Add JSON AST dumping #12387
Add JSON AST dumping #12387
Conversation
| io::println(json); | ||
| } | ||
| return; | ||
| } |
There was a problem hiding this comment.
The compile_input function has managed to remain pretty clean since it was created, perhaps the json emission would be better suited inside of the phases themselves?
There was a problem hiding this comment.
This isn't possible without exiting the entire process in the phases, which they avoid by having compile_input return early for them if necessary.
There was a problem hiding this comment.
(That was the first thing I implemented, before noticing that issue)
There was a problem hiding this comment.
Couldn't you leave the modifications to stop_after_phase_X and then move the emission logic to the end of the phases themselves?
There was a problem hiding this comment.
oh, duh. will-do
On Wed, Feb 19, 2014 at 2:52 AM, Alex Crichton notifications@github.comwrote:
In src/librustc/driver/driver.rs:
@@ -531,7 +541,13 @@ pub fn compile_input(sess: Session, cfg: ast::CrateConfig, input: &Input,
write_out_deps(sess, input, &outputs, &expanded_crate).unwrap();
if stop_after_phase_2(sess) { return; }if stop_after_phase_2(sess) {if sess.opts.debugging_opts & session::AST_JSON != 0 {let json = json::PrettyEncoder::str_encode(&expanded_crate);io::println(json);}return;}Couldn't you leave the modifications to stop_after_phase_X and then move
the emission logic to the end of the phases themselves?—
Reply to this email directly or view it on GitHubhttps://github.com//pull/12387/files#r9856752
.
This is mostly useful for working on rustc, when one is unfamiliar with the AST a particular construct will produce. It's a -Z flag as it's very much for debugging. Closes #10485
| front::assign_node_ids_and_map::assign_node_ids_and_map(sess, krate)); | ||
|
|
||
| if sess.opts.debugging_opts & session::AST_JSON != 0 { | ||
| let mut stdout = io::stdout(); |
There was a problem hiding this comment.
Shouldn't this be buffered? (Otherwise every single write, including each and every : separator, will be a syscall.)
There was a problem hiding this comment.
Isn't stdout line buffered by default?
On Wed, Feb 19, 2014 at 7:39 PM, Huon Wilson notifications@github.comwrote:
In src/librustc/driver/driver.rs:
@@ -220,8 +231,16 @@ pub fn phase_2_configure_and_expand(sess: Session,
krate = time(time_passes, "prelude injection", krate, |krate|
front::std_inject::maybe_inject_prelude(sess, krate));
- time(time_passes, "assinging node ids and indexing ast", krate, |krate|
front::assign_node_ids_and_map::assign_node_ids_and_map(sess, krate))- let (krate, map) = time(time_passes, "assinging node ids and indexing ast", krate, |krate|
front::assign_node_ids_and_map::assign_node_ids_and_map(sess, krate));
- if sess.opts.debugging_opts & session::AST_JSON != 0 {
let mut stdout = io::stdout();Shouldn't this be buffered? (Otherwise every single write, including each
and every : separator will be a syscall.)—
Reply to this email directly or view it on GitHubhttps://github.com//pull/12387/files#r9890950
.
There was a problem hiding this comment.
The task-local stdout handle is line-buffered by default, but those created through io::stdout() are not buffered.
There was a problem hiding this comment.
feh.
On Wed, Feb 19, 2014 at 7:55 PM, Alex Crichton notifications@github.comwrote:
In src/librustc/driver/driver.rs:
@@ -220,8 +231,16 @@ pub fn phase_2_configure_and_expand(sess: Session,
krate = time(time_passes, "prelude injection", krate, |krate|
front::std_inject::maybe_inject_prelude(sess, krate));
- time(time_passes, "assinging node ids and indexing ast", krate, |krate|
front::assign_node_ids_and_map::assign_node_ids_and_map(sess, krate))- let (krate, map) = time(time_passes, "assinging node ids and indexing ast", krate, |krate|
front::assign_node_ids_and_map::assign_node_ids_and_map(sess, krate));
- if sess.opts.debugging_opts & session::AST_JSON != 0 {
let mut stdout = io::stdout();The task-local stdout handle is line-buffered by default, but those
created through io::stdout() are not buffered.—
Reply to this email directly or view it on GitHubhttps://github.com//pull/12387/files#r9891396
.
|
@cmr, also, while you're fixing the buffering ;), it would be good to add a simple run-make test, to ensure the flags keep doing something approximately sane. |
|
Awesome! |
fix(ide-db): correct single-file module rename
Fixes a bug where rust-analyzer would emit `WorkspaceEdit`s with paths to dirs instead of files for the following project layout.
lib.rs
```rust
mod foo;
```
foo.rs
```rust
mod bar {
struct Bar;
}
```
Also fixes emitted paths for modules with mod.rs.
The bug resulted in panic in helix editor when attempting to rename a module.
See the commits