aboutsummaryrefslogtreecommitdiff
diff options
context:
space:
mode:
-rw-r--r--rfc.md32
1 files changed, 27 insertions, 5 deletions
diff --git a/rfc.md b/rfc.md
index 48d200fe7..c6cb4320f 100644
--- a/rfc.md
+++ b/rfc.md
@@ -393,21 +393,43 @@ plan is suggested:
393- No harm will be done as long as the new libsyntax exists as an 393- No harm will be done as long as the new libsyntax exists as an
394 experiemt on crates.io. However, actually using it in the compiler 394 experiemt on crates.io. However, actually using it in the compiler
395 and other tools would require massive refactorings. 395 and other tools would require massive refactorings.
396
397- Proposed syntax tree requires to keep the original source code
398 available, which might increase memory usage of the
399 compiler. However, it should be possible to throw the original tree
400 and source code away after conversion to HIR.
396 401
397# Rationale and alternatives 402# Rationale and alternatives
398[alternatives]: #alternatives 403[alternatives]: #alternatives
399 404
400- Incrementally add more information about source code to the current AST. 405- Incrementally add more information about source code to the current
401- Move the current libsyntax to crates.io as is. 406 AST.
407
408- Move the current libsyntax to crates.io as is. In the past, there
409 were several failed attempts to do that.
410
402- Explore alternative representations for the parse tree. 411- Explore alternative representations for the parse tree.
403- Use parser generator instead of hand written parser. 412
413- Use parser generator instead of hand written parser. Using the
414 parser from libsyntax directly would be easier, and hand-written
415 LL-style parsers usually have much better error recovery than
416 generated LR-style ones.
404 417
405# Unresolved questions 418# Unresolved questions
406[unresolved]: #unresolved-questions 419[unresolved]: #unresolved-questions
407 420
408- Is it at all possible to represent Rust parser as a pure function of 421- Is it at all possible to represent Rust parser as a pure function of
409 the source code? 422 the source code? It seems like the answer is yes, because the
423 language and especially macros were cleverly designed with this
424 use-case in mind.
425
426
410- Is it possible to implement macro expansion using the proposed 427- Is it possible to implement macro expansion using the proposed
411 framework? 428 framework? This is the main question of this RFC. The proposed
429 solution of synthesizing source code on the fly seems workable: it's
430 not that different from the current implementation, which
431 synthesizes token trees.
432
433
412- How to actually phase out current libsyntax, if libsyntax2.0 turns 434- How to actually phase out current libsyntax, if libsyntax2.0 turns
413 out to be a success? 435 out to be a success?