Bert Frees comments

Results 564 comments of


                                            Bert Frees

UEB tables produce unusable output for unknown characters

> But, if the output is ordinary ASCII braille then something very weird is going on: The same test, is somehow BACK translating the digits back to punctuation: That doesn't...

Potential memory leaks and strange translations when using 'replace' or 'correct' opcodes

@egli Do you know what is the state of this issue?

Potential memory leaks and strange translations when using 'replace' or 'correct' opcodes

But can you explain what you have done?

Integrate with OSS-Fuzz

@egli Why is this issue open?

UEB rules for contracting compound words are not always applied

`sufword` matches characters when they are either a word or at the beginning of a word.

UEB rules for contracting compound words are not always applied

I think: `match %[_.] abc %[_.a] xyz`

UEB rules for contracting compound words are not always applied

I think redefining the character classes of some characters, or defining additional classes on characters (with `attribute` rules), makes most sense. In case some characters were specifically _not_ defined as...

the "there" contraction is mistranslated under these conditions

Thanks @jrbowden. Do we need to add more tests, or can this just be closed?

the "there" contraction is mistranslated under these conditions

I'll put it differently: would the table benefit from additional tests, or are there already sufficient tests to prevent any future regressions regarding this fixed issue?

the "there" contraction is mistranslated under these conditions

This issue was fixed in 3.28.