While preparing our toolchain updates for Firefox 78 ESR. I realized that --enable-strip is gone for Windows and the build now breaks with that configure option set. However, we used it to cut down the size of our Tor Browser .exe files that way. Not setting that option anymore seems to increase the size by about 66%, which is excessive for bundles we want to ship, while e.g. on Linux just an 8% increase compared to Firefox 68 ESR-based bundles can be seen (the 8% are kind of not unusual).

The actual failure with --enable-strip looks like:

 0:08.12   File "/var/tmp/build/firefox-e2c9144979c2/configure.py", line 181, in <module>
 0:08.12     sys.exit(main(sys.argv))
 0:08.12   File "/var/tmp/build/firefox-e2c9144979c2/configure.py", line 52, in main
 0:08.12     sandbox.run(os.path.join(os.path.dirname(__file__), 'moz.configure'))
 0:08.12   File "/var/tmp/build/firefox-e2c9144979c2/python/mozbuild/mozbuild/configure/__init__.py", line 455, in run
 0:08.12     self._value_for(option)
 0:08.12   File "/var/tmp/build/firefox-e2c9144979c2/python/mozbuild/mozbuild/configure/__init__.py", line 545, in _value_for
 0:08.12     return self._value_for_option(obj)
 0:08.12   File "/var/tmp/build/firefox-e2c9144979c2/python/mozbuild/mozbuild/util.py", line 1018, in method_call
 0:08.12     cache[args] = self.func(instance, *args)
 0:08.12   File "/var/tmp/build/firefox-e2c9144979c2/python/mozbuild/mozbuild/configure/__init__.py", line 613, in _value_for_option
 0:08.12     % option_string.split('=', 1)[0])
 0:08.12 mozbuild.configure.options.InvalidOptionError: --enable-strip is not available in this configuration
 0:08.19 *** Fix above errors and then restart with\
 0:08.19                "./mach build"
 0:08.19 make: *** [client.mk:115: configure] Error 1

This is due to bug 1639815, and I assume, this line in particular.

glandium, is the exclusion of Windows intentional here? Is there a significant amount of work required to allow Windows to set --enable-strip?

Set release status flags based on info from the regressing bug 1639815

Bugbug thinks this bug should belong to this component, but please revert this change in case of error.

The component has been changed since the backlog priority was decided, so we're resetting it.
For more information, please visit auto_nag documentation.

Backed out for build bustages.

Push with failure: https://treeherder.mozilla.org/#/jobs?repo=autoland&resultStatus=testfailed%2Cbusted%2Cexception&revision=09a5f4dcd92ae8a3e3526ad51fb3d32785ad6e7b&failure_classification_id=2&selectedTaskRun=HCj-po3KTFmN8e3CDD7VTg.0

Log: https://treeherder.mozilla.org/logviewer.html#/jobs?job_id=312683690&repo=autoland&lineNumber=990
https://treeherder.mozilla.org/logviewer.html#/jobs?job_id=312684018&repo=autoland&lineNumber=561

Backout: https://hg.mozilla.org/integration/autoland/rev/eb557787e80df69548cc34cc052a6f0e785f1bc6

https://hg.mozilla.org/mozilla-central/rev/aafba1319774
https://hg.mozilla.org/mozilla-central/rev/0ecdc8f7b354

Thanks glandium! Do you feel that's okay to have this backported to ESR78? If so, could you file the respective request? That would help us by having to ship two patches less in our pile of already existing patches. Thanks!

Comment on attachment 9169320 [details]
Bug 1651680 - Replace PKG_SKIP_STRIP with PKG_STRIP.

ESR Uplift Approval Request

• If this is not a sec:{high,crit} bug, please state case for ESR consideration: Fixes a build regression on TOR builds on esr78.
• User impact if declined:
• Fix Landed on Version:
• Risk to taking this patch: Low
• Why is the change risky/not risky? (and alternatives if risky): Practically speaking, this is a no-op for builds other than mingw.
• String or UUID changes made by this patch:

Comment on attachment 9169320 [details]
Bug 1651680 - Replace PKG_SKIP_STRIP with PKG_STRIP.

Makes life better for our Tor friends. Approved for 78.2esr.

https://hg.mozilla.org/releases/mozilla-esr78/rev/0af3b12bc9a0
https://hg.mozilla.org/releases/mozilla-esr78/rev/3ce564d77e5f