Bug 28149 - Fail to test GCC on simulation due to libjava build errors.
Fail to test GCC on simulation due to libjava build errors.
Status: NEW
Product: gcc
Classification: Unclassified
Component: java
4.2.0
: P3 normal
: ---
Assigned To: Not yet assigned to anyone
:
Depends on:
Blocks:
  Show dependency treegraph
 
Reported: 2006-06-24 07:50 UTC by Jim Huang
Modified: 2007-02-14 15:37 UTC (History)
2 users (show)

See Also:
Host: i486-linux-gnu
Target: arm-elf
Build: i486-linux-gnu
Known to work:
Known to fail:
Last reconfirmed: 2006-06-27 15:30:56


Attachments
Fix build errors in libjava when target = arm-elf simulation (1.54 KB, patch)
2006-06-24 07:52 UTC, Jim Huang
Details | Diff

Note You need to log in before you can comment on or make changes to this bug.
Description Jim Huang 2006-06-24 07:50:41 UTC
While testing GCC on a simulator[1], there are some build errors causing the process failed. The files are involving in:
(1) libjava/java/util/natVMTimeZone.cc
(2) libjava/classpath/native/fdlibm/mprec.c

Attachment attempts to fix the issue and drive testing GCC on simulation continue.

[1] Reference: http://gcc.gnu.org/simtest-howto.html
Comment 1 Jim Huang 2006-06-24 07:52:09 UTC
Created attachment 11739 [details]
Fix build errors in libjava when target = arm-elf simulation
Comment 2 Tom Tromey 2006-06-27 15:30:56 UTC
The mprec part of this patch looks fine to me.
This should go in Classpath as well as libgcj.

The timezone part of this patch looks odd.
I would expect that we would need another case in there
for "__timezone" -- both a configure check and another
#if.
Comment 3 Jim Huang 2006-06-29 06:43:37 UTC
(In reply to comment #2)
> The timezone part of this patch looks odd.
> I would expect that we would need another case in there
> for "__timezone" -- both a configure check and another
> #if.

I compared automake detecting report between gcc/gcj and GNU Classpath, and I found that macro definition HAVE_UNDERSCORE_TIMEZONE was set to different values. For the former, that is 1, and becomes unset for the later. GNU Classpath bootstrapped with Automake 1.9.6 works for me, and configure script outputs the right results.

Still, I don't find other clues.
Comment 4 Manuel López-Ibáñez 2007-02-14 15:37:14 UTC
(In reply to comment #3)
> (In reply to comment #2)
> > The timezone part of this patch looks odd.
> > I would expect that we would need another case in there
> > for "__timezone" -- both a configure check and another
> > #if.
> 
> I compared automake detecting report between gcc/gcj and GNU Classpath, and I
> found that macro definition HAVE_UNDERSCORE_TIMEZONE was set to different
> values. For the former, that is 1, and becomes unset for the later. GNU
> Classpath bootstrapped with Automake 1.9.6 works for me, and configure script
> outputs the right results.
> 

Maybe you should ask about that in the gcc@gcc.gnu.org and java@gcc.gnu.org mailing lists, so your question will reach a wider audience.

I had to disable java to bootstrapped+test using a simulator. :(