge according to Tom and Andrew.

>  - Apply the Thread optimization change

not used in libgcj; can be delayed.

>  - http://gcc.gnu.org/bugzilla/show_bug.cgi?id=36763

outstanding?

are there any new issues?

how should the changelogs be handled? Merging all entries into one, or just
updating the date when the branch is merged?

are we in a state where the merge can be proposed/announcd on gcc-patches?

  Matthias

Re: [classpath-0.98 merge branch] update and status

[Andrew John Hughes]

On 17:32 Sun 28 Sep     , Matthias Klose wrote:

snip...
> 
> >  - Apply the Thread optimization change
> 
> not used in libgcj; can be delayed.
> 

Would be nice for completeness.

> >  - http://gcc.gnu.org/bugzilla/show_bug.cgi?id=36763
> 
> outstanding?
> 

Is this not already on trunk? (it's the previous merge)
There are plenty of VM 'holes' like these already on trunk
so I can't see this being a blocker.

> are there any new issues?
> 

aph cleared up the main one, which was CPStringBuilder.

> how should the changelogs be handled? Merging all entries into one, or just
> updating the date when the branch is merged?
> 

Which ChangeLog(s) do you mean?  I'd prefer to keep the individual entries
to be honest.  Merging them would only make it harder to track down issues later.

> are we in a state where the merge can be proposed/announcd on gcc-patches?
> 

Did you see the Mauve results I posted?  There are a few GCJ-specific regressions
(from 4.3), but not sure whether these are also on trunk.

>   Matthias
-- 
Andrew :)

Support Free Java!
Contribute to GNU Classpath and the OpenJDK
http://www.gnu.org/software/classpath
http://openjdk.java.net
PGP Key: 94EFD9D8 (http://subkeys.pgp.net)
Fingerprint = F8EF F1EA 401E 2E60 15FA  7927 142C 2591 94EF D9D8

Re: [classpath-0.98 merge branch] update and status

[Matthias Klose]

Andrew John Hughes schrieb:
> On 17:32 Sun 28 Sep     , Matthias Klose wrote:
>>>  - http://gcc.gnu.org/bugzilla/show_bug.cgi?id=36763
>> outstanding?
>>
> 
> Is this not already on trunk? (it's the previous merge)
> There are plenty of VM 'holes' like these already on trunk
> so I can't see this being a blocker.

libgcj has some classes which shadow the classpath classes. while the merge is
ok, it doesn't affect the libgcj implementation. I did fil bug reports where I
did notice these diversions and didn't merge these into libgcj.

>> are we in a state where the merge can be proposed/announcd on gcc-patches?
>>
> 
> Did you see the Mauve results I posted?  There are a few GCJ-specific regressions
> (from 4.3), but not sure whether these are also on trunk.

how should we proceed? please file reports on gcc/libgcj. do you see these for
classpath builds as well?

  Matthias