rustc_parse_format: Fix character indices in find_skips

[osa1]

Fixes#81006

[rust-highfive]

r? @estebank

(rust-highfive has picked a reviewer for you, use r? to override)

[osa1]

I have two different regression tests that fail in different ways without this patch. I'll add those once I figure out how to make them pass with the specified error message.

[osa1]

I added two tests, both fail with errors like:

running 1 test
F
failures:
---- [ui] ui/issues/issue-81006-2.rs stdout ----
error: /home/omer/rust/rust/src/test/ui/issues/issue-81006-2.rs:6: unexpected error: '6:23: 6:25: 1 positional argument in format string, but no arguments were given'
error: 1 unexpected errors found, 0 expected errors not found
status: exit code: 1
command: "/home/omer/rust/rust/build/x86_64-unknown-linux-gnu/stage1/bin/rustc" "/home/omer/rust/rust/src/test/ui/issues/issue-81006-2.rs" "-Zthreads=1" "--target=x86_64-unknown-linux-gnu" "--error-format" "json" "-Zui-testing" "-Zdeduplicate-diagnostics=no" "-Zemit-future-incompat-report" "--emit" "metadata" "-C" "prefer-dynamic" "--out-dir" "/home/omer/rust/rust/build/x86_64-unknown-linux-gnu/test/ui/issues/issue-81006-2" "-A" "unused" "-Crpath" "-O" "-Cdebuginfo=0" "-Zunstable-options" "-Lnative=/home/omer/rust/rust/build/x86_64-unknown-linux-gnu/native/rust-test-helpers" "-L" "/home/omer/rust/rust/build/x86_64-unknown-linux-gnu/test/ui/issues/issue-81006-2/auxiliary"
unexpected errors (from JSON output): [
Error {
line_num: 6,
kind: Some(
Error,
),
msg: "6:23: 6:25: 1 positional argument in format string, but no arguments were given",
},
]
thread '[ui] ui/issues/issue-81006-2.rs' panicked at 'explicit panic', src/tools/compiletest/src/runtest.rs:1491:13
note: run with `RUST_BACKTRACE=1` environment variable to display a backtrace

Any ideas why? These are expect-fail tests and the stderr files match the actual error messages.

[osa1]

All fixed, this should be good to go

[estebank]

Great work @osa1! Could you squash your commits into a single one? After that ping me here and I'll merge it ASAP :)

[osa1]

@estebank ah, sorry, I thought CI squashes commits automatically. Just did it myself. Thanks!

[estebank]

For t-compiler: Nominating for beta backport to minimize time where the error is user visible. The change is small, seems quite safe and the current ICE precludes an actual usable error, leaving users that hit this to their own devices.

[estebank]

Thanks for the quick turnaround!

@bors r+ rollup

[bors]

📌 Commit 28501c0 has been approved by estebank

[m-ou-se]

This failed in a rollup (#81111 (comment)), as src/test/ui/issues/issue-81006.rs added by this PR is making that directory too big:

tidy error: following path contains more than 2669 entries, you should move the test to some relevant subdirectory (current: 2671): /checkout/src/test/ui/issues

That doesn't mean that this PR is to blame, as the blame is shared with a few thousand other files. But we can't merge it like this. :(

See #73494 and #80942 (comment).

@bors r-

[osa1]

Moved the test to ui/macros.

[m-ou-se]

@bors r=estebank

[bors]

📌 Commit 9111e9d has been approved by estebank