Ian Lance Taylor wrote:
> I'm personally reluctant to codify it, because it's really hard to
> codify good judgment. But if you say in your patch how you tested it,
> the reviewers should be able to consider whether that is sufficient.
I agree.
I always claim that my most valuable contribution
Joern Rennecke writes:
> So, back to the original question. Is this a suitable bootstrap substitute
> for testing patches?
I think it can be. You have to use good judgment, of course. I know
you know this, but if the patch is going to change the generated code on
a specific target, then a boo
Quoting Paolo Bonzini :
--enable-werror-always is today what you're looking for, I think.
I just tried, and it seems to work for my i686-pc-linux-gnu X ia64-linux-gnu
build. Not only did it supply -Werror, make all-gcc already completed,
while the bootstrap on gcc60 has been chugging along fr
On 07/01/2010 03:34 PM, Joern Rennecke wrote:
Quoting Paolo Bonzini :
On 07/01/2010 03:26 PM, Joern Rennecke wrote:
Well, what we want for a bootstrap replacement is that it gives
errors for everything where a bootstrap gives errors.
--enable-werror-always?
No, we don't want -Werror for
Quoting Paolo Bonzini :
On 07/01/2010 03:26 PM, Joern Rennecke wrote:
Well, what we want for a bootstrap replacement is that it gives errors
for everything where a bootstrap gives errors.
--enable-werror-always?
No, we don't want -Werror for files that are excluded from -Werror for
a boot
On 07/01/2010 03:26 PM, Joern Rennecke wrote:
Quoting Paolo Bonzini :
Sorry, I meant that it should not give any warning, not that -Werror is
in use.
Well, what we want for a bootstrap replacement is that it gives errors
for everything where a bootstrap gives errors.
--enable-werror-always?
Quoting Paolo Bonzini :
Sorry, I meant that it should not give any warning, not that -Werror is
in use.
Well, what we want for a bootstrap replacement is that it gives errors
for everything where a bootstrap gives errors.
On 07/01/2010 02:57 PM, Joern Rennecke wrote:
Quoting Paolo Bonzini :
On 07/01/2010 02:27 PM, Joern Rennecke wrote:
When risks of the patch mostly involve type checking or things that
could be caught with a simple compilation, could we relax this
testing requirement to do a cross-build of all-
Quoting Paolo Bonzini :
On 07/01/2010 02:27 PM, Joern Rennecke wrote:
When risks of the patch mostly involve type checking or things that
could be caught with a simple compilation, could we relax this
testing requirement to do a cross-build of all-gcc all-target-libgcc
with a recent fully boots
On 07/01/2010 02:27 PM, Joern Rennecke wrote:
When risks of the patch mostly involve type checking or things that
could be caught with a simple compilation, could we relax this
testing requirement to do a cross-build of all-gcc all-target-libgcc
with a recent fully bootstrapped compiler, with -We
We generally require bootstraps for patches to native-capable targets.
This is quite time consuming for targets like rs6000 or ia64 where
the available machines in the compile farm are have low processing speed
and/or memory, and for rs6000 also suffer issues with mpc / gmp / mpfr
libraries
and
11 matches
Mail list logo
