Respect `-o` output file in `--print` by dtolnay · Pull Request #110785 · rust-lang/rust

dtolnay · 2023-04-25T01:02:16Z

Previously something like rustc --print=cfg -o cfg.txt just ignored the -o option and would print to stdout as usual.

File output simplifies integration in non-Cargo build systems — in my case Buck2. With this PR we get to skip a layer of Python wrapper or Bash wrapper to redirect rustc's stdout into the location dictated by the build system.

rustbot · 2023-04-25T01:02:22Z

r? @jackh726

(rustbot has picked a reviewer for you, use r? to override)

dtolnay · 2023-04-25T21:58:16Z

Rebased to resolve conflict with Fix ICE on --print=... i/o errors #110783

jyn514 · 2023-04-26T04:22:37Z

This makes sense to me, since we currently abort compilation early if you pass --print:

; rustc main.rs -o foo --print=target-list >/dev/null; file foo
foo: cannot open `foo' (No such file or directory)

That said, are we sure we want to commit to that behavior? Could there be a scenario where it makes sense to emit the --print info but still continue compilation? In that case, the -o argument would cause a collision, we'd try and output the binary and the print info to the same file.

apiraino · 2023-05-17T14:12:53Z

r? compiler

apiraino · 2023-06-07T10:28:13Z

(sorry for the pings)

r? compiler

fee1-dead · 2023-07-14T08:36:23Z

Thanks. Could you add an ui test for the llvm print changes? After that r=me.

bors · 2023-07-15T01:28:23Z

⌛ Testing commit 89a32a4 with merge 6cd0af32cea94ac27cc4e4135acdd5b7e111ee53...

bors · 2023-07-15T01:30:43Z

💔 Test failed - checks-actions

rust-log-analyzer · 2023-07-15T02:19:29Z

The job armhf-gnu failed! Check out the build log: (web)(plain)

Click to see the possible cause of the failure (guessed by this bot)

curl: (22) The requested URL returned error: 404 error: failed to download llvm from ci
help: old builds get deleted after a certain time
help: if trying to compile an old commit of rustc, disable `download-ci-llvm` in config.toml:
[llvm]
download-ci-llvm = false
Build completed unsuccessfully in 0:00:00

jyn514 · 2023-07-15T13:20:48Z

@bors retry #113713

bjorn3 · 2023-07-15T14:50:40Z

Doesn't this break -o foo --emit=link --print native-static-libs? Both --emit=link and --print native-static-libs would be fighting over writing foo.

jyn514 · 2023-07-15T14:58:28Z

@bors r-

bjorn3 · 2023-07-15T15:03:49Z

Maybe --print could be changed to allow explicitly specifying a path just like for --emit?

dtolnay · 2023-07-15T17:57:29Z

Doesn't this break -o foo --emit=link --print native-static-libs? Both --emit=link and --print native-static-libs would be fighting over writing foo.

Good question. This PR does not apply to native-static-libs. That's handled differently from all the other --print options: --print normally means print to stdout, whereas --print=native-static-libs gets emitted via diagnostics machinery instead (sess.emit_note) to stderr, and continues to work that way after this PR. As far as I can tell the intended way to use native-static-libs is in combination with --error-format=json and then searching for a note with prefix "native-static-libs:" among the JSON messages.

So this PR does not break it or cause fighting over writing the same output file with multiple contents. native-static-libs will be inconsistent vs the other --print options by not respecting -o, but has always been inconsistent already, so I'm not sure it ends up being worse.

Maybe --print could be changed to allow explicitly specifying a path just like for --emit?

I am open to this. --print sysroot=sysroot.txt --print cfg=cfg.txt

I think -o makes sense to apply to --print regardless of whether --print KIND=PATH is supported. -o means "set a location for the output that was asked for". In rustc --emit=metadata -o PATH, the -o means set the location for the rmeta output. In rustc --emit=asm -o PATH, the -o means set the location for the asm output. In rustc --print=cfg -o PATH, the configuration data is the output being asked for, and I think -o should apply.

If --print + -o does get supported, I'm not sure I would be able to justify --print KIND=PATH in addition. @bjorn3 do you feel that -o should not apply to --print, or do you feel that both --print + -o and --print KIND=PATH should work?

I can see that for native-static-libs, rustc --emit link=libfoo.a --print native-static-libs=native-static-libs.txt can be potentially more convenient than grabbing notes from a JSON diagnostic.

bjorn3 · 2023-07-15T18:25:59Z

This PR does not apply to native-static-libs.

Right. It also doesn't apply to --print link-args it seems even though that one doesn't use a diagnostic. That one doesn't do an early exit unlike most --print either.

If --print + -o does get supported, I'm not sure I would be able to justify --print KIND=PATH in addition. @bjorn3 do you feel that -o should not apply to --print, or do you feel that both --print + -o and --print KIND=PATH should work?

I think --print + -o should not be supported. For --emit + -o rustc changes the file extension if there are multiple outputs. For --print there are no obvious file extensions to use. I don't have any strong objection though. If it does get supported I think both --print + -o and --print KIND=PATH should be supported.

I can see that for native-static-libs, rustc --emit link=libfoo.a --print native-static-libs=native-static-libs.txt can be potentially more convenient than grabbing notes from a JSON diagnostic.

Indeed.

dtolnay

Closing in favor of --print KIND=PATH—#113780. Thank you for the review! I agree that is a better approach.

Support `--print KIND=PATH` command line syntax As is already done for `--emit KIND=PATH` and `-L KIND=PATH`. In the discussion of rust-lang#110785, it was pointed out that `--print KIND=PATH` is nicer than trying to apply the single global `-o` path to `--print`'s output, because in general there can be multiple print requests within a single rustc invocation, and anyway `-o` would already be used for a different meaning in the case of `link-args` and `native-static-libs`. I am interested in using `--print cfg=PATH` in Buck2. Currently Buck2 works around the lack of support for `--print KIND=PATH` by [indirecting through a Python wrapper script](https://github.com/facebook/buck2/blob/d43cf3a51a31f00be2c2248e78271b0fef0452b4/prelude/rust/tools/get_rustc_cfg.py) to redirect rustc's stdout into the location dictated by the build system. From skimming Cargo's usages of `--print`, it definitely seems like it would benefit from `--print KIND=PATH` too. Currently it is working around the lack of this by inserting `--crate-name=___ --print=crate-name` so that it can look for a line containing `___` as a delimiter between the 2 other `--print` informations it actually cares about. This is commented as a "HACK" and "abuse". https://github.com/rust-lang/cargo/blob/31eda6f7c360d9911f853b3014e057db61238f3e/src/cargo/core/compiler/build_context/target_info.rs#L242 (FYI `@weihanglo` as you dealt with this recently in rust-lang/cargo#11633.) Mentioning reviewers active in rust-lang#110785: `@fee1-dead` `@jyn514` `@bjorn3`

…li-obk Stabilize `PATH` option for `--print KIND=PATH` This PR propose stabilizing the `PATH` option for `--print KIND=PATH`. This option was previously added in rust-lang#113780 (as insta-stable before being un-stablized in rust-lang#114139). Description of the `PATH` option: > A filepath may optionally be specified for each requested information kind, in the format `--print KIND=PATH`, just like for `--emit`. When a path is specified, information will be written there instead of to stdout. ------ Description of the original PR [\[link\]](rust-lang#113780 (comment)): > **Support --print KIND=PATH command line syntax** > > As is already done for `--emit KIND=PATH` and `-L KIND=PATH`. > > In the discussion of rust-lang#110785, it was pointed out that `--print KIND=PATH` is nicer than trying to apply the single global `-o path` to `--print`'s output, because in general there can be multiple print requests within a single rustc invocation, and anyway `-o` would already be used for a different meaning in the case of `link-args` and `native-static-libs`. > > I am interested in using `--print cfg=PATH` in Buck2. Currently Buck2 works around the lack of support for `--print KIND=PATH` by [indirecting through a Python wrapper script](https://github.com/facebook/buck2/blob/d43cf3a51a31f00be2c2248e78271b0fef0452b4/prelude/rust/tools/get_rustc_cfg.py) to redirect rustc's stdout into the location dictated by the build system. > > From skimming Cargo's usages of `--print`, it definitely seems like it would benefit from `--print KIND=PATH` too. Currently it is working around the lack of this by inserting `--crate-name=___ --print=crate-name` so that it can look for a line containing `___` as a delimiter between the 2 other `--print` informations it actually cares about. This is commented as a "HACK" and "abuse". https://github.com/rust-lang/cargo/blob/31eda6f7c360d9911f853b3014e057db61238f3e/src/cargo/core/compiler/build_context/target_info.rs#L242 ----- cc `@dtolnay` r? `@jackh726`

Stabilize `PATH` option for `--print KIND=PATH` This PR propose stabilizing the `PATH` option for `--print KIND=PATH`. This option was previously added in rust-lang/rust#113780 (as insta-stable before being un-stablized in rust-lang/rust#114139). Description of the `PATH` option: > A filepath may optionally be specified for each requested information kind, in the format `--print KIND=PATH`, just like for `--emit`. When a path is specified, information will be written there instead of to stdout. ------ Description of the original PR [\[link\]](rust-lang/rust#113780 (comment)): > **Support --print KIND=PATH command line syntax** > > As is already done for `--emit KIND=PATH` and `-L KIND=PATH`. > > In the discussion of rust-lang/rust#110785, it was pointed out that `--print KIND=PATH` is nicer than trying to apply the single global `-o path` to `--print`'s output, because in general there can be multiple print requests within a single rustc invocation, and anyway `-o` would already be used for a different meaning in the case of `link-args` and `native-static-libs`. > > I am interested in using `--print cfg=PATH` in Buck2. Currently Buck2 works around the lack of support for `--print KIND=PATH` by [indirecting through a Python wrapper script](https://github.com/facebook/buck2/blob/d43cf3a51a31f00be2c2248e78271b0fef0452b4/prelude/rust/tools/get_rustc_cfg.py) to redirect rustc's stdout into the location dictated by the build system. > > From skimming Cargo's usages of `--print`, it definitely seems like it would benefit from `--print KIND=PATH` too. Currently it is working around the lack of this by inserting `--crate-name=___ --print=crate-name` so that it can look for a line containing `___` as a delimiter between the 2 other `--print` informations it actually cares about. This is commented as a "HACK" and "abuse". https://github.com/rust-lang/cargo/blob/31eda6f7c360d9911f853b3014e057db61238f3e/src/cargo/core/compiler/build_context/target_info.rs#L242 ----- cc `@dtolnay` r? `@jackh726`

rustbot assigned jackh726 Apr 25, 2023

rustbot added S-waiting-on-review Status: Awaiting review from the assignee but also interested parties. T-compiler Relevant to the compiler team, which will review and decide on the PR/issue. labels Apr 25, 2023

Mark-Simulacrum added the relnotes Marks issues that should be documented in the release notes of the next release. label Apr 25, 2023

dtolnay force-pushed the printoutfile branch from fb865ca to a6e0f93Compare April 25, 2023 21:58

rustbot assigned wesleywiser and unassigned jackh726 May 17, 2023

rustbot assigned fee1-dead and unassigned wesleywiser Jun 7, 2023

fee1-dead reviewed Jun 7, 2023
View reviewed changes

Comment threadcompiler/rustc_driver_impl/src/lib.rs Outdated
Comment threadcompiler/rustc_driver_impl/src/lib.rs Outdated

fee1-dead added S-waiting-on-author Status: This is awaiting some action (such as code changes or more information) from the author. and removed S-waiting-on-review Status: Awaiting review from the assignee but also interested parties. labels Jun 7, 2023

dtolnay added S-waiting-on-author Status: This is awaiting some action (such as code changes or more information) from the author. and removed S-waiting-on-author Status: This is awaiting some action (such as code changes or more information) from the author. labels Jul 13, 2023

Add test of --print + -o
34d3fcb

dtolnay force-pushed the printoutfile branch from a6e0f93 to f4c3009Compare July 14, 2023 00:45

dtolnay added 5 commits July 13, 2023 17:47

Respect -o output file in --print
a32f1ae

Implement -o in rustc_driver_impl using pretty::write_or_print
62ecfd3

Implement printing to file in codegen_backend.print
294a4f5

Implement printing to file in llvm_util
299d8b9

Implement printing to file in PassWrapper
df5dede

dtolnay force-pushed the printoutfile branch from f4c3009 to df5dedeCompare July 14, 2023 00:48

dtolnay added S-waiting-on-review Status: Awaiting review from the assignee but also interested parties. and removed S-waiting-on-author Status: This is awaiting some action (such as code changes or more information) from the author. labels Jul 14, 2023

bors added the S-waiting-on-bors Status: Waiting on bors to run and complete tests. Bors will change the label on completion. label Jul 15, 2023

bors added S-waiting-on-review Status: Awaiting review from the assignee but also interested parties. and removed S-waiting-on-bors Status: Waiting on bors to run and complete tests. Bors will change the label on completion. labels Jul 15, 2023

bors added S-waiting-on-bors Status: Waiting on bors to run and complete tests. Bors will change the label on completion. and removed S-waiting-on-review Status: Awaiting review from the assignee but also interested parties. labels Jul 15, 2023

bors added S-waiting-on-author Status: This is awaiting some action (such as code changes or more information) from the author. and removed S-waiting-on-bors Status: Waiting on bors to run and complete tests. Bors will change the label on completion. labels Jul 15, 2023

dtolnay added S-waiting-on-author Status: This is awaiting some action (such as code changes or more information) from the author. and removed S-waiting-on-author Status: This is awaiting some action (such as code changes or more information) from the author. labels Jul 16, 2023

dtolnay mentioned this pull request Jul 17, 2023
Support --print KIND=PATH command line syntax #113780
Merged

dtolnay commented Jul 17, 2023
View reviewed changes

dtolnay closed this Jul 17, 2023

dtolnay deleted the printoutfile branch July 17, 2023 06:33

Urgau mentioned this pull request Jul 28, 2023
Stabilize PATH option for --print KIND=PATH#114183
Merged

dtolnay removed the S-waiting-on-author Status: This is awaiting some action (such as code changes or more information) from the author. label Jan 14, 2025

Uh oh!

Uh oh!

Conversation

dtolnay commented Apr 25, 2023• edited Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

Uh oh!

rustbot commented Apr 25, 2023

Uh oh!

dtolnay commented Apr 25, 2023

Uh oh!

jyn514 commented Apr 26, 2023

Uh oh!

apiraino commented May 17, 2023

Uh oh!

apiraino commented Jun 7, 2023

Uh oh!

Uh oh!

Uh oh!

fee1-dead commented Jul 14, 2023

Uh oh!

bors commented Jul 15, 2023

Uh oh!

bors commented Jul 15, 2023

Uh oh!

rust-log-analyzer commented Jul 15, 2023

Uh oh!

jyn514 commented Jul 15, 2023

Uh oh!

bjorn3 commented Jul 15, 2023

Uh oh!

jyn514 commented Jul 15, 2023

Uh oh!

bjorn3 commented Jul 15, 2023

Uh oh!

dtolnay commented Jul 15, 2023

Uh oh!

bjorn3 commented Jul 15, 2023

Uh oh!

dtolnay left a comment

Choose a reason for hiding this comment

Uh oh!

Reviewers

Assignees

Labels

Projects

Milestone

Development

Uh oh!

11 participants

dtolnay commented Apr 25, 2023•
edited
Loading