Incorrect report of ambiguous unicode characters

[sergeych]

Description

The wiki page viewer (and editor) show a warring on ambiguous unicode chararcters I tried for most non-latin1 characters. Please remove it at all. It is hardly useful even would it be functional (and it is not). It is not possible to make a good guess on which unicode is OK for a project with many localisations. The locale of the developer (reporter, wiki writer) computer/browser often will be different from the locale that is used to report a problem (in a localised text for example) or describe a test case described in the wiki page.

Gitea Version

1.18.5

Can you reproduce the bug on the Gitea demo site?

No

Log Gist

No response

Screenshots

Git Version

1.18.5

Operating System

Linux ** 4.19.0-21-amd64 #1 SMP Debian 4.19.249-2 (2022-06-30) x86_64

How are you running Gitea?

downloaded your binaries

Database

PostgreSQL

[sergeych]

looks like it is fixed to not yet included in the try.gitea version (1.20...). Hope it is fixed ;)

[lunny]

I think this is a known issue.

[wxiaoguang]

Known issues have been fixed. This could be a new issue.

If you could provide a reproducible page on try.gitea.io, maybe it would help some maintainers to take a look at it.

[wxiaoguang]

More or less related to this one, but not the same one:

• https://try.gitea.io/DarrenPIngram/test/src/branch/main/test
• Autocomplete creating problems with text entry, strange unicode (or user error) #20451

[sergeych]

Known issues have been fixed. This could be a new issue.

If you could provide a reproducible page on try.gitea.io, maybe it would help some maintainers to take a look at it.

I was not able to reproduce it on try.gitea - but the version there is a beta, while I'm using the latest stable. Hope it is fixed in beta (but strange it is not yet fixed in fresh new stable), but it is very likely its just a new "feature" not yet ported to the try.gitea. I sincerely hate the idea of trying to analyze unicode in the text field... Let browser do it. Practically no reason to spend effort to just make the life of gitea user just more complicated.

[wxiaoguang]

I was not able to reproduce it on try.gitea

Could you share the text contents of your screenshot? Wrap them here by ```

while I'm using the latest stable

Do you mean 1.18.x? If there is a chance, could you try 1.19.0-rc1? Most changes in try.gita.io are also in 1.19, but not in 1.18.

[ARNik]

Have the same issue with 1.19.0

[sergeych]

So any plan to fix it? Just remove the clumsy and useless check.

[wxiaoguang]

Could you share the text contents of your screenshot? Wrap them here by ```

---

How about introducing a global config option to disable the check for all repositories?

[lunny]

I like to add a global option to disable it but keep it open default as before.

[sergeych]

with due respect what is the reason to leave it on by default if it just does not works false reporting absolutely normal sequences?

…

On Sat, Apr 1, 2023 at 4:13 AM Lunny Xiao ***@***.***> wrote: I like to add a global option to disable it but keep it open default as before. — Reply to this email directly, view it on GitHub <#23149 (comment)>, or unsubscribe <https://github.com/notifications/unsubscribe-auth/AACJIIITP53YHHTBTHRSFADW66MNDANCNFSM6AAAAAAVH4EK3E> . You are receiving this because you authored the thread.Message ID: ***@***.***>

[wxiaoguang]

For security:

• https://www.securityweek.com/trojan-source-attack-abuses-unicode-inject-vulnerabilities-code/
• https://threatpost.com/trojan-source-invisible-bugs-source-code/175891/
• GitHub also warns on ambiguous unicode characters, while it does better

ps: Could you share the text contents of your screenshot? Wrap them here by ``` ? At least, it's easier for developer to reproduce your issue and do improvements in the future.

[sergeych]

The text was: ~~~

Важно: право на операцию импорта теперь имеют только те пользователи, у

которых имя (не логин!) кончаются на "$" ~~~

…

On Sat, Apr 1, 2023 at 4:36 AM wxiaoguang ***@***.***> wrote: For security: - https://www.securityweek.com/trojan-source-attack-abuses-unicode-inject-vulnerabilities-code/ - https://threatpost.com/trojan-source-invisible-bugs-source-code/175891/ - GitHub also warns on ambiguous unicode characters, while it does better ps: Could you share the text contents of your screenshot? Wrap them here by ``` ? At least, it's easier for developer to reproduce your issue and do improvements in the future. — Reply to this email directly, view it on GitHub <#23149 (comment)>, or unsubscribe <https://github.com/notifications/unsubscribe-auth/AACJIIMOYEB3Q3WDVMX6MKTW66PEBANCNFSM6AAAAAAVH4EK3E> . You are receiving this because you authored the thread.Message ID: ***@***.***>

[silverwind]

I assume wiki needs the same treatment as #22016, cc @zeripath.

[silverwind]

This should be an easy fix. Hide the warning on wiki pages via template condition, similar to what #22016 did for file render view.