pymake considers that a colon followed by a slash or backslash is not to be treated as the usual target/dependencies separator. But it doesn't do that for includedep files. But these files contain full windows paths, including the drive letter, which means the drive letter is considered a target, and everything that follows, dependencies.

Comment on attachment 802718 [details] [diff] [review] Fix pymake handling of colon in includedep files [Approval Request Comment] Bug caused by (feature/regressing bug #): bug 852249 User impact if declined: Windows builds may fail after some classes of changes because dependencies are not fully defined. Testing completed (on m-c, etc.): Landed on m-i today. Tested locally by avih. Risk to taking this patch (and alternatives if risky): Should be low risk String or IDL/UUID changes made by this patch: None

(In reply to Mike Hommey [:glandium] from comment #3) > User impact if declined: Windows builds may fail after some classes of > changes because dependencies are not fully defined. This has been around for a while - isn't the risk of this breaking something in the 25 timeframe very unlikely since it'll already be frozen? And wouldn't we know if it broke?

Comment on attachment 802718 [details] [diff] [review] Fix pymake handling of colon in includedep files 26 is on Aurora now and from Alex's comment it looks unlikely this will get uplift approval for Beta but if there's a significant reward please re-nominate for approval-beta