Unescaped final Soft characters in LinkFormattingTest.txt #527

[macchiati]

There was a bad parameter passed in when processing the wiki URLs for use in the test data file. This exposed a bug in how that parameter was used. This corrects that, and gives the parameters more appropriate names (using an enum instead of a naked boolean).

Fixes https://github.com/unicode-org/properties/issues/527

NOTE: I could also regenerate the v17 file if needed. It has ~10 fixes also.

• Approver: Feel free to merge on my behalf
  • rebase & merge one or more commits
  • squash & merge multiple commits into one

[markusicu]

NOTE: I could also regenerate the v17 file if needed. It has ~10 fixes also.

I think we should treat published files as immutable, as we normally do.

[macchiati]

NOTE: I could also regenerate the v17 file if needed. It has ~10 fixes also.

I think we should treat published files as immutable, as we normally do.

Of course. The question is whether we should make a 17.1 file available, or document the problem so that people don't stumble over it until they implement 18.

[markusicu]

output lgtm now.

looks like spotless is unhappy, and maybe some other CI check :-/

[macchiati]

I think this should do it; I am also escaping any final Soft characters in the full-escaped string.

[markusicu]

lgtm tnx

[macchiati]

BTW, if we want to be more explicit, we could make the following change in the header of the test file.

OLD

# The fully-escaped field percent-escapes all literal syntax characters and all characters above ASCII.

NEW

# The fully-escaped field percent-escapes all code points based on https://url.spec.whatwg.org/#percent-encoded-bytes.
# This means all literal syntax characters in each Part and all code points above ASCII.
# It also percent-escapes the last character, if it is Link_Term:Soft.

[markusicu]

BTW, if we want to be more explicit, we could make the following change in the header of the test file.

good idea