update backtrace formating text tips

[F001]

Follow up with #38165, improve consistency.

[rust-highfive]

Thanks for the pull request, and welcome! The Rust team is excited to review your changes, and you should hear from @arielb1 (or someone else) soon.

If any changes to this PR are deemed necessary, please add them as extra commits. This ensures that the reviewer can see what has changed since they last reviewed the code. Due to the way GitHub handles out-of-date commits, this should also make it reasonably obvious what issues have or haven't been addressed. Large or tricky changes may require several passes of review and changes.

Please see the contribution instructions for more information.

[Mark-Simulacrum]

Looks like a test needs to be updated.

[00:50:03] failures:
[00:50:03] 
[00:50:03] ---- [run-pass] run-pass/multi-panic.rs stdout ----
[00:50:03] 	
[00:50:03] error: test run failed!
[00:50:03] status: exit code: 101
[00:50:03] command: /checkout/obj/build/x86_64-unknown-linux-gnu/test/run-pass/multi-panic.stage2-x86_64-unknown-linux-gnu 
[00:50:03] stdout:
[00:50:03] ------------------------------------------
[00:50:03] 
[00:50:03] ------------------------------------------
[00:50:03] stderr:
[00:50:03] ------------------------------------------
[00:50:03] thread 'main' panicked at 'assertion failed: `(left == right)` (left: `Some("note: Run with `RUST_BACKTRACE=full` for a backtrace.")`, right: `Some("note: Run with `RUST_BACKTRACE=1` for a backtrace.")`)', /checkout/src/test/run-pass/multi-panic.rs:19
[00:50:03] note: Run with `RUST_BACKTRACE=full` for a backtrace.
[00:50:03] 
[00:50:03] ------------------------------------------
[00:50:03] 
[00:50:03] thread '[run-pass] run-pass/multi-panic.rs' panicked at 'explicit panic', /checkout/src/tools/compiletest/src/runtest.rs:2472
[00:50:03] note: Run with `RUST_BACKTRACE=full` for a backtrace.
[00:50:03] 
[00:50:03] 
[00:50:03] failures:
[00:50:03]     [run-pass] run-pass/multi-panic.rs
[00:50:03] 
[00:50:03] test result: FAILED. 2682 passed; 1 failed; 5 ignored; 0 measured

[F001]

@Mark-Simulacrum Fixed. Do you have other concerns?

[Mark-Simulacrum]

Out of interest, why did you add RUST_BACKTRACE=no as an option? Otherwise, I'm going to leave for @arielb1 to review the heart of the changes, I don't really know this area of the compiler.

[F001]

Well, this option is not added by this PR. It has been added by the PR #38165, but I think not all the occurrences are updated accordingly.

[arielb1]

r? @alexcrichton

[alexcrichton]

Thanks for the PR! Could you expand a bit on the rationle though? I'm not sure why we'd suggest RUST_BACKTRACE=full to start with, it seems like you'd only want to reach for that if you saw a backtrace and already determined that it may not be good enough.

[F001]

The rationale is that, according to #38165, the options are defined as:

• RUST_BACKTRACE=full prints all the informations (current behaviour)
• RUST_BACKTRACE=(short|yes) is the default
• RUST_BACKTRACE=(0|no) disables the backtrace.
• RUST_BACKTRACE= is equivalent to short for backward compatibility.

So I think the suggestion should be updated to RUST_BACKTRACE=(full|short|yes), and the old style RUST_BACKTRACE=(1|0) should not be the official suggestion.

I'm not sure which one is the best either, (full|short|yes) all seems reasonable to me. They are all better than (1|0).

[F001]

@alexcrichton If you think full is not appropriate in the suggestion, I'll update it to yes.

[F001]

Another alternative, use longer sentence for the suggestion:

note: Run with RUST_BACKTRACE=yes for a backtrace, and run with RUST_BACKTRACE=full for a verbose backtrace.

[alexcrichton]

Ah yes sorry, so I'm not personally a fan of the simultaneous change in interface here either, I'd prefer to not change how we parse values just yet. What I'm worried about though is that this is recommending RUST_BACKTRACE=full before RUST_BACKTRACE=1, which seems backwards?

[F001]

Yeah, I understand your concern. Now I have changed the recommending to RUST_BACKTRACE=yes which has the same behavior as RUST_BACKTRACE=1.

The motivation is that right now option 1 is supported only for backward compatibility, the option short|yes is more official with the same behavior.

I would like to hear suggestion from @Yamakaky too, do you think we should update the recommending words?

[Yamakaky]

My rational for only suggesting the short back trace was that new users will not want the full one in almost any case, and advanced users will know about the full version.

[F001]

Yeah~ I agree to suggesting the short style too. I just think yes is better than 1.

[Yamakaky]

I do too! I think yes/short, full and no should be good, with the old values as compat.

[F001]

@alexcrichton PR is updated.

[alexcrichton]

Ok not recommending "full" sounds good to me, but otherwise I'm still not sure I understand the rationale for this PR? Why change from "1" to "yes"? Why add "no" as a way to turn it off?

[F001]

I thought that the options of RUST_BACKTRACE had been changed to (full|short|yes|no) as described in #38165 , and it has been an agreement in core team. But It turns out to be not true.

I'm sorry for the confusion I caused. Therefore, I'm going to close this PR.