1515075 - Build failure: Error in script "maintenanceservice_installer.nsi" on line 184 -- aborting creation process

Status: RESOLVED FIXED

(Firefox :: Installer, defect)

Description

[Honza Bambas (:mayhemer)]

Win64 mingw build (standard dev config on windows)
m-c@d62e952be812

Build log snippet (./mach build):

 0:38.66 Processing script file: "maintenanceservice_installer.nsi" (ACP)
 0:38.86 !insertmacro: macro "_Or" requires 4 parameter(s), passed 2!
 0:38.86 Error in script "maintenanceservice_installer.nsi" on line 184 -- aborting creation process
 0:38.89 mozmake.EXE[5]: *** [c:/Mozilla/src/mozilla-central/toolkit/mozapps/installer/windows/nsis/makensis.mk:105: maintenanceservice_installer] Error 1
 0:38.90 mozmake.EXE[4]: *** [Makefile:22: libs] Error 2
 0:38.91 mozmake.EXE[3]: *** [c:/Mozilla/src/mozilla-central/config/recurse.mk:101: browser/libs] Error 2
 0:38.91 mozmake.EXE[2]: *** [c:/Mozilla/src/mozilla-central/config/recurse.mk:34: libs] Error 2
 0:38.92 mozmake.EXE[1]: *** [c:/Mozilla/src/mozilla-central/config/rules.mk:431: default] Error 2
 0:38.93 mozmake.EXE: *** [client.mk:125: build] Error 2

The affected line is probably here:
https://searchfox.org/mozilla-central/rev/13788edbabb04d004e4a1ceff41d4de68a8320a2/browser/installer/windows/nsis/maintenanceservice_installer.nsi#184


.mozconfig:
mk_add_options "export COMPILE_PDB_FLAG="
mk_add_options "export HOST_PDB_FLAG="
mk_add_options "export MOZ_DEBUG_FLAGS=-Z7"

ac_add_options --build-backends=-VisualStudio
ac_add_options --host=x86_64-pc-mingw32
ac_add_options --target=x86_64-pc-mingw32

export MOZ_DEBUG_SYMBOLS=1

ac_add_options --enable-application=browser
ac_add_options --enable-optimize
ac_add_options --disable-debug

ac_add_options --disable-logrefcnt
ac_add_options --disable-crashreporter
ac_add_options --disable-profiling

ac_add_options --enable-tests
ac_add_options --enable-warnings-as-errors
ac_add_options --enable-chrome-format=flat

Comment 1

[Robert Strong (they/them - no direct email)]

I suspect a clobber would fix this but could you try copying
<srcdir>/toolkit/mozapps/installer/windows/nsis/overrides.nsh 
to
<objdir>/browser/installer/windows/instgen/

and recompiling?

Comment 2

[Robert Strong (they/them - no direct email)]

I just verified that copying in the old overrides.nsh will cause this.

Comment 3

[Honza Bambas (:mayhemer)]

OK, so it seems that enforcing clobber is what we need here?  At least to w/a the problem locally.  I have just rebuilt the tree of the last working cset... :)

Comment 4

[Molly Howell (she/her) (no longer active)]

I'm surprised a clobber would be required for this change; I wonder if the need for that is another symptom of bug 1485057.

Comment 5

[Honza Bambas (:mayhemer)]

If you want me to do any experiments, let me know.

Comment 6

[Robert Strong (they/them - no direct email)]

(In reply to Matt Howell (he/him) [:mhowell] from comment #4)
> I'm surprised a clobber would be required for this change; I wonder if the
> need for that is another symptom of bug 1485057.
It might be similar in that when building the helper.exe it should have copied over the new overrides.nsh. The maintenanceservice_installer.exe should likely also copy over the new overrides.nsh but that was never added to makensis.mk like it was for helper.exe.

I found that

Comment 7

[Robert Strong (they/them - no direct email)]

I found that changing makensis.mk to include the copy of files like it has for helper.exe fixes this bug for me.

Comment 8

[Robert Strong (they/them - no direct email)]

Attachment: patch rev1

Comment 9

[Robert Strong (they/them - no direct email)]

Pushed to try
https://treeherder.mozilla.org/#/jobs?repo=try&revision=3cb86e095abf7652fdc4dbe308ea54466b03fc63

This won't fix bug 1485057 but it does fix this bug for me locally.

To check:
copy an old overrides.nsh into <objdir>/browser/installer/windows/instgen/
run mach build to verify that an incremental build fails
apply patch
run mach build to verify that an incremental build succeeds

Comment 10

[Jonathan Kew [:jfkthame]]

https://treeherder.mozilla.org/#/jobs?repo=try&revision=044a6eb886fbfba64438fa4c628549cfeeb17063

Comment 11

[Michael Shal [:mshal]]

Comment on attachment 9032205 [details] [diff] [review]
patch rev1

• $(RM) -r $(CONFIG_DIR)
• $(MKDIR) $(CONFIG_DIR)

It's not great that we do this in 3 different targets now, but from what I can tell they should all run serially so it shouldn't present a problem for now. It might make sense in the future to move these rules into a python script and use GENERATED_FILES to create maintenanceservice_installer.exe, helper.exe, and maybe setup.exe

Comment 12

[Robert Strong (they/them - no direct email)]

(In reply to Jonathan Kew (:jfkthame) from comment #10)

https://treeherder.mozilla.org/#/
jobs?repo=try&revision=044a6eb886fbfba64438fa4c628549cfeeb17063
jfkthame, it isn't clear to me whether you think those build failures are related to this patch. It looks to me like the installers built fine and there were errors unrelated to this patch.

Comment 13

[Jonathan Kew [:jfkthame]]

Huh, I don't think I consciously wrote comment 10 -- at least I don't recall doing so! I think treeherder helpfully posted that because I had the patch from this bug at the top of my patch queue when I pushed the try job. So yes, the errors there aren't related to this bug/patch at all.

Comment 14

[Pulsebot]

Pushed by [e-mail]:
https://hg.mozilla.org/integration/mozilla-inbound/rev/cbda9c487dca
Build failure: Error in script maintenanceservice_installer.nsi on line 184 -- aborting creation process. r=mshal

Comment 15

[Arthur Iakab [arthur_iakab]]

https://hg.mozilla.org/mozilla-central/rev/cbda9c487dca