Remove canned hello_world binary

Remove canned hello_world binary

This change removes canned hello_world binaries for ARM, X86-32 and X86-64.
Instead, build.scons file has been extended to build hello_world.nexe for
each of the architectures, including MIPS now.

BUG= run_trusted_mmap_test was failing for MIPS due to missing binary
TEST= run smoke tests for all architectures

Committed: http://src.chromium.org/viewvc/native_client?view=rev&revision=12314

[petarj, 2013-10-15 12:49:07]

PTAL.

[Mark Seaborn, 2013-10-15 23:38:45]

https://codereview.chromium.org/26971007/diff/1/src/trusted/service_runtime/b...
File src/trusted/service_runtime/build.scons (left):

https://codereview.chromium.org/26971007/diff/1/src/trusted/service_runtime/b...
src/trusted/service_runtime/build.scons:848: untrusted_env =
env.MakeUntrustedNativeEnv()
Better to leave this in.  Otherwise you're relying on untrusted_env being
assigned in a block above which is conditionalised under a different condition,
which is a bit dodgy.

https://codereview.chromium.org/26971007/diff/1/src/trusted/service_runtime/b...
File src/trusted/service_runtime/build.scons (right):

https://codereview.chromium.org/26971007/diff/1/src/trusted/service_runtime/b...
src/trusted/service_runtime/build.scons:557: if not env.Bit('bitcode') or
env.Bit('target_mips32'):
Why would this work for env.Bit('target_mips32') but not for env.Bit('bitcode')
in general?  Can't this be enabled even if 'bitcode' is true?

It might be that you want to check "pnacl_generate_pexe" rather than "bitcode"
here.  Or maybe you just need to apply env.GetTranslatedNexe() so that this
works even for pnacl_generate_pexe=1.

https://codereview.chromium.org/26971007/diff/1/src/trusted/service_runtime/b...
src/trusted/service_runtime/build.scons:562: hello_world_nexe =
untrusted_env.ComponentProgram(
Can you do:
untrusted_env.File('$STAGING_DIR/hello_world$PROGSUFFIX')
so that hello_world.nexe isn't rebuilt here?

https://codereview.chromium.org/26971007/diff/1/src/trusted/service_runtime/b...
src/trusted/service_runtime/build.scons:595: mmap_test_args = mmap_test_args +
[hello_world_nexe]
This is kind of misleading: since this test wouldn't be running ld.so, it
wouldn't load hello_world, so the hello_world argument is immaterial.  Maybe
remove the hello_world argument or just disable this test for
env.Bit('nacl_glibc').

https://codereview.chromium.org/26971007/diff/1/src/trusted/service_runtime/b...
src/trusted/service_runtime/build.scons:894: node = env.CommandSelLdrTestNacl(
Why is this moved?

[petarj, 2013-10-16 00:24:37]

https://codereview.chromium.org/26971007/diff/1/src/trusted/service_runtime/b...
File src/trusted/service_runtime/build.scons (right):

https://codereview.chromium.org/26971007/diff/1/src/trusted/service_runtime/b...
src/trusted/service_runtime/build.scons:557: if not env.Bit('bitcode') or
env.Bit('target_mips32'):
On 2013/10/15 23:38:45, Mark Seaborn wrote:
> Why would this work for env.Bit('target_mips32') but not for
env.Bit('bitcode')
> in general?  Can't this be enabled even if 'bitcode' is true?

This is a workaround to the issue with NO-NATIVE-CC-INVOCATION-ALLOWED
that comes out with call to MakeUntrustedNativeEnv().

https://codereview.chromium.org/26971007/diff/1/src/trusted/service_runtime/b...
src/trusted/service_runtime/build.scons:894: node = env.CommandSelLdrTestNacl(
On 2013/10/15 23:38:45, Mark Seaborn wrote:
> Why is this moved?

It is not this part that was moved, but the other tests left the
"canned binaries" section.

[petarj, 2013-10-16 23:36:07]

New patch uploaded. PTAL.

https://codereview.chromium.org/26971007/diff/1/src/trusted/service_runtime/b...
File src/trusted/service_runtime/build.scons (left):

https://codereview.chromium.org/26971007/diff/1/src/trusted/service_runtime/b...
src/trusted/service_runtime/build.scons:848: untrusted_env =
env.MakeUntrustedNativeEnv()
On 2013/10/15 23:38:45, Mark Seaborn wrote:
> Better to leave this in.  Otherwise you're relying on untrusted_env being
> assigned in a block above which is conditionalised under a different
condition,
> which is a bit dodgy.

untrusted_env is not assigned under condition anymore.

https://codereview.chromium.org/26971007/diff/1/src/trusted/service_runtime/b...
File src/trusted/service_runtime/build.scons (right):

https://codereview.chromium.org/26971007/diff/1/src/trusted/service_runtime/b...
src/trusted/service_runtime/build.scons:562: hello_world_nexe =
untrusted_env.ComponentProgram(
On 2013/10/15 23:38:45, Mark Seaborn wrote:
> Can you do:
> untrusted_env.File('$STAGING_DIR/hello_world$PROGSUFFIX')
> so that hello_world.nexe isn't rebuilt here?

Done.

https://codereview.chromium.org/26971007/diff/1/src/trusted/service_runtime/b...
src/trusted/service_runtime/build.scons:595: mmap_test_args = mmap_test_args +
[hello_world_nexe]
On 2013/10/15 23:38:45, Mark Seaborn wrote:
> This is kind of misleading: since this test wouldn't be running ld.so, it
> wouldn't load hello_world, so the hello_world argument is immaterial.  Maybe
> remove the hello_world argument or just disable this test for
> env.Bit('nacl_glibc').

Done.

[Mark Seaborn, 2013-10-17 00:06:39]

https://codereview.chromium.org/26971007/diff/34001/src/trusted/service_runti...
File src/trusted/service_runtime/build.scons (right):

https://codereview.chromium.org/26971007/diff/34001/src/trusted/service_runti...
src/trusted/service_runtime/build.scons:558: hello_world_nexe =
untrusted_env.File('$STAGING_DIR/hello_world.nexe')
Can you use $PROGSUFFIX instead of .nexe to match the case below, or change the
other case to match this?

https://codereview.chromium.org/26971007/diff/34001/src/trusted/service_runti...
src/trusted/service_runtime/build.scons:563: not env.Bit('nacl_glibc')):
Make this "env.Bit('nacl_static_link')" (no "not"), since that better captures
the intent.

https://codereview.chromium.org/26971007/diff/34001/src/trusted/service_runti...
src/trusted/service_runtime/build.scons:816: and not env.Bit('nacl_glibc')):
Not sure why this wouldn't work with dynamic linking.  Maybe doing
untrusted_env.CommandSelLdrTestNacl() instead of env.CommandSelLdrTestNacl()
would make it work in that case?

https://codereview.chromium.org/26971007/diff/34001/src/trusted/service_runti...
src/trusted/service_runtime/build.scons:921: if env.Bit('target_x86') and not
env.Bit('nacl_glibc'):
Same here.

[petarj, 2013-10-17 18:52:49]

https://codereview.chromium.org/26971007/diff/34001/src/trusted/service_runti...
File src/trusted/service_runtime/build.scons (right):

https://codereview.chromium.org/26971007/diff/34001/src/trusted/service_runti...
src/trusted/service_runtime/build.scons:558: hello_world_nexe =
untrusted_env.File('$STAGING_DIR/hello_world.nexe')
On 2013/10/17 00:06:40, Mark Seaborn wrote:
> Can you use $PROGSUFFIX instead of .nexe to match the case below, or change
the
> other case to match this?

Using $PROGSUFFIX causes a conflict for bitcode=1 because the suffix is
'.nonfinal.pexe',
so we need to call GetTranslatedNexe(). This again in conflict with hello_world
in
nacl_scons in stubout_mode folder.

https://codereview.chromium.org/26971007/diff/34001/src/trusted/service_runti...
src/trusted/service_runtime/build.scons:563: not env.Bit('nacl_glibc')):
On 2013/10/17 00:06:40, Mark Seaborn wrote:
> Make this "env.Bit('nacl_static_link')" (no "not"), since that better captures
> the intent.

Done.

https://codereview.chromium.org/26971007/diff/34001/src/trusted/service_runti...
src/trusted/service_runtime/build.scons:816: and not env.Bit('nacl_glibc')):
On 2013/10/17 00:06:40, Mark Seaborn wrote:
> Not sure why this wouldn't work with dynamic linking.  Maybe doing
> untrusted_env.CommandSelLdrTestNacl() instead of env.CommandSelLdrTestNacl()
> would make it work in that case?

untrusted_env has an issue with bitcode=1 while calling GetFinalizedPexe()
(called from CommandSelLdrTestNacl()).

[petarj, 2013-10-22 23:14:26]

On 2013/10/17 18:52:49, petarj wrote:
>
https://codereview.chromium.org/26971007/diff/34001/src/trusted/service_runti...
> File src/trusted/service_runtime/build.scons (right):
> 
>
https://codereview.chromium.org/26971007/diff/34001/src/trusted/service_runti...
> src/trusted/service_runtime/build.scons:558: hello_world_nexe =
> untrusted_env.File('$STAGING_DIR/hello_world.nexe')
> On 2013/10/17 00:06:40, Mark Seaborn wrote:
> > Can you use $PROGSUFFIX instead of .nexe to match the case below, or change
> the
> > other case to match this?
> 
> Using $PROGSUFFIX causes a conflict for bitcode=1 because the suffix is
> '.nonfinal.pexe',
> so we need to call GetTranslatedNexe(). This again in conflict with
hello_world
> in
> nacl_scons in stubout_mode folder.
> 
>
https://codereview.chromium.org/26971007/diff/34001/src/trusted/service_runti...
> src/trusted/service_runtime/build.scons:563: not env.Bit('nacl_glibc')):
> On 2013/10/17 00:06:40, Mark Seaborn wrote:
> > Make this "env.Bit('nacl_static_link')" (no "not"), since that better
captures
> > the intent.
> 
> Done.
> 
>
https://codereview.chromium.org/26971007/diff/34001/src/trusted/service_runti...
> src/trusted/service_runtime/build.scons:816: and not env.Bit('nacl_glibc')):
> On 2013/10/17 00:06:40, Mark Seaborn wrote:
> > Not sure why this wouldn't work with dynamic linking.  Maybe doing
> > untrusted_env.CommandSelLdrTestNacl() instead of env.CommandSelLdrTestNacl()
> > would make it work in that case?
> 
> untrusted_env has an issue with bitcode=1 while calling GetFinalizedPexe()
> (called from CommandSelLdrTestNacl()).

Any additional comments here?

Regards,
Petar

[Mark Seaborn, 2013-10-24 16:40:24]

LGTM

[commit-bot: I haz the power, 2013-10-24 17:11:35]

CQ is trying da patch. Follow status at
https://nativeclient-status.appspot.com/cq/petarj@mips.com/26971007/90001

[commit-bot: I haz the power, 2013-10-24 17:18:19]

Try job failure for 26971007-90001 (retry) on
nacl-precise_64-newlib-x86_32-pnacl-spec for steps "annotate, build
prerequisites [x86-32 bitcode] ".
It's a second try, previously, steps "annotate, build prerequisites [x86-32
bitcode] " failed.
http://build.chromium.org/p/tryserver.nacl/buildstatus?builder=nacl-precise_6...

[commit-bot: I haz the power, 2013-10-24 18:37:31]

CQ is trying da patch. Follow status at
https://nativeclient-status.appspot.com/cq/petarj@mips.com/26971007/90001

[commit-bot: I haz the power, 2013-10-24 18:43:29]

Try job failure for 26971007-90001 (retry) on
nacl-precise_64-newlib-x86_32-pnacl-spec for steps "annotate, build
prerequisites [x86-32 bitcode] ".
It's a second try, previously, steps "annotate, build prerequisites [x86-32
bitcode] " failed.
http://build.chromium.org/p/tryserver.nacl/buildstatus?builder=nacl-precise_6...

[commit-bot: I haz the power, 2013-10-24 22:02:23]

CQ is trying da patch. Follow status at
https://nativeclient-status.appspot.com/cq/petarj@mips.com/26971007/90001

[commit-bot: I haz the power, 2013-10-25 02:37:41]

Change committed as 12314