Fix find_width_of_character_at_span bounds check

[etaoins]

Commit 0bd9667 added bounds checking of our current target byte position to prevent infinite loops. Unfortunately it was comparing the file-relative target versus the global file_start_pos and file_end_pos.

The result is failing to detect multibyte characters unless their file-relative offset fit within their global offset. This causes other parts of the compiler to generate spans pointing to the middle of a
multibyte character which will ultimately panic in bytepos_to_file_charpos.

Fix by comparing the target to the total file size when moving forward and doing checked subtraction when moving backwards. This should preserve the intent of the bounds check while removing the offset confusion.

cc @davidtwco

Fixes#48508

[rust-highfive]

r? @nikomatsakis

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

[estebank]

Could you add a test to catch any regressions?

[etaoins]

@estebank Added a new commit with a test

[estebank]

@bors r+

[bors]

📌 Commit c237d4f has been approved by estebank

[nikomatsakis]

r? @estebank

[Manishearth]

Fails pretty:

[01:46:36] failures:
[01:46:36] [01:46:36] ---- [pretty] run-pass/issue-48508.rs stdout ----
[01:46:36] [01:46:36] error: pretty-printed source does not typecheck
[01:46:36] status: exit code: 101
[01:46:36] command: "/checkout/obj/build/x86_64-unknown-linux-gnu/stage2/bin/rustc" "-" "-Zno-trans" "--out-dir" "/checkout/obj/build/x86_64-unknown-linux-gnu/test/run-pass/issue-48508.pretty-out" "--target=x86_64-unknown-linux-gnu" "-L" "/checkout/obj/build/x86_64-unknown-linux-gnu/test/run-pass" "-L" "/checkout/obj/build/x86_64-unknown-linux-gnu/test/run-pass/issue-48508.stage2-x86_64-unknown-linux-gnu.pretty.aux" "-Crpath" "-O" "-Zmiri" "-Zunstable-options" "-Lnative=/checkout/obj/build/x86_64-unknown-linux-gnu/native/rust-test-helpers" "-g"
[01:46:36] stdout:
[01:46:36] ------------------------------------------
[01:46:36] [01:46:36] ------------------------------------------
[01:46:36] stderr:
[01:46:36] ------------------------------------------
[01:46:36] error[E0425]: cannot find function `other` in module `other_file`
[01:46:36] --> <anon>:25:25
[01:46:36] |
[01:46:36] 25 | fn main() { other_file::other(); }
[01:46:36] | ^^^^^ not found in `other_file`
[01:46:36] [01:46:36] error: aborting due to previous error
[01:46:36] [01:46:36] If you want more information on this error, try using "rustc --explain E0425"
[01:46:36] [01:46:36] ------------------------------------------
[01:46:36] [01:46:36] thread '[pretty] run-pass/issue-48508.rs' panicked at 'explicit panic', tools/compiletest/src/runtest.rs:2893:9
[01:46:36] note: Run with `RUST_BACKTRACE=1` for a backtrace.
[01:46:36] [01:46:36] [01:46:36] failures:
[01:46:36] [pretty] run-pass/issue-48508.rs
[01:46:36] [01:46:36] test result: �[31mFAILED�(B�[m. 2879 passed; 1 failed; 49 ignored; 0 measured; 0 filtered out
[01:46:36] [01:46:36] thread 'main' panicked at 'Some tests failed', tools/compiletest/src/main.rs:476:22

[etaoins]

@Manishearth That looks like a bug in pretty, no? The test compiles and runs correctly so I'm not sure why pretty can't typecheck it.

Is it worth committing this without the test?

cc @estebank

[Manishearth]

Shouldn't this file be in the auxiliary folder?

[etaoins]

Everything in auxillary is treated as a crate. This needs to be a module to trigger #48508

[Manishearth]

Ah, ok

[Manishearth]

Perhaps. The path flag may not work if the pretty file is compiled elsewhere. ignore-pretty is an acceptable fix

[etaoins]

@Manishearth@estebank

Looks like this was triggering #37195. Pushed an ignore-pretty to work around it.

[Manishearth]

@bors r=estebank

[bors]

📌 Commit 363d604 has been approved by estebank