1253431 - Port SDK related install targets to moz.build

Status: RESOLVED FIXED

(Firefox Build System :: General, defect)

Description

[Michael Shal [:mshal]]

I think this requires a new SdkFiles object derived from FinalTargetFiles in order to install things into dist/sdk. Some of the INSTALL_TARGETS also install things generated during the compile tier, so we depend on bug 1253430.

Comment 1

[Michael Shal [:mshal]]

Attachment: MozReview Request: Bug 1253431 part 1 - Add SDK_FILES to moz.build; r?gps

This adds support for an SDK_FILES variable in moz.build, which creates
a FinalTargetFiles object to install files into dist/sdk/

Review commit: https://reviewboard.mozilla.org/r/38307/diff/#index_header
See other reviews: https://reviewboard.mozilla.org/r/38307/

Comment 2

[Michael Shal [:mshal]]

Attachment: MozReview Request: Bug 1253431 part 2 - Use SDK_FILES instead of INSTALL_TARGETS; r?gps

Review commit: https://reviewboard.mozilla.org/r/38309/diff/#index_header
See other reviews: https://reviewboard.mozilla.org/r/38309/

Comment 3

[Michael Shal [:mshal]]

Attachment: MozReview Request: Bug 1253431 part 3 - Move SDK_BINARY files in xpcom/idl-parser/xpidl to moz.build; r?gps

We can just generate xpidllex.py/xpidlyacc.py in the current directory
rather than one directory higher, and specify this directory as an
include path to xpidl-process.py

Review commit: https://reviewboard.mozilla.org/r/38311/diff/#index_header
See other reviews: https://reviewboard.mozilla.org/r/38311/

Comment 4

[Michael Shal [:mshal]]

Attachment: MozReview Request: Bug 1253431 part 4 - Convert header.py to a GENERATED_FILES script; r?gps

We only ever execute this in one place, so we can just have the main
action do the --regen --cachedir=. mode of operation.

Review commit: https://reviewboard.mozilla.org/r/38313/diff/#index_header
See other reviews: https://reviewboard.mozilla.org/r/38313/

Comment 5

[Michael Shal [:mshal]]

Attachment: MozReview Request: Bug 1253431 part 5 - Remove build/unix/Makefile.in; r?gps

Review commit: https://reviewboard.mozilla.org/r/38315/diff/#index_header
See other reviews: https://reviewboard.mozilla.org/r/38315/

Comment 6

[Michael Shal [:mshal]]

Attachment: MozReview Request: Bug 1253431 part 6 - Remove SDK_BINARY from js/src; r?gps

It has no effect anyway, since it is set after including config/rules.mk

Review commit: https://reviewboard.mozilla.org/r/38317/diff/#index_header
See other reviews: https://reviewboard.mozilla.org/r/38317/

Comment 7

[Michael Shal [:mshal]]

Attachment: MozReview Request: Bug 1253431 part 7 - Remove SDK_BINARY; r?gps

Review commit: https://reviewboard.mozilla.org/r/38319/diff/#index_header
See other reviews: https://reviewboard.mozilla.org/r/38319/

Comment 8

[Michael Shal [:mshal]]

try: https://treeherder.mozilla.org/#/jobs?repo=try&revision=e92169592069
(Note this patchset requires bug 1253430 to build successfully)

Comment 9

[Gregory Szorc [:gps]]

Comment on attachment 8726911 [details]
MozReview Request: Bug 1253431 part 1 - Add SDK_FILES to moz.build; r?gps

https://reviewboard.mozilla.org/r/38307/#review35079

Nit: you don't need to write "part N" in commit messages in MozReview because MozReview preserves the commit ordering in the UI since it is basing things off version control.

Comment 10

[Gregory Szorc [:gps]]

Comment on attachment 8726912 [details]
MozReview Request: Bug 1253431 part 2 - Use SDK_FILES instead of INSTALL_TARGETS; r?gps

https://reviewboard.mozilla.org/r/38309/#review35083

The patch reminds me how much I hate xpcom/idl-parser/xpidl/Makefile.in.

Comment 11

[Gregory Szorc [:gps]]

Comment on attachment 8726913 [details]
MozReview Request: Bug 1253431 part 3 - Move SDK_BINARY files in xpcom/idl-parser/xpidl to moz.build; r?gps

https://reviewboard.mozilla.org/r/38311/#review35093

Very nice.

Comment 12

[Gregory Szorc [:gps]]

Comment on attachment 8726914 [details]
MozReview Request: Bug 1253431 part 4 - Convert header.py to a GENERATED_FILES script; r?gps

https://reviewboard.mozilla.org/r/38313/#review35097

::: xpcom/idl-parser/xpidl/header.py
(Diff revision 1)
> -    idl = p.parse(open(file).read(), filename=file)
> -    idl.resolve(options.incdirs, p)
> -    print_header(idl, outfd, file)
> -
> -    if closeoutfd:
> -        outfd.close()
> -
> -    if options.depfile is not None:
> -        dirname = os.path.dirname(options.depfile)
> -        if dirname:
> -            try:
> -                os.makedirs(dirname)
> -            except:
> -                pass
> -        depfd = open(options.depfile, 'w')
> -        deps = [dep.replace('\\', '/') for dep in idl.deps]
> -
> -        print >>depfd, "%s: %s" % (options.outfile, " ".join(deps))
> -        for dep in deps:
> -            print >>depfd, "%s:" % dep

Is all this code really dead?! I guess we have mozbuild.action.xpidl_process, so I guess so!

Comment 13

[Gregory Szorc [:gps]]

Comment on attachment 8726914 [details]
MozReview Request: Bug 1253431 part 4 - Convert header.py to a GENERATED_FILES script; r?gps

https://reviewboard.mozilla.org/r/38313/#review35099

Comment 14

[Gregory Szorc [:gps]]

Comment on attachment 8726915 [details]
MozReview Request: Bug 1253431 part 5 - Remove build/unix/Makefile.in; r?gps

https://reviewboard.mozilla.org/r/38315/#review35101

Comment 15

[Gregory Szorc [:gps]]

Comment on attachment 8726916 [details]
MozReview Request: Bug 1253431 part 6 - Remove SDK_BINARY from js/src; r?gps

https://reviewboard.mozilla.org/r/38317/#review35103

Wow.

Comment 16

[Gregory Szorc [:gps]]

Comment on attachment 8726917 [details]
MozReview Request: Bug 1253431 part 7 - Remove SDK_BINARY; r?gps

https://reviewboard.mozilla.org/r/38319/#review35105

Awesome series.

::: python/mozbuild/mozbuild/frontend/emitter.py:833
(Diff revision 1)
> +        generated_files.update(['%s%s' % (k, self.config.substs.get('BIN_SUFFIX', '')) for k in self._binaries.keys()])

I'm kinda surprised this doesn't invalidate any tests!

Comment 17

[Michael Shal [:mshal]]

(In reply to Gregory Szorc [:gps] from comment #9)
> Nit: you don't need to write "part N" in commit messages in MozReview
> because MozReview preserves the commit ordering in the UI since it is basing
> things off version control.

Thanks for the heads up! I'll do that from now on.

Comment 18

[Michael Shal [:mshal]]

(In reply to Gregory Szorc [:gps] from comment #12)
> Is all this code really dead?! I guess we have
> mozbuild.action.xpidl_process, so I guess so!

Apparently so - we only run header.py once in the whole build.

Comment 19

[Pulsebot]

https://hg.mozilla.org/integration/mozilla-inbound/rev/dab431efeb01
https://hg.mozilla.org/integration/mozilla-inbound/rev/63e1cbf27de9
https://hg.mozilla.org/integration/mozilla-inbound/rev/e11026ba0b4a
https://hg.mozilla.org/integration/mozilla-inbound/rev/e78e4cf27dc8
https://hg.mozilla.org/integration/mozilla-inbound/rev/5f87a0086c0f
https://hg.mozilla.org/integration/mozilla-inbound/rev/54c8a7c98f30
https://hg.mozilla.org/integration/mozilla-inbound/rev/647ab3b76705

Comment 20

[Pulsebot]

https://hg.mozilla.org/integration/mozilla-inbound/rev/ac43f8ec3958

Comment 21

[Carsten Book [:Tomcat]]

https://hg.mozilla.org/mozilla-central/rev/dab431efeb01
https://hg.mozilla.org/mozilla-central/rev/63e1cbf27de9
https://hg.mozilla.org/mozilla-central/rev/e11026ba0b4a
https://hg.mozilla.org/mozilla-central/rev/e78e4cf27dc8
https://hg.mozilla.org/mozilla-central/rev/5f87a0086c0f
https://hg.mozilla.org/mozilla-central/rev/54c8a7c98f30
https://hg.mozilla.org/mozilla-central/rev/647ab3b76705
https://hg.mozilla.org/mozilla-central/rev/ac43f8ec3958