Bump serde to 1.0

[SimonSapin]

This is a rebase of #1359 with bitflags added to the tidy duplicate crates exception list, and version numbers incremented some more to catch up with master. Original work by @Eijebong.

Fixes #1359.

---

This change is 

[SimonSapin]

@bors-servo r=glennw,SimonSapin

[bors-servo]

📌 Commit da57667 has been approved by glennw,SimonSapin

[bors-servo]

⌛ Testing commit da57667 with merge 0fccde8...

[SimonSapin]

@bors-servo r=glennw,SimonSapin

[bors-servo]

📌 Commit dfeefbd has been approved by glennw,SimonSapin

[bors-servo]

⌛ Testing commit dfeefbd with merge 0030d93...

[bors-servo]

☀️ Test successful - status-travis
Approved by: glennw,SimonSapin
Pushing 0030d93 to master...

[staktrace]

@SimonSapin Is there a reason this also bumps core-text to 5.0 from 4.0? It makes webrender fail linking on OS 10.7 builders (which we are still using in m-c).

[staktrace]

Argh. it looks like core-text 4.0.0 uses core-graphics 0.7.0 which uses serde 0.9.15. core-text 5.0.0 on other other hand uses core-graphics 0.8.1 which doesn't need serde.

[SimonSapin]

I don’t know, I carried that particular change over from #1359. I assume that it’s a cascading effect of upgrading serde being a breaking change in public API, which causes breaking changes in public APIs of dependent crates, etc. What’s the linking issue?

[staktrace]

core-text 5.0.0 uses the CTFontCopyDefaultCascadeListForLanguages symbol which is only available on OS X 10.8+. So it causes a linker failure, e.g. try push

[SimonSapin]

So this update pull in unrelated changes that had landed in the meantime in core-text-rs? Sounds like that change is what should be fixed.

[staktrace]

Yeah I'm following up on servo/core-text-rs#57 (comment) as well. That's probably a better way to resolve this.