Inconsistent link escaping mechanism

[Dirk Stöcker]

In Trac a link can be escaped with a !, like in ([42])

![42]

The problem is, that this depends on the content of the brackets. e.g. this wont work (![foo]) and print the ! sign as well.

![foo]

This means the escape mechanism on pages depends on the future changes of the wiki (in this case the status will change as soon as somebody adds a page called foo.

Especially strange as it also depends on potential link status like for (![FooBar]) which cannot only be escaped with another method ([FooBar]).

![FooBar] doesn't work
[!FooBar] works

My suggestion. ![ should always escape the contents. If the ! should be printed a second ! needs to be added !![ (like in ![42]):

!![42]

That way it would be consistent and not depending on possible future changes.

[Jun Omae]

[...] is link to a revision in a repository. If the revision part doesn't match with [0-9]+ or [0-9a-f]+, the text is not convert a link.

• [42] (Numeric only)
• [42/mirror] (Numeric with repository name)
• [92f432b3] (Hex only)
• [92f432b3/jomae.git] (Hex with repository name)

A user could use the following workarounds:

• Enclose with `...` or {{{...}}} (e.g. `[42]` → [42])
• [[html()] macro (e.g. [[html([42])]] → [42])

I think that it might be good to add a syntax like `...` without any markups.

[Dirk Stöcker]

Replying to Jun Omae:

[...] is link to a revision in a repository. If the revision part doesn't match with [0-9]+ or [0-9a-f]+, the text is not convert a link.

Don't focus too much on the example. The effect is true for other working links (but not CamelCase): [/Clients] ![/Clients] vs [Clients] [Clients]

I think that it might be good to add a syntax like `...` without any markups.

That would be another solution.