Port SeaMonkey bug 735618. (Untested (yet).) Is there a syntax to add/merge "#ifndef MOZ_WIN32_REDIST" to "#if _MSC_VER != xx00" checks?

Comment on attachment 606120 [details] [diff] [review] (Av1) Stop trying to package MSVC dlls when WIN32_REDIST_DIR isn't set [Backed out: Comment 12] And do you want to remove the "#ifdef XP_WIN32" in package-manifest.in?

(In reply to Serge Gautherie (:sgautherie) from comment #1) > (Untested (yet).) Succeeded as https://tbpl.mozilla.org/?tree=Try&rev=1ab2ce9d7277 > Is there a syntax to add/merge "#ifndef MOZ_WIN32_REDIST" to "#if _MSC_VER > != xx00" checks? Eventually found Preprocessor.py and Expression.py: the answer is 'no'. (Though I may investigate a workaround.)

(In reply to Serge Gautherie (:sgautherie) from comment #2) > And do you want to remove the "#ifdef XP_WIN32" in package-manifest.in? As this depends on http://mxr.mozilla.org/mozilla-central/source/build/win32/Makefile.in#68 I believe its safer to leave it in.

(In reply to Serge Gautherie (:sgautherie) from comment #3) > (Though I may investigate a workaround.) Here it is: I suggest to push the merge of the two patches. Succeeded as https://tbpl.mozilla.org/?tree=Try&rev=9590d17db91d

Comment on attachment 606120 [details] [diff] [review] (Av1) Stop trying to package MSVC dlls when WIN32_REDIST_DIR isn't set [Backed out: Comment 12] It occurred to me that Ted is not listed as a Firefox peer: should probably find someone in that list to review Firefox-only (packaging) patches.

Comment on attachment 606120 [details] [diff] [review] (Av1) Stop trying to package MSVC dlls when WIN32_REDIST_DIR isn't set [Backed out: Comment 12] I hereby formally delegate full review authority to Ted!

Comment on attachment 606120 [details] [diff] [review] (Av1) Stop trying to package MSVC dlls when WIN32_REDIST_DIR isn't set [Backed out: Comment 12] Whoever reviews these patches first.

Comment on attachment 606120 [details] [diff] [review] (Av1) Stop trying to package MSVC dlls when WIN32_REDIST_DIR isn't set [Backed out: Comment 12] <khuey> sgautherie: I think ted should look at that [bug]

Comment on attachment 606120 [details] [diff] [review] (Av1) Stop trying to package MSVC dlls when WIN32_REDIST_DIR isn't set [Backed out: Comment 12] Review of attachment 606120 [details] [diff] [review]: ----------------------------------------------------------------- I decided to go ahead and do these myself.

Comment on attachment 606120 [details] [diff] [review] (Av1) Stop trying to package MSVC dlls when WIN32_REDIST_DIR isn't set [Backed out: Comment 12] https://hg.mozilla.org/mozilla-central/rev/b972b89518c3 Cv1 = Av1 + Bv1. [Approval Request Comment] Regression caused by (bug #): (old) User impact if declined: None, but blocks landing bug 713132. Testing completed (on m-c, etc.): Try, this comment. Risk to taking this patch (and alternatives if risky): None, buildtime (+ install) only. String changes made by this patch: None.

Turns out our preprocessor doesn't understand #if FOO == bar https://hg.mozilla.org/mozilla-central/rev/9989b866c3a8

Historically we have treated the package-manifest.in file as part of the Build Config module, despite it living in the browser/installer directory.

(In reply to Ted Mielczarek [:ted] from comment #13) > Historically we have treated the package-manifest.in file as part of the > Build Config module, despite it living in the browser/installer directory. FWIW, I consider Firefox's package manifest it to be very much part of the Firefox module, given its purpose. Of course I'm happy to delegate responsibility for reviewing non-controversial changes to build config peers (or anyone else who fully understands the implication of such changes, really).

Doesn't really matter to me, honestly. :) It's sort of a gray area, but most changes to it are trivial anyway.

Cv1, but moving new logic to Makefile, as Preprocessor.py supports only 1 level of substitution :-| http://ftp.mozilla.org/pub/mozilla.org/firefox/try-builds/sgautherie.bz@free.fr-58b0579a1cac/try-win64/firefox-14.0a1.en-US.win64-x86_64.zip removed-files is correct now.

(In reply to Serge Gautherie (:sgautherie) from comment #16) > as Preprocessor.py supports only 1 level of substitution :-| Ftr, https://developer.mozilla.org/en/Build/Text_Preprocessor http://mxr.mozilla.org/mozilla-central/source/config/Preprocessor.py http://mxr.mozilla.org/mozilla-central/source/config/Expression.py

Try run for 58b0579a1cac is complete. Detailed breakdown of the results available here: https://tbpl.mozilla.org/?tree=Try&rev=58b0579a1cac Results (out of 14 total builds): success: 14 Builds (or logs if builds failed) available at: http://ftp.mozilla.org/pub/mozilla.org/firefox/try-builds/sgautherie.bz@free.fr-58b0579a1cac

Comment on attachment 608985 [details] [diff] [review] (Cv2) Stop trying to package MSVC dlls when WIN32_REDIST_DIR isn't set Review of attachment 608985 [details] [diff] [review]: ----------------------------------------------------------------- ::: browser/installer/Makefile.in @@ +95,5 @@ > DEFINES += -DMOZ_CHILD_PROCESS_NAME=$(MOZ_CHILD_PROCESS_NAME) > > +# Set MSVC dlls version to package, if any. > +ifndef WIN32_REDIST_DIR > +DEFINES += -DMOZ_MSVC_REDIST=-1 This doesn't seem necessary. If MOZ_MSVC_REDIST is not defined, then all your #if conditions will evaluate to false.

Cv2, with comment 19 suggestion(s).

Comment on attachment 610544 [details] [diff] [review] (Cv3) Stop trying to package MSVC dlls when WIN32_REDIST_DIR isn't set [Checked in: Comment 21] http://hg.mozilla.org/mozilla-central/rev/92fe907ddac8