Linux: Add a check to the startup script to check for SSE2 support.

Linux: Add a check to the startup script to check for SSE2 support.

BUG=348761
NOTRY=true

Committed: https://src.chromium.org/viewvc/chrome?view=rev&revision=271529

[Nico, 2014-05-20 01:27:09]

lgtm

[Lei Zhang, 2014-05-20 01:31:31]

The CQ bit was checked by thestig@chromium.org

[commit-bot: I haz the power, 2014-05-20 01:32:43]

CQ is trying da patch. Follow status at
 https://chromium-status.appspot.com/cq/thestig@chromium.org/296663002/1

[commit-bot: I haz the power, 2014-05-20 01:34:29]

Change committed as 271529

[witold.baryluk, 2016-05-09 20:55:48]

witold.baryluk@gmail.com changed reviewers:
+ witold.baryluk@gmail.com

[witold.baryluk, 2016-05-09 20:55:48]

Hi. This is not the best approach to test SSE2 support.

Chromium runs also on archs other than i386 and amd64. Notably it works on arm,
aarch32 and mips. But this prevents me from using chromium.

Please wrap this check in the detection of the Intel architecture first.

I.e. uname --machine will return something like: x86_64, i686 or aarch64, armv6l
, armv7l depending on the machine.

Other option is to run file(1) on the chromium binary, but file is likely to not
be installed on some machines, and it is hackish.


So, it would be better to only do what you are trying to do here.

[Lei Zhang, 2016-05-09 21:10:01]

On 2016/05/09 20:55:48, witold.baryluk wrote:
> Hi. This is not the best approach to test SSE2 support.
> 
> Chromium runs also on archs other than i386 and amd64. Notably it works on
arm,
> aarch32 and mips. But this prevents me from using chromium.
> 
> Please wrap this check in the detection of the Intel architecture first.
> 
> I.e. uname --machine will return something like: x86_64, i686 or aarch64,
armv6l
> , armv7l depending on the machine.
> 
> Other option is to run file(1) on the chromium binary, but file is likely to
not
> be installed on some machines, and it is hackish.
> 
> 
> So, it would be better to only do what you are trying to do here.

AFAIK, chrome/installer/linux/common/wrapper is only used in Google Chrome on
Linux, which only ships for amd64 as of this writing. If some Chromium package
copied it and is using it for a non-x86 build, that's the packager's problem.

Can you file a bug on crbug.com for your issue? Happy to continue discussing
your problem there.

[Lei Zhang, 2016-05-09 21:24:00]

On 2016/05/09 21:10:01, Lei Zhang wrote:
> AFAIK, chrome/installer/linux/common/wrapper is only used in Google Chrome on
> Linux, which only ships for amd64 as of this writing. If some Chromium package
> copied it and is using it for a non-x86 build, that's the packager's problem.

And that reminds me, Google Chrome no longer needs this check, so I'm removing
this code in https://codereview.chromium.org/1961163002/

You should still file a bug for your issue with Chromium though.

[witold.baryluk, 2016-05-09 21:54:58]

Yes, it is probably packaging issue. I had no problem with this just few
weeks ago on arm64 with chromium (using Debian testing/unstable sources for
compilation + own patches for other issues).

I will try to trouble shoot if further.


2016-05-09 23:24 GMT+02:00 <thestig@chromium.org>:

> On 2016/05/09 21:10:01, Lei Zhang wrote:
> > AFAIK, chrome/installer/linux/common/wrapper is only used in Google
> Chrome on
> > Linux, which only ships for amd64 as of this writing. If some Chromium
> package
> > copied it and is using it for a non-x86 build, that's the packager's
> problem.
>
> And that reminds me, Google Chrome no longer needs this check, so I'm
> removing
> this code in https://codereview.chromium.org/1961163002/
>
> You should still file a bug for your issue with Chromium though.
>
> https://codereview.chromium.org/296663002/
>

-- 
You received this message because you are subscribed to the Google Groups
"Chromium-reviews" group.
To unsubscribe from this group and stop receiving emails from it, send an email
to chromium-reviews+unsubscribe@chromium.org.

[witold.baryluk, 2016-05-09 21:58:16]

Indeed, it looks like Debian provides own launcher wrapper with similar
problem and almost same code. -
https://anonscm.debian.org/cgit/pkg-chromium/pkg-chromium.git/tree/debian/scr...

I will poke them about it and/or send fixes.

Thanks for the quick response.


2016-05-09 23:54 GMT+02:00 Witold Baryluk <witold.baryluk@gmail.com>:

> Yes, it is probably packaging issue. I had no problem with this just few
> weeks ago on arm64 with chromium (using Debian testing/unstable sources for
> compilation + own patches for other issues).
>
> I will try to trouble shoot if further.
>
>
> 2016-05-09 23:24 GMT+02:00 <thestig@chromium.org>:
>
>> On 2016/05/09 21:10:01, Lei Zhang wrote:
>> > AFAIK, chrome/installer/linux/common/wrapper is only used in Google
>> Chrome on
>> > Linux, which only ships for amd64 as of this writing. If some Chromium
>> package
>> > copied it and is using it for a non-x86 build, that's the packager's
>> problem.
>>
>> And that reminds me, Google Chrome no longer needs this check, so I'm
>> removing
>> this code in https://codereview.chromium.org/1961163002/
>>
>> You should still file a bug for your issue with Chromium though.
>>
>> https://codereview.chromium.org/296663002/
>>
>
>

-- 
You received this message because you are subscribed to the Google Groups
"Chromium-reviews" group.
To unsubscribe from this group and stop receiving emails from it, send an email
to chromium-reviews+unsubscribe@chromium.org.