Cleanup rmeta::MacroDef

[Centril]

Avoid using rountrip parsing in the encoder and in fn load_macro_untracked.

The main reason I was interested in this was to remove rustc_parse as a dependency of rustc_metadata but it seems like this had other benefits as well.

Fixes #49511.

r? @eddyb
cc @matthewjasper @estebank @petrochenkov

[Mark-Simulacrum]

Overall though I think this is reasonable, I guess :)

[Centril]

All the tests passed (before I adjusted the comment), so this should be ready for review now. :)

[petrochenkov]

Do not forget to run perf on this.
Stringification was a cheat, but it could be a performance-friendly cheat.

[Centril]

Let's try a perf run now (and maybe one later after doing some more changes according to #66364 (comment)):

@bors try @rust-timer queue

[rust-timer]

Awaiting bors try build completion

[bors]

⌛ Trying commit a2ec40b4bf40b57c3df20d4c679804812d3be8e6 with merge 8391cac23555ae55fb9f4d7fc909c58c3f1313bf...

[Centril]

@bors retry

[bors]

⌛ Testing commit bafa5cc with merge 7736d6d4c0c6ca1c4028ac34b424375d0a4f8068...

[rust-highfive]

Your PR failed (pretty log, raw log). Through arcane magic we have determined that the following fragments from the build log may contain information about the problem.

Click to expand the log.

I'm a bot! I can only do what humans tell me to, so if this was not helpful or you have suggestions for improvements, please ping or otherwise contact @rust-lang/infra. (Feature Requests)

[bors]

💔 Test failed - checks-azure

[Centril]

@bors retry

[bors]

⌛ Testing commit bafa5cc with merge c6224a4e9dbab86d6b743397f3c64dad86098a9f...

[Centril]

@bors retry

[rust-highfive]

Your PR failed (pretty log, raw log). Through arcane magic we have determined that the following fragments from the build log may contain information about the problem.

Click to expand the log.

I'm a bot! I can only do what humans tell me to, so if this was not helpful or you have suggestions for improvements, please ping or otherwise contact @rust-lang/infra. (Feature Requests)

[bors]

⌛ Testing commit bafa5cc with merge 5cffb60e74b9d4dca3beef2dd4ba4913a848a7e1...

[emilyalbini]

@bors retry

Yield to the beta release

[bors]

⌛ Testing commit bafa5cc with merge 1581278...

[bors]

☀️ Test successful - checks-azure
Approved by: petrochenkov,eddyb
Pushing 1581278 to master...

[Mark-Simulacrum]

Tagging as relnotes (so we don't forget about it -- it may not go into the final version), though that may be a false positive. @tomaka reported some breakage here (https://twitter.com/tomaka17/status/1238204339505770499) though it may be entirely in unstable features and unobservable on stable, in which case we wouldn't need to document it in relnotes.

@eddyb tells me @petrochenkov is the person to ask on whether there's something to say in relnotes here :)

[Mark-Simulacrum]

Further discussion on Discord informed me that perhaps there is not a significant change here and the only exposed changes are to unstable APIs in proc_macro anyway, so untagging.

[petrochenkov]

Spans can affect many things, starting from lints (as this PR itself shows).
I guess the beta crater run will show the full extent of changes this PR caused.

[petrochenkov]

Actually, running a separate crater check for this PR would be reasonable.