715397 - ability to build m-c on 10.7 (lion)

Status: RESOLVED FIXED

(Firefox Build System :: General, defect)

Description

[Dustin J. Mitchell [:dustin] (he/him)]

This is a follow-on to bug 674655, which gives us the ability to build 10.5-compatible (32-bit) firefox using the 10.6 SDK.

Once that's in place, to satisfy bug 715337, we need to be able to build with that SDK on a 10.7 system.

As I understand it, 10.7 requires XCode 4.x (and since we're using gcc, that means 4.1, which is not the latest), and XCode 4.x only has the 10.6 and 10.7 SDKs, so bug 674655 is definitely a prerequisite here.

Comment 1

[Rafael Ávila de Espíndola (:espindola) (not reading bugmail)]

I did a try push to
https://tbpl.mozilla.org/?tree=Try&rev=35b3fc2ac72e

The 10.5 debug build already failed as expected since it is running on a 10.5 machine with no 10.6 sdk.

If the rest is green we should be able to switch to the 10.6 soon after the debug builds are moved out of 10.5 machines.

Comment 2

[Rafael Ávila de Espíndola (:espindola) (not reading bugmail)]

A new push is at

https://tbpl.mozilla.org/?tree=Try&rev=84fad4f51ae3

It should fix the "Symbol not found: _SCDynamicStoreCreate" errors.

Comment 3

[Rafael Ávila de Espíndola (:espindola) (not reading bugmail)]

And yet a new one for the libstdc++ issue:

https://tbpl.mozilla.org/?tree=Try&rev=c071c5aaa866

looks like we are also not correctly detecting how to handle page protection.  I will try to debug that tomorrow.

Comment 4

[Rafael Ávila de Espíndola (:espindola) (not reading bugmail)]

A new try push with the non-hack version of the libstdc++ fix is at

https://tbpl.mozilla.org/?tree=Try&rev=0d5234270db4

Comment 5

[John Ford [:jhford] CET/CEST Berlin Time]

(In reply to Rafael Ávila de Espíndola (:espindola) from comment #4)
> A new try push with the non-hack version of the libstdc++ fix is at
> 
> https://tbpl.mozilla.org/?tree=Try&rev=0d5234270db4

Rafael, this looks really neat!  How far out do you think this bug is?

Comment 6

[Rafael Ávila de Espíndola (:espindola) (not reading bugmail)]

There is a new try on top of m-i in:

https://tbpl.mozilla.org/?tree=Try&rev=c7fec001ba99

The expected result is for all to be green, except for a failure in the 10.5 debug build which is running on a 10.5 machine. If that is really the case, this bug would then be fully blocking on us moving the 10.5 debug builds to a 10.6 machine.

Comment 7

[Rafael Ávila de Espíndola (:espindola) (not reading bugmail)]

A new push to try is in https://tbpl.mozilla.org/?tree=Try&rev=f3c8c19da938

We are being "conservative" back in bug 698827, so we can't actually check that in, but should be interesting to see if a debug build finds any other problems with the 10.6 sdk.

Comment 8

[Rafael Ávila de Espíndola (:espindola) (not reading bugmail)]

Attachment: Use the 10.6 sdk

We have to wait for bug 698827 to be fixed for this to land, but we should be able to do the code review in parallel.

https://tbpl.mozilla.org/?tree=Try&rev=c7fec001ba99

Comment 9

[Rafael Ávila de Espíndola (:espindola) (not reading bugmail)]

note: the new try push is https://tbpl.mozilla.org/?tree=Try&rev=f3c8c19da938
sorry for the confusion.

Comment 10

[(not currently active) Ted Mielczarek]

Comment on attachment 589377 [details] [diff] [review]
Use the 10.6 sdk

Review of attachment 589377 [details] [diff] [review]:
-----------------------------------------------------------------

::: build/macosx/universal/mozconfig.common
@@ +46,1 @@
>  ac_add_app_options x86_64 --with-macos-sdk=/Developer/SDKs/MacOSX10.6.sdk

You can combine these two lines into:
ac_add_options --with-macos-sdk=/Developer/SDKs/MacOSX10.6.sdk

Comment 11

[Rafael Ávila de Espíndola (:espindola) (not reading bugmail)]

https://tbpl.mozilla.org/?tree=Mozilla-Inbound&rev=429503206f26

Comment 12

[Marco Bonardo [:mak]]

https://hg.mozilla.org/mozilla-central/rev/429503206f26

Comment 13

[Mike Taylor [:bear]]

I believe we are done here - if not please reopen.