Make it possible to disable fatal linker warnings

Make it possible to disable fatal linker warnings

This came up in Gentoo Linux bug https://bugs.gentoo.org/show_bug.cgi?id=506268

Fatal linker warnings were originally introduced in https://codereview.chromium.org/196943020

BUG=352985
R=jochen@chromium.org

Committed: https://src.chromium.org/viewvc/chrome?view=rev&revision=262093

[jochen (gone - plz use gerrit), 2014-04-04 15:02:39]

https://codereview.chromium.org/224713007/diff/1/build/common.gypi
File build/common.gypi (right):

https://codereview.chromium.org/224713007/diff/1/build/common.gypi#newcode3077
build/common.gypi:3077: ['os_posix==1 and (chromeos==0 or target_arch!="arm")',
{
can you move the 'and (...)' part to the condition as well (and thec omment)?

[Paweł Hajdan Jr., 2014-04-04 15:09:33]

PTAL

https://codereview.chromium.org/224713007/diff/1/build/common.gypi
File build/common.gypi (right):

https://codereview.chromium.org/224713007/diff/1/build/common.gypi#newcode3077
build/common.gypi:3077: ['os_posix==1 and (chromeos==0 or target_arch!="arm")',
{
On 2014/04/04 15:02:40, jochen wrote:
> can you move the 'and (...)' part to the condition as well (and thec omment)?

Done. Note that now this might unexpectedly break on these configuration if it
regressed and we have no trybot coverage.

[jochen (gone - plz use gerrit), 2014-04-04 15:49:50]

On 2014/04/04 15:09:33, Paweł Hajdan Jr. wrote:
> PTAL
> 
> https://codereview.chromium.org/224713007/diff/1/build/common.gypi
> File build/common.gypi (right):
> 
> https://codereview.chromium.org/224713007/diff/1/build/common.gypi#newcode3077
> build/common.gypi:3077: ['os_posix==1 and (chromeos==0 or
target_arch!="arm")',
> {
> On 2014/04/04 15:02:40, jochen wrote:
> > can you move the 'and (...)' part to the condition as well (and thec
omment)?
> 
> Done. Note that now this might unexpectedly break on these configuration if it
> regressed and we have no trybot coverage.



I think the new condition block also needs is_posix==1

[Paweł Hajdan Jr., 2014-04-04 17:07:37]

On 2014/04/04 15:49:50, jochen wrote:
> I think the new condition block also needs is_posix==1

Could you explain this more?

1) The new condition block is nested inside the existing os_posix==1

2) It's not is_posix (maybe coming from gn) but os_posix, and so is already
there

Please let me know if I missed something obvious. The trybots don't indicate
weird failures on non-posix...

[jochen (gone - plz use gerrit), 2014-04-04 18:27:12]

On 2014/04/04 17:07:37, Paweł Hajdan Jr. wrote:
> On 2014/04/04 15:49:50, jochen wrote:
> > I think the new condition block also needs is_posix==1
> 
> Could you explain this more?
> 
> 1) The new condition block is nested inside the existing os_posix==1
> 
> 2) It's not is_posix (maybe coming from gn) but os_posix, and so is already
> there

typing on a phone is hard...

> 
> Please let me know if I missed something obvious. The trybots don't indicate
> weird failures on non-posix...

I missed that this was nested in the other condition. Anyway, instead of nesting
the condition in the condition, can you just pull this out into it's own block?

[Paweł Hajdan Jr., 2014-04-07 07:53:24]

On 2014/04/04 18:27:12, jochen wrote:
> I missed that this was nested in the other condition. Anyway, instead of
nesting
> the condition in the condition, can you just pull this out into it's own
block?

Could you explain more what are advantages of that?

Note that this would duplicate the os_posix==1 part of the condition, possibly
making it easier for the two conditions to get out of sync.

[jochen (gone - plz use gerrit), 2014-04-07 07:58:57]

On 2014/04/07 07:53:24, Paweł Hajdan Jr. wrote:
> On 2014/04/04 18:27:12, jochen wrote:
> > I missed that this was nested in the other condition. Anyway, instead of
> nesting
> > the condition in the condition, can you just pull this out into it's own
> block?
> 
> Could you explain more what are advantages of that?
> 
> Note that this would duplicate the os_posix==1 part of the condition, possibly
> making it easier for the two conditions to get out of sync.

it's a bug that the relro stuff was lumped together with the warnings. It should
be clear that these are two different things.

[jochen (gone - plz use gerrit), 2014-04-07 08:09:16]

lgtm, thanks

[Paweł Hajdan Jr., 2014-04-07 08:17:29]

Committed patchset #3 manually as r262093 (presubmit successful).