Use lsb_release in install-build-deps.sh to determine linux distro.

Use lsb_release in install-build-deps.sh to determine linux distro.

Parsing /etc/issue is unreliable.

I tried logging into a compute engine bot to verify that
output of lsb_release and it reported precise correctly
(mine your the /etc/issue was also the same as precise
and didn't contain GCEL).

BUG=376363
R=thestig@chromium.org

Committed: https://src.chromium.org/viewvc/chrome?view=rev&revision=272292

[Sam Clegg, 2014-05-22 19:17:44]

+thakis (as lei doesn't seem to be around).

[Sam Clegg, 2014-05-22 19:32:31]

-thakis (lei is back!)

[Lei Zhang, 2014-05-22 19:36:39]

lgtm

https://codereview.chromium.org/298863006/diff/1/build/install-build-deps.sh
File build/install-build-deps.sh (right):

https://codereview.chromium.org/298863006/diff/1/build/install-build-deps.sh#...
build/install-build-deps.sh:67: if ! type -P lsb_release > /dev/null; then
we usually do "which foo" instead of "type -P foo"

[Sam Clegg, 2014-05-22 19:39:46]

On 2014/05/22 19:36:39, Lei Zhang wrote:
> lgtm
> 
> https://codereview.chromium.org/298863006/diff/1/build/install-build-deps.sh
> File build/install-build-deps.sh (right):
> 
>
https://codereview.chromium.org/298863006/diff/1/build/install-build-deps.sh#...
> build/install-build-deps.sh:67: if ! type -P lsb_release > /dev/null; then
> we usually do "which foo" instead of "type -P foo"

According to stack overflow the 'type' shell builtin is better:
http://stackoverflow.com/questions/592620/how-to-check-if-a-program-exists-fr...

If there is a consistent use of which we could stick with that if you prefer?

[Lei Zhang, 2014-05-22 19:47:15]

On 2014/05/22 19:39:46, Sam Clegg wrote:
> On 2014/05/22 19:36:39, Lei Zhang wrote:
> > lgtm
> > 
> > https://codereview.chromium.org/298863006/diff/1/build/install-build-deps.sh
> > File build/install-build-deps.sh (right):
> > 
> >
>
https://codereview.chromium.org/298863006/diff/1/build/install-build-deps.sh#...
> > build/install-build-deps.sh:67: if ! type -P lsb_release > /dev/null; then
> > we usually do "which foo" instead of "type -P foo"
> 
> According to stack overflow the 'type' shell builtin is better:
>
http://stackoverflow.com/questions/592620/how-to-check-if-a-program-exists-fr...
> 
> If there is a consistent use of which we could stick with that if you prefer?

A lot of our scripts are in portable shell and not bash, which may explain why
we use "which" in general. I'd imagine most of the OSes we run on have a sane
which command that doesn't do the stupid things stack overflow talks about. We
also don't care much about performance here.

[Sam Clegg, 2014-05-22 19:56:11]

On 2014/05/22 19:47:15, Lei Zhang wrote:
> On 2014/05/22 19:39:46, Sam Clegg wrote:
> > On 2014/05/22 19:36:39, Lei Zhang wrote:
> > > lgtm
> > > 
> > >
https://codereview.chromium.org/298863006/diff/1/build/install-build-deps.sh
> > > File build/install-build-deps.sh (right):
> > > 
> > >
> >
>
https://codereview.chromium.org/298863006/diff/1/build/install-build-deps.sh#...
> > > build/install-build-deps.sh:67: if ! type -P lsb_release > /dev/null; then
> > > we usually do "which foo" instead of "type -P foo"
> > 
> > According to stack overflow the 'type' shell builtin is better:
> >
>
http://stackoverflow.com/questions/592620/how-to-check-if-a-program-exists-fr...
> > 
> > If there is a consistent use of which we could stick with that if you
prefer?
> 
> A lot of our scripts are in portable shell and not bash, which may explain why
> we use "which" in general. I'd imagine most of the OSes we run on have a sane
> which command that doesn't do the stupid things stack overflow talks about. We
> also don't care much about performance here.

OK. Done.

[Lei Zhang, 2014-05-22 20:00:07]

lgtm, don't forget NOTRY=true

[Sam Clegg, 2014-05-22 20:07:22]

Committed patchset #2 manually as r272292 (presubmit successful).