Make use_nss a gyp var so that aura can build w/ nss.

Make use_nss a gyp var so that aura can build w/ nss.

BUG=None
TEST=None

Committed: http://src.chromium.org/viewvc/chrome?view=rev&revision=104100

[mattm, 2011-10-05 02:27:51]

http://codereview.chromium.org/8135024/diff/1001/build/common.gypi
File build/common.gypi (right):

http://codereview.chromium.org/8135024/diff/1001/build/common.gypi#newcode82
build/common.gypi:82: ['OS=="linux" or OS=="freebsd" or OS=="openbsd" or
OS=="solaris" or OS=="android"', {
android shouldn't be in this list (in the conditions down below it sets
use_openssl=1)

http://codereview.chromium.org/8135024/diff/1001/build/common.gypi#newcode82
build/common.gypi:82: ['OS=="linux" or OS=="freebsd" or OS=="openbsd" or
OS=="solaris" or OS=="android"', {
I'm not sure I understand the gyp stuff that requires this multilevel defines
thing, but I think this could go in the main block further down?

http://codereview.chromium.org/8135024/diff/1001/build/common.gypi#newcode98
build/common.gypi:98: 'use_nss%': '<(use_nss)',
And this could probably go next to use_openssl?

http://codereview.chromium.org/8135024/diff/1001/chrome/chrome_browser.gypi
File chrome/chrome_browser.gypi (right):

http://codereview.chromium.org/8135024/diff/1001/chrome/chrome_browser.gypi#n...
chrome/chrome_browser.gypi:4133: }],
unrelated?

http://codereview.chromium.org/8135024/diff/1001/chrome/chrome_browser.gypi#n...
chrome/chrome_browser.gypi:4162: ['use_nss==0 and use_openssl==1', {
use_openssl==0

http://codereview.chromium.org/8135024/diff/1001/chrome/chrome_exe.gypi
File chrome/chrome_exe.gypi (right):

http://codereview.chromium.org/8135024/diff/1001/chrome/chrome_exe.gypi#newco...
chrome/chrome_exe.gypi:147: }],
unrelated?

[DaveMoore, 2011-10-05 03:15:52]

Check it again...I had an extra file accidentally and changed the way I'm
including the source files.
On 2011/10/05 02:27:51, mattm wrote:
> http://codereview.chromium.org/8135024/diff/1001/build/common.gypi
> File build/common.gypi (right):
> 
> http://codereview.chromium.org/8135024/diff/1001/build/common.gypi#newcode82
> build/common.gypi:82: ['OS=="linux" or OS=="freebsd" or OS=="openbsd" or
> OS=="solaris" or OS=="android"', {
> android shouldn't be in this list (in the conditions down below it sets
> use_openssl=1)
> 
> http://codereview.chromium.org/8135024/diff/1001/build/common.gypi#newcode82
> build/common.gypi:82: ['OS=="linux" or OS=="freebsd" or OS=="openbsd" or
> OS=="solaris" or OS=="android"', {
> I'm not sure I understand the gyp stuff that requires this multilevel defines
> thing, but I think this could go in the main block further down?
> 
> http://codereview.chromium.org/8135024/diff/1001/build/common.gypi#newcode98
> build/common.gypi:98: 'use_nss%': '<(use_nss)',
> And this could probably go next to use_openssl?
> 
> http://codereview.chromium.org/8135024/diff/1001/chrome/chrome_browser.gypi
> File chrome/chrome_browser.gypi (right):
> 
>
http://codereview.chromium.org/8135024/diff/1001/chrome/chrome_browser.gypi#n...
> chrome/chrome_browser.gypi:4133: }],
> unrelated?
> 
>
http://codereview.chromium.org/8135024/diff/1001/chrome/chrome_browser.gypi#n...
> chrome/chrome_browser.gypi:4162: ['use_nss==0 and use_openssl==1', {
> use_openssl==0
> 
> http://codereview.chromium.org/8135024/diff/1001/chrome/chrome_exe.gypi
> File chrome/chrome_exe.gypi (right):
> 
>
http://codereview.chromium.org/8135024/diff/1001/chrome/chrome_exe.gypi#newco...
> chrome/chrome_exe.gypi:147: }],
> unrelated?

[mattm, 2011-10-05 03:44:47]

http://codereview.chromium.org/8135024/diff/3004/build/common.gypi
File build/common.gypi (right):

http://codereview.chromium.org/8135024/diff/3004/build/common.gypi#newcode90
build/common.gypi:90: 'use_nss%': 1,
Did you see the question about whether these could go in one of the other
less-nested condition blocks?

http://codereview.chromium.org/8135024/diff/3004/chrome/chrome_browser.gypi
File chrome/chrome_browser.gypi (left):

http://codereview.chromium.org/8135024/diff/3004/chrome/chrome_browser.gypi#o...
chrome/chrome_browser.gypi:323: 'browser/certificate_manager_model.h',
These should stay here and be excluded when we don't need them.  (There's a
general idiom of doing that.   Some IDE like VS let you search among all files
even ones the platform doesn't use, but only if they are listed here and then
excluded later.)

http://codereview.chromium.org/8135024/diff/3004/chrome/chrome_browser.gypi
File chrome/chrome_browser.gypi (right):

http://codereview.chromium.org/8135024/diff/3004/chrome/chrome_browser.gypi#n...
chrome/chrome_browser.gypi:4178: 'sources': [
Should keep it the excluding way

[DaveMoore, 2011-10-05 04:02:57]

http://codereview.chromium.org/8135024/diff/3004/build/common.gypi
File build/common.gypi (right):

http://codereview.chromium.org/8135024/diff/3004/build/common.gypi#newcode90
build/common.gypi:90: 'use_nss%': 1,
On 2011/10/05 03:44:47, mattm wrote:
> Did you see the question about whether these could go in one of the other
> less-nested condition blocks?

Done.

http://codereview.chromium.org/8135024/diff/3004/chrome/chrome_browser.gypi
File chrome/chrome_browser.gypi (left):

http://codereview.chromium.org/8135024/diff/3004/chrome/chrome_browser.gypi#o...
chrome/chrome_browser.gypi:323: 'browser/certificate_manager_model.h',
On 2011/10/05 03:44:47, mattm wrote:
> These should stay here and be excluded when we don't need them.  (There's a
> general idiom of doing that.   Some IDE like VS let you search among all files
> even ones the platform doesn't use, but only if they are listed here and then
> excluded later.)

Done.

http://codereview.chromium.org/8135024/diff/3004/chrome/chrome_browser.gypi
File chrome/chrome_browser.gypi (right):

http://codereview.chromium.org/8135024/diff/3004/chrome/chrome_browser.gypi#n...
chrome/chrome_browser.gypi:4178: 'sources': [
On 2011/10/05 03:44:47, mattm wrote:
> Should keep it the excluding way

Done.

[mattm, 2011-10-05 04:10:13]

lgtm

[joth, 2011-10-05 15:44:01]

FYI This broke the use_openssl build, because now both use_openssl=1 and
use_nss=1 are getting set
http://chromegw.corp.google.com/i/chromium.fyi/builders/Chromium%20Linux%20Re...

Not sure if its reasonable to require use_nss=0 is also set whenever
use_openssl=1 is GYP_DEFINES.. is so I'll add that to the buildbot spec. Or
else, will need some restructuring of the gypi to work this out.

[mattm, 2011-10-05 19:11:47]

My bad, didn't think about what would happen when manually forcing openssl.  I
suppose it would be possible to add use_openssl=0 to the conditions for setting
use_nss, I will take a look at that.

On 2011/10/05 15:44:01, joth wrote:
> FYI This broke the use_openssl build, because now both use_openssl=1 and
> use_nss=1 are getting set
>
http://chromegw.corp.google.com/i/chromium.fyi/builders/Chromium%2520Linux%25...
> 
> Not sure if its reasonable to require use_nss=0 is also set whenever
> use_openssl=1 is GYP_DEFINES.. is so I'll add that to the buildbot spec. Or
> else, will need some restructuring of the gypi to work this out.

[joth, 2011-10-05 19:19:46]

Thanks both for following up on this. I did look at setting the builder to
explicitly set use_nss=0 via GYP_DEFINES but it's not going to be quick to
propogate this through to various other build configs that duplicate these
sections, so if you can see to handle it in the gypi that would be
great. (I've got to head home now so won't get to this for another day
anyway)

I guess use_openssl will need pulling into one of the inner dicts, so it can
be 'seen' by the correct conditional.

Thanks!

On 5 October 2011 20:11, <mattm@chromium.org> wrote:

> My bad, didn't think about what would happen when manually forcing openssl.
>  I
> suppose it would be possible to add use_openssl=0 to the conditions for
> setting
> use_nss, I will take a look at that.
>
>
> On 2011/10/05 15:44:01, joth wrote:
>
>> FYI This broke the use_openssl build, because now both use_openssl=1 and
>> use_nss=1 are getting set
>>
>
> http://chromegw.corp.google.**com/i/chromium.fyi/builders/**
>
Chromium%2520Linux%2520Redux/**builds/239<http://chromegw.corp.google.com/i/chromium.fyi/builders/Chromium%2520Linux%2520Redux/builds/239>
>
>
>  Not sure if its reasonable to require use_nss=0 is also set whenever
>> use_openssl=1 is GYP_DEFINES.. is so I'll add that to the buildbot spec.
>> Or
>> else, will need some restructuring of the gypi to work this out.
>>
>
>
>
>
http://codereview.chromium.**org/8135024/<http://codereview.chromium.org/8135...
>