As documented in the blocked bugs, Fx38 and later currently requires gcc-4.7 or later to compile. This is just to make the failure explicit. No GNU/Linux package maintainers seem to have an issue with this. There is a slight question mark over the host compiler for some B2G / Android builds, but this won't affect the host compiler's version anyway.

Here was a try push that built everything on try. This actually used the patch from bug 1142352 and bumped the host compiler to 4.7 as well. I think the B2G L failure is unrelated. https://treeherder.mozilla.org/#/jobs?repo=try&revision=656cbf6fe1e6

(In reply to Bob Owen (:bobowen) from comment #3) > I think the B2G L failure is unrelated. Looks like bug 1142495. Old parent rev?

(In reply to Ryan VanderMeulen [:RyanVM UTC-4] from comment #4) > (In reply to Bob Owen (:bobowen) from comment #3) > > I think the B2G L failure is unrelated. > > Looks like bug 1142495. Old parent rev? Ah yeah, thanks Ryan. Looks like that try push was just a few hours before that fix hit m-c.

Comment on attachment 8578701 [details] [diff] [review] Part 1: Bump our minimum supported GCC version for Gecko up from 4.6 to 4.7. Review of attachment 8578701 [details] [diff] [review]: ----------------------------------------------------------------- Please make sure bug 1142352 is updated accordingly.

remote: https://hg.mozilla.org/integration/mozilla-inbound/rev/dd58b992baf8 remote: https://hg.mozilla.org/integration/mozilla-inbound/rev/ab26af10cb05 Thanks Mike.

Bob, does this mean that we no longer use gcc 4.6 anywhere? If yes, there is a number of new C++11 features which we can now allow to use in our code.

(In reply to :Ehsan Akhgari (not reading bugmail, needinfo? me!) from comment #8) > Bob, does this mean that we no longer use gcc 4.6 anywhere? If yes, there > is a number of new C++11 features which we can now allow to use in our code. Barring the caveat in my dev-platform post about the host compiler I believe it does, yes. I was going to wait to see if there is agreement to get this uplifted before updating the two pages that I know of (below). But, I suppose it doesn't matter much as Gecko 38 already doesn't compile with 4.6, these patches just make it fail faster. https://developer.mozilla.org/en-US/docs/Using_CXX_in_Mozilla_code https://developer.mozilla.org/en-US/docs/Mozilla/Developer_guide/Build_Instructions/Linux_Prerequisites

Comment on attachment 8578701 [details] [diff] [review] Part 1: Bump our minimum supported GCC version for Gecko up from 4.6 to 4.7. Approval Request Comment [Feature/regressing bug #]: Bugs 1102195 and 1128413 introduced things to Fx38 that would not compile with gcc-4.6 [User impact if declined]: The impact will be to people attempting to build Firefox 38 with gcc-4.6. This takes a while to fail currently. This makes it fail during configure. [Describe test coverage new/current, TreeHerder]: n/a [Risks and why]: Low: the build already fails with gcc-4.6, this just makes it explicit. [String/UUID change made/needed]: None

Comment on attachment 8578703 [details] [diff] [review] Part 2: Bump the MFBT GCC requirement to 4.7. See comment 10.

https://hg.mozilla.org/mozilla-central/rev/dd58b992baf8 https://hg.mozilla.org/mozilla-central/rev/ab26af10cb05

Comment on attachment 8578701 [details] [diff] [review] Part 1: Bump our minimum supported GCC version for Gecko up from 4.6 to 4.7. As a compiler guy, I am glad to see this change! C++11 \o/

MDN pages mentioned in Comment 9 updated.

https://hg.mozilla.org/releases/mozilla-aurora/rev/ab5f08d31c49 https://hg.mozilla.org/releases/mozilla-aurora/rev/75b29a50fa63

Comment on attachment 8578703 [details] [diff] [review] Part 2: Bump the MFBT GCC requirement to 4.7. I forgot to approve this one too.

They both landed already.