[NaCl SDK] Switch ppapi_simple to C library

[NaCl SDK] Switch ppapi_simple to C library

Actually, this creates two libraries:
ppapi_simple
ppapi_simple_cpp

The second must be used if you want to use ppapi_cpp. If you use the C
interfaces exclusively, you can use ppapi_simple instead.

BUG=

Committed: https://crrev.com/cd3968da9da9621a56afaf185ebc416a5955cb09
Cr-Commit-Position: refs/heads/master@{#317704}

[binji, 2015-02-19 19:52:50]

binji@chromium.org changed reviewers:
+ bradnelson@chromium.org, sbc@chromium.org

[binji, 2015-02-19 21:01:30]

PTAL

[Sam Clegg, 2015-02-19 21:24:20]

https://codereview.chromium.org/914983003/diff/60001/native_client_sdk/src/li...
File native_client_sdk/src/libraries/ppapi_simple/ppapi_simple_c.a.linkerscript
(right):

https://codereview.chromium.org/914983003/diff/60001/native_client_sdk/src/li...
native_client_sdk/src/libraries/ppapi_simple/ppapi_simple_c.a.linkerscript:1:
INPUT(libppapi_simple_c_entry.a libppapi_simple_real.a)
Can these just be one library?

https://codereview.chromium.org/914983003/diff/60001/native_client_sdk/src/li...
File native_client_sdk/src/libraries/ppapi_simple/ppapi_simple_c.so.linkerscript
(right):

https://codereview.chromium.org/914983003/diff/60001/native_client_sdk/src/li...
native_client_sdk/src/libraries/ppapi_simple/ppapi_simple_c.so.linkerscript:1:
INPUT(libppapi_simple_c_entry.a -lppapi_simple_real)
Why use -lppapi_simple_real but not -lppapi_simple_c_entry?

https://codereview.chromium.org/914983003/diff/60001/native_client_sdk/src/li...
File native_client_sdk/src/libraries/ppapi_simple/ps_context_2d.c (right):

https://codereview.chromium.org/914983003/diff/60001/native_client_sdk/src/li...
native_client_sdk/src/libraries/ppapi_simple/ps_context_2d.c:136: 
nit: Remove trailing newline

https://codereview.chromium.org/914983003/diff/60001/native_client_sdk/src/li...
File native_client_sdk/src/libraries/ppapi_simple/ps_entrypoints_cpp.cc (right):

https://codereview.chromium.org/914983003/diff/60001/native_client_sdk/src/li...
native_client_sdk/src/libraries/ppapi_simple/ps_entrypoints_cpp.cc:42: extern
"C" {
These should be declared as extern "C" in the headers, no?  so it should not be
needed here.

https://codereview.chromium.org/914983003/diff/60001/native_client_sdk/src/li...
File native_client_sdk/src/libraries/ppapi_simple/ps_main.c (right):

https://codereview.chromium.org/914983003/diff/60001/native_client_sdk/src/li...
native_client_sdk/src/libraries/ppapi_simple/ps_main.c:1: /* Copyright (c) 2015
The Chromium Authors. All rights reserved.
nit: might as well drop (c) if you are going to change the date

https://codereview.chromium.org/914983003/diff/60001/native_client_sdk/src/li...
native_client_sdk/src/libraries/ppapi_simple/ps_main.c:47: 
nit: trailing newline.

https://codereview.chromium.org/914983003/diff/60001/native_client_sdk/src/re...
File native_client_sdk/src/resources/Makefile.example.template (right):

https://codereview.chromium.org/914983003/diff/60001/native_client_sdk/src/re...
native_client_sdk/src/resources/Makefile.example.template:75: [[]]
Is this needed?

https://codereview.chromium.org/914983003/diff/60001/native_client_sdk/src/to...
File native_client_sdk/src/tools/host_gcc.mk (right):

https://codereview.chromium.org/914983003/diff/60001/native_client_sdk/src/to...
native_client_sdk/src/tools/host_gcc.mk:176: endef
Hmm..  in my version of this change I just have the dsc file copy the linker
scripts into place, just like they do for headers.   This is more complex but
allows people to mess with the linker scripts more easily I guess.

[Sam Clegg, 2015-02-20 17:33:12]

lgtm

https://codereview.chromium.org/914983003/diff/80001/native_client_sdk/src/li...
File native_client_sdk/src/libraries/ppapi_simple/ps_entrypoints_c.c (right):

https://codereview.chromium.org/914983003/diff/80001/native_client_sdk/src/li...
native_client_sdk/src/libraries/ppapi_simple/ps_entrypoints_c.c:14: extern
PPB_GetInterface g_ps_get_interface;
Can we use headers declarations here instead?  (a private header if needs be)

[binji, 2015-02-20 17:33:49]

https://codereview.chromium.org/914983003/diff/60001/native_client_sdk/src/li...
File native_client_sdk/src/libraries/ppapi_simple/ppapi_simple_c.a.linkerscript
(right):

https://codereview.chromium.org/914983003/diff/60001/native_client_sdk/src/li...
native_client_sdk/src/libraries/ppapi_simple/ppapi_simple_c.a.linkerscript:1:
INPUT(libppapi_simple_c_entry.a libppapi_simple_real.a)
On 2015/02/19 21:24:20, Sam Clegg wrote:
> Can these just be one library?

I don't think so. The entrypoint needs to be static, AFAICT, otherwise using
EXTERN in the linkerscript seems to have no effect. I suppose we could make the
"real" library static too...

OK, just tried this but pnacl-ld barfs when using -l, so I should probably only
add it where it is needed.

https://codereview.chromium.org/914983003/diff/60001/native_client_sdk/src/li...
File native_client_sdk/src/libraries/ppapi_simple/ppapi_simple_c.so.linkerscript
(right):

https://codereview.chromium.org/914983003/diff/60001/native_client_sdk/src/li...
native_client_sdk/src/libraries/ppapi_simple/ppapi_simple_c.so.linkerscript:1:
INPUT(libppapi_simple_c_entry.a -lppapi_simple_real)
On 2015/02/19 21:24:20, Sam Clegg wrote:
> Why use -lppapi_simple_real but not -lppapi_simple_c_entry?

I only used -l where necessary -- without the NEEDED field in the .nexe ends up
with an absolute path when dynamic linking. But I'll switch them all to -l.

https://codereview.chromium.org/914983003/diff/60001/native_client_sdk/src/li...
File native_client_sdk/src/libraries/ppapi_simple/ps_context_2d.c (right):

https://codereview.chromium.org/914983003/diff/60001/native_client_sdk/src/li...
native_client_sdk/src/libraries/ppapi_simple/ps_context_2d.c:136: 
On 2015/02/19 21:24:20, Sam Clegg wrote:
> nit: Remove trailing newline

Done.

https://codereview.chromium.org/914983003/diff/60001/native_client_sdk/src/li...
File native_client_sdk/src/libraries/ppapi_simple/ps_entrypoints_cpp.cc (right):

https://codereview.chromium.org/914983003/diff/60001/native_client_sdk/src/li...
native_client_sdk/src/libraries/ppapi_simple/ps_entrypoints_cpp.cc:42: extern
"C" {
On 2015/02/19 21:24:20, Sam Clegg wrote:
> These should be declared as extern "C" in the headers, no?  so it should not
be
> needed here.

True for PPP_*, but the other definitions are not defined in a header. I didn't
want to put them in the public interface.

https://codereview.chromium.org/914983003/diff/60001/native_client_sdk/src/li...
File native_client_sdk/src/libraries/ppapi_simple/ps_main.c (right):

https://codereview.chromium.org/914983003/diff/60001/native_client_sdk/src/li...
native_client_sdk/src/libraries/ppapi_simple/ps_main.c:1: /* Copyright (c) 2015
The Chromium Authors. All rights reserved.
On 2015/02/19 21:24:20, Sam Clegg wrote:
> nit: might as well drop (c) if you are going to change the date

Done.

https://codereview.chromium.org/914983003/diff/60001/native_client_sdk/src/li...
native_client_sdk/src/libraries/ppapi_simple/ps_main.c:47: 
On 2015/02/19 21:24:20, Sam Clegg wrote:
> nit: trailing newline.

Done.

https://codereview.chromium.org/914983003/diff/60001/native_client_sdk/src/re...
File native_client_sdk/src/resources/Makefile.example.template (right):

https://codereview.chromium.org/914983003/diff/60001/native_client_sdk/src/re...
native_client_sdk/src/resources/Makefile.example.template:75: [[]]
On 2015/02/19 21:24:20, Sam Clegg wrote:
> Is this needed?

Yes, otherwise the following lines are included in the for loop. They're just
comments, but it still is wrong. I guess I could put it in another CL

https://codereview.chromium.org/914983003/diff/60001/native_client_sdk/src/to...
File native_client_sdk/src/tools/host_gcc.mk (right):

https://codereview.chromium.org/914983003/diff/60001/native_client_sdk/src/to...
native_client_sdk/src/tools/host_gcc.mk:176: endef
On 2015/02/19 21:24:20, Sam Clegg wrote:
> Hmm..  in my version of this change I just have the dsc file copy the linker
> scripts into place, just like they do for headers.   This is more complex but
> allows people to mess with the linker scripts more easily I guess.

Yeah, it's more complex but I think it's more explicit. Also, we've always had a
unwritten rule that you can rebuild everything in the lib directory, so putting
sources there would break that.

[binji, 2015-02-20 18:32:32]

New patchsets have been uploaded after l-g-t-m from sbc@chromium.org

[binji, 2015-02-20 18:58:36]

PTAL, removed linkerscript stuff

[Sam Clegg, 2015-02-20 19:33:52]

lgtm

m

[binji, 2015-02-20 21:05:36]

New patchsets have been uploaded after l-g-t-m from sbc@chromium.org

[binji, 2015-02-24 00:47:13]

The CQ bit was checked by binji@chromium.org

[commit-bot: I haz the power, 2015-02-24 00:49:04]

CQ is trying da patch. Follow status at
 https://chromium-cq-status.appspot.com/patch-status/914983003/180001

[commit-bot: I haz the power, 2015-02-24 01:01:17]

Committed patchset #10 (id:180001)

[commit-bot: I haz the power, 2015-02-24 01:02:06]

Patchset 10 (id:??) landed as
https://crrev.com/cd3968da9da9621a56afaf185ebc416a5955cb09
Cr-Commit-Position: refs/heads/master@{#317704}