Re: [oe-commits] Koen Kooi : gcc 4.3.3: fix regression from 4.3. 1 that caused fortran to get disabled for linux-gnueabi

Re: [oe-commits] Koen Kooi : gcc 4.3.3: fix regression from 4.3. 1 that caused fortran to get disabled for linux-gnueabi

[Tom Rini]

On Wed, May 13, 2009 at 10:56:03PM +0200, GIT User account wrote:

> Author: Koen Kooi <koen@openembedded.org>
> Date:   Wed May 13 22:49:19 2009 +0200
> 
> gcc 4.3.3: fix regression from 4.3.1 that caused fortran to get disabled for linux-gnueabi

This came up on IRC today for something else.  Why do we only build a
fortran compiler on ARM (EABI)?  Should it just be disabled on a few
platforms perhaps?

-- 
Tom Rini

Re: [oe-commits] Koen Kooi : gcc 4.3.3: fix regression from 4.3. 1 that caused fortran to get disabled for linux-gnueabi

[Koen Kooi]

On 13-05-09 23:13, Tom Rini wrote:
> On Wed, May 13, 2009 at 10:56:03PM +0200, GIT User account wrote:
>
>> Author: Koen Kooi<koen@openembedded.org>
>> Date:   Wed May 13 22:49:19 2009 +0200
>>
>> gcc 4.3.3: fix regression from 4.3.1 that caused fortran to get disabled for linux-gnueabi
>
> This came up on IRC today for something else.  Why do we only build a
> fortran compiler on ARM (EABI)?  Should it just be disabled on a few
> platforms perhaps?

When we first started playing with things like this (2005-ish) some 
people didn't want java, fortran or obj-c compilers, so everything 
defaulted to off.
And gcc is crappy enough that java, fortran and objc just don't 
crosscompile for every arch (check the arm fortran patches).

Having said that, I rather like complete toolchains, so if it was up to 
me, java, fortran and objc would be turned on by default for every arch.

regards,

Koen

Re: [oe-commits] Koen Kooi : gcc 4.3.3: fix regression from 4.3. 1 that caused fortran to get disabled for linux-gnueabi

[Phil Blundell]

On Wed, 2009-05-13 at 14:13 -0700, Tom Rini wrote:
> On Wed, May 13, 2009 at 10:56:03PM +0200, GIT User account wrote:
> 
> > Author: Koen Kooi <koen@openembedded.org>
> > Date:   Wed May 13 22:49:19 2009 +0200
> > 
> > gcc 4.3.3: fix regression from 4.3.1 that caused fortran to get disabled for linux-gnueabi
> 
> This came up on IRC today for something else.  Why do we only build a
> fortran compiler on ARM (EABI)?  Should it just be disabled on a few
> platforms perhaps?

Yes: all these options (JAVA, FORTRAN, OBJC) ought really to be
defaulting to on and only suppressed when necessary for the specific
platform/version combinations where they're known to be broken beyond
economic repair.  

In fact gcc-configure-common.inc is doing the right thing, but it does
seem that some joker has decided to turn them off by default in the
gcc-4.x.x.inc files.  I'd certainly be in favour of deleting those
assignments for new versions of gcc-4.x.x.inc, and perhaps even
retrospectively deleting them for the existing versions too.

p.