rust/src/libsyntax/ext/tt/transcribe.rs
blake2-ppc 5cfad6fbae syntax: Shrink enum Token and enum nonterminal
`enum Token` was 192 bytes (64-bit), as pointed out by pnkfelix; the only
bloating variant being `INTERPOLATED(nonterminal)`.

Updating `enum nonterminal` to use ~ where variants included big types,
shrunk size_of(Token) to 32 bytes (64-bit).

I am unsure if the `nt_ident` variant should have an indirection, with
ast::ident being only 16 bytes (64-bit), but without this, enum Token
would be 40 bytes.

A dumb benchmark says that compilation time is unchanged, while peak
memory usage for compiling std.rs is down 3%

Before::

    $ time ./x86_64-unknown-linux-gnu/stage1/bin/rustc --cfg stage1 src/libstd/std.rs
    19.00user 0.39system 0:19.41elapsed 99%CPU (0avgtext+0avgdata 627820maxresident)k
    0inputs+28896outputs (0major+228665minor)pagefaults 0swaps
    $ time ./x86_64-unknown-linux-gnu/stage1/bin/rustc -O --cfg stage1 src/libstd/std.rs
    31.64user 0.34system 0:32.02elapsed 99%CPU (0avgtext+0avgdata 629876maxresident)k
    0inputs+22432outputs (0major+229411minor)pagefaults 0swaps

After::

    $ time ./x86_64-unknown-linux-gnu/stage1/bin/rustc --cfg stage1 src/libstd/std.rs
    19.07user 0.45system 0:19.55elapsed 99%CPU (0avgtext+0avgdata 609384maxresident)k
    0inputs+28896outputs (0major+221997minor)pagefaults 0swaps

    $ time ./x86_64-unknown-linux-gnu/stage1/bin/rustc -O --cfg stage1 src/libstd/std.rs
    31.90user 0.34system 0:32.28elapsed 99%CPU (0avgtext+0avgdata 612080maxresident)k
    0inputs+22432outputs (0major+223726minor)pagefaults 0swaps
2013-08-11 06:56:07 +02:00

306 lines
9.8 KiB
Rust

// Copyright 2012 The Rust Project Developers. See the COPYRIGHT
// file at the top-level directory of this distribution and at
// http://rust-lang.org/COPYRIGHT.
//
// Licensed under the Apache License, Version 2.0 <LICENSE-APACHE or
// http://www.apache.org/licenses/LICENSE-2.0> or the MIT license
// <LICENSE-MIT or http://opensource.org/licenses/MIT>, at your
// option. This file may not be copied, modified, or distributed
// except according to those terms.
use ast;
use ast::{token_tree, tt_delim, tt_tok, tt_seq, tt_nonterminal,ident};
use codemap::{span, dummy_sp};
use diagnostic::span_handler;
use ext::tt::macro_parser::{named_match, matched_seq, matched_nonterminal};
use parse::token::{EOF, INTERPOLATED, IDENT, Token, nt_ident};
use parse::token::{ident_to_str};
use parse::lexer::TokenAndSpan;
use std::hashmap::HashMap;
use std::option;
///an unzipping of `token_tree`s
struct TtFrame {
forest: @mut ~[ast::token_tree],
idx: uint,
dotdotdoted: bool,
sep: Option<Token>,
up: Option<@mut TtFrame>,
}
pub struct TtReader {
sp_diag: @span_handler,
// the unzipped tree:
stack: @mut TtFrame,
/* for MBE-style macro transcription */
interpolations: HashMap<ident, @named_match>,
repeat_idx: ~[uint],
repeat_len: ~[uint],
/* cached: */
cur_tok: Token,
cur_span: span
}
/** This can do Macro-By-Example transcription. On the other hand, if
* `src` contains no `tt_seq`s and `tt_nonterminal`s, `interp` can (and
* should) be none. */
pub fn new_tt_reader(sp_diag: @span_handler,
interp: Option<HashMap<ident,@named_match>>,
src: ~[ast::token_tree])
-> @mut TtReader {
let r = @mut TtReader {
sp_diag: sp_diag,
stack: @mut TtFrame {
forest: @mut src,
idx: 0u,
dotdotdoted: false,
sep: None,
up: option::None
},
interpolations: match interp { /* just a convienience */
None => HashMap::new(),
Some(x) => x
},
repeat_idx: ~[],
repeat_len: ~[],
/* dummy values, never read: */
cur_tok: EOF,
cur_span: dummy_sp()
};
tt_next_token(r); /* get cur_tok and cur_span set up */
return r;
}
fn dup_tt_frame(f: @mut TtFrame) -> @mut TtFrame {
@mut TtFrame {
forest: @mut (*f.forest).clone(),
idx: f.idx,
dotdotdoted: f.dotdotdoted,
sep: f.sep.clone(),
up: match f.up {
Some(up_frame) => Some(dup_tt_frame(up_frame)),
None => None
}
}
}
pub fn dup_tt_reader(r: @mut TtReader) -> @mut TtReader {
@mut TtReader {
sp_diag: r.sp_diag,
stack: dup_tt_frame(r.stack),
repeat_idx: r.repeat_idx.clone(),
repeat_len: r.repeat_len.clone(),
cur_tok: r.cur_tok.clone(),
cur_span: r.cur_span,
interpolations: r.interpolations.clone(),
}
}
fn lookup_cur_matched_by_matched(r: &mut TtReader,
start: @named_match)
-> @named_match {
fn red(ad: @named_match, idx: &uint) -> @named_match {
match *ad {
matched_nonterminal(_) => {
// end of the line; duplicate henceforth
ad
}
matched_seq(ref ads, _) => ads[*idx]
}
}
r.repeat_idx.iter().fold(start, red)
}
fn lookup_cur_matched(r: &mut TtReader, name: ident) -> @named_match {
match r.interpolations.find_copy(&name) {
Some(s) => lookup_cur_matched_by_matched(r, s),
None => {
r.sp_diag.span_fatal(r.cur_span, fmt!("unknown macro variable `%s`",
ident_to_str(&name)));
}
}
}
#[deriving(Clone)]
enum lis {
lis_unconstrained,
lis_constraint(uint, ident),
lis_contradiction(~str),
}
fn lockstep_iter_size(t: &token_tree, r: &mut TtReader) -> lis {
fn lis_merge(lhs: lis, rhs: lis) -> lis {
match lhs {
lis_unconstrained => rhs.clone(),
lis_contradiction(_) => lhs.clone(),
lis_constraint(l_len, ref l_id) => match rhs {
lis_unconstrained => lhs.clone(),
lis_contradiction(_) => rhs.clone(),
lis_constraint(r_len, _) if l_len == r_len => lhs.clone(),
lis_constraint(r_len, ref r_id) => {
let l_n = ident_to_str(l_id);
let r_n = ident_to_str(r_id);
lis_contradiction(fmt!("Inconsistent lockstep iteration: \
'%s' has %u items, but '%s' has %u",
l_n, l_len, r_n, r_len))
}
}
}
}
match *t {
tt_delim(ref tts) | tt_seq(_, ref tts, _, _) => {
do tts.iter().fold(lis_unconstrained) |lis, tt| {
let lis2 = lockstep_iter_size(tt, r);
lis_merge(lis, lis2)
}
}
tt_tok(*) => lis_unconstrained,
tt_nonterminal(_, name) => match *lookup_cur_matched(r, name) {
matched_nonterminal(_) => lis_unconstrained,
matched_seq(ref ads, _) => lis_constraint(ads.len(), name)
}
}
}
// return the next token from the TtReader.
// EFFECT: advances the reader's token field
pub fn tt_next_token(r: &mut TtReader) -> TokenAndSpan {
// XXX(pcwalton): Bad copy?
let ret_val = TokenAndSpan {
tok: r.cur_tok.clone(),
sp: r.cur_span,
};
loop {
{
let stack = &mut *r.stack;
let forest = &mut *stack.forest;
if stack.idx < forest.len() {
break;
}
}
/* done with this set; pop or repeat? */
if ! r.stack.dotdotdoted
|| { *r.repeat_idx.last() == *r.repeat_len.last() - 1 } {
match r.stack.up {
None => {
r.cur_tok = EOF;
return ret_val;
}
Some(tt_f) => {
if r.stack.dotdotdoted {
r.repeat_idx.pop();
r.repeat_len.pop();
}
r.stack = tt_f;
r.stack.idx += 1u;
}
}
} else { /* repeat */
r.stack.idx = 0u;
r.repeat_idx[r.repeat_idx.len() - 1u] += 1u;
match r.stack.sep.clone() {
Some(tk) => {
r.cur_tok = tk; /* repeat same span, I guess */
return ret_val;
}
None => ()
}
}
}
loop { /* because it's easiest, this handles `tt_delim` not starting
with a `tt_tok`, even though it won't happen */
// XXX(pcwalton): Bad copy.
match r.stack.forest[r.stack.idx].clone() {
tt_delim(tts) => {
r.stack = @mut TtFrame {
forest: tts,
idx: 0u,
dotdotdoted: false,
sep: None,
up: option::Some(r.stack)
};
// if this could be 0-length, we'd need to potentially recur here
}
tt_tok(sp, tok) => {
r.cur_span = sp;
r.cur_tok = tok;
r.stack.idx += 1u;
return ret_val;
}
tt_seq(sp, tts, sep, zerok) => {
// XXX(pcwalton): Bad copy.
let t = tt_seq(sp, tts, sep.clone(), zerok);
match lockstep_iter_size(&t, r) {
lis_unconstrained => {
r.sp_diag.span_fatal(
sp, /* blame macro writer */
"attempted to repeat an expression \
containing no syntax \
variables matched as repeating at this depth");
}
lis_contradiction(ref msg) => {
/* FIXME #2887 blame macro invoker instead*/
r.sp_diag.span_fatal(sp, (*msg));
}
lis_constraint(len, _) => {
if len == 0 {
if !zerok {
r.sp_diag.span_fatal(sp, /* FIXME #2887 blame invoker
*/
"this must repeat at least \
once");
}
r.stack.idx += 1u;
return tt_next_token(r);
} else {
r.repeat_len.push(len);
r.repeat_idx.push(0u);
r.stack = @mut TtFrame {
forest: tts,
idx: 0u,
dotdotdoted: true,
sep: sep,
up: Some(r.stack)
};
}
}
}
}
// FIXME #2887: think about span stuff here
tt_nonterminal(sp, ident) => {
match *lookup_cur_matched(r, ident) {
/* sidestep the interpolation tricks for ident because
(a) idents can be in lots of places, so it'd be a pain
(b) we actually can, since it's a token. */
matched_nonterminal(nt_ident(~sn,b)) => {
r.cur_span = sp; r.cur_tok = IDENT(sn,b);
r.stack.idx += 1u;
return ret_val;
}
matched_nonterminal(ref other_whole_nt) => {
// XXX(pcwalton): Bad copy.
r.cur_span = sp;
r.cur_tok = INTERPOLATED((*other_whole_nt).clone());
r.stack.idx += 1u;
return ret_val;
}
matched_seq(*) => {
r.sp_diag.span_fatal(
r.cur_span, /* blame the macro writer */
fmt!("variable '%s' is still repeating at this depth",
ident_to_str(&ident)));
}
}
}
}
}
}