Add support for QNX Neutrino part2

Add support for QNX Neutrino part2

pqnx.h was not added in the previous CL by mistake
(https://codereview.chromium.org/195973005/ ).

A spurious tab in icu.gyp was removed.

README.chromium was updated to clarify how pqnx.h was generated.
platform.qnx.patch was made against a copy of plinux.h

BUG=NONE
TEST=icu is built properly on qnx. After DEPS roll, net_unittests:*IDN* and base_unittests:*String* all pass on qnx.`
R=mark@chromium.org

Committed: https://src.chromium.org/viewvc/chrome?view=rev&revision=258166

[Mark Mentovai, 2014-03-18 19:08:29]

LGTM. However, consider:

In
https://codereview.chromium.org/195973005/diff/40001/patches/platform.qnx.patch:

The patch shows up as a diff between a/source/common/unicode/pandroid.h and
b/source/common/unicode/pqnx.h, but pandroid.h is generated from a diff from
plinux.h (see README.chromium and patches/pandroid.h.patch).

README.chromium does not make clear that this is how platform.qnx.patch works.
And perhaps platform.qnx.patch should be based on plinux.h and not pandroid.h,
which is another plinux.h derivative, especially when one of the things
platform.qnx.patch does is revert one of the changes made by pandroid.h.patch.

[cosmin.truta, 2014-03-18 19:27:19]

On 2014/03/18 19:08:29, Mark Mentovai wrote:
> README.chromium does not make clear that this is how platform.qnx.patch works.
> And perhaps platform.qnx.patch should be based on plinux.h and not pandroid.h,
> which is another plinux.h derivative, especially when one of the things
> platform.qnx.patch does is revert one of the changes made by pandroid.h.patch.

When I presented pqnx.h as a diff from pandroid.h, I did so because I found
that pandroid.h is closest to pqnx.h. They both lack NL_LANGINFO_CODESET.
But that is, indeed, the only difference, and in other respects, plinux.h
is closer.

I am okay with rebasing the pqnx.h diff onto plinux.h.

[jungshik at Google, 2014-03-20 00:01:41]

On 2014/03/18 19:27:19, cosmin.truta wrote:
> On 2014/03/18 19:08:29, Mark Mentovai wrote:
> > README.chromium does not make clear that this is how platform.qnx.patch
works.
> > And perhaps platform.qnx.patch should be based on plinux.h and not
pandroid.h,
> > which is another plinux.h derivative, especially when one of the things
> > platform.qnx.patch does is revert one of the changes made by
pandroid.h.patch.

Thank you, Mark, for the comments. 
I rebased pqnx.h against plinux.h and updated README.chromium accordingly. 


> When I presented pqnx.h as a diff from pandroid.h, I did so because I found
> that pandroid.h is closest to pqnx.h. They both lack NL_LANGINFO_CODESET.
> But that is, indeed, the only difference, and in other respects, plinux.h
> is closer.
> 
> I am okay with rebasing the pqnx.h diff onto plinux.h.

Thanks. 

JFYI,  I'll drop everything (I hope I can :-)) and dedicate the whole week to
upgrading ICU to 52 so that this won't be necessary any more (there'll be a
single platform.h that supports all the platforms we care about. The upstream
made that change a while ago).

[jungshik at Google, 2014-03-20 00:02:20]

Committed patchset #2 manually as r258166 (presubmit successful).

[Mark Mentovai, 2014-03-20 13:44:30]

LGTM