Run on Rust stable

[SimonSapin]

Without --features unstable:

• Atom is 16 bytes instead of 8. (It has a drop flag.)
• ns! and atom! are giant generated macros instead of plugins, and so may increase compile times.

r? @glennw, particularly c4fa256.

Needed for servo/html5ever#53

[glennw]

Reviewed 13 of 14 files at r1.
Review status: 13 of 14 files reviewed at latest revision, 3 unresolved discussions, all commit checks successful.

---

^{.travis.yml, line 8 [r1] (raw file):}
Is this saying we don't run the tests on anything but nightly? Is there a reason for that?

---

^{src/atom/mod.rs, line 91 [r1] (raw file):}
I think this still handles the ABA case mentioned in the comments. Might be worth a double check though.

---

^{src/atom/mod.rs, line 234 [r1] (raw file):}
Do we still need this check for the optimized case?

---

Comments from the review on Reviewable.io

[SimonSapin]

Review status: 10 of 14 files reviewed at latest revision, 3 unresolved discussions, some commit checks pending.

---

^{.travis.yml, line 8 [r1] (raw file):}
Yes, because benchmarks and some tests require unstable features. I’ve added a commit to disable just those when --feature unstable is not given.

---

^{src/atom/mod.rs, line 234 [r1] (raw file):}
I don’t understand. Need what in what case? There is no refcount to decrement if an atom is not dynamic.

---

Comments from the review on Reviewable.io

[SimonSapin]

Review status: 10 of 14 files reviewed at latest revision, 3 unresolved discussions, all commit checks successful.

---

^{src/atom/mod.rs, line 91 [r1] (raw file):}
I believe the logic is the same as before in all three cases:

• ptr is null: allocate a new entry and return a pointer to that entry.
• ptr is not null, and the refcount was zero before we incremented it: decrement it again, allocate a new entry and return a pointer to that entry.
• ptr is not null, and the refcount was positive: return ptr unchanged.

---

Comments from the review on Reviewable.io

[bors-servo]

☔ The latest upstream changes (presumably #96) made this pull request unmergeable. Please resolve the merge conflicts.

[glennw]

Reviewed 1 of 14 files at r1.
Review status: 9 of 13 files reviewed at latest revision, 3 unresolved discussions, all commit checks successful.

---

Comments from the review on Reviewable.io

[glennw]

@SimonSapin looks good to me.

[SimonSapin]

@bors-servo r=glennw

[bors-servo]

📌 Commit f7fd964 has been approved by glennw

[bors-servo]

⌛ Testing commit f7fd964 with merge dd1b626...

[bors-servo]

☀️ Test successful - travis