tigefa4u projects Trac

Robust coupling of Wiki lists and Wiki quotes

With [3316], new indent levels are only introduced when they need to be.

However, the implicit relationship that existed between the
indentation level of the lists and the indentation level of the blocks
was lost.

We therefore keep track of the indentation level introduced by
either of them, referred to as "tabstops".

That way, the current indentation is remembered even when
empty lines are introduced for clarity, and the appropriate
indenting is used for quotes following lists or quotes:

{{{
* first level
* second level

A second level quote
}}}

Tab characters are now converted before the line is parsed.

Fixes #124.

More robust parsing and formatting of Wiki lists (#124)

Previously, each indent of 2 spaces would introduce a new indentation level.
This lead to a confusing output in a variety of situations,
especially in changelog style commit messages.

Now, new indent levels are only introduced when they need to be,
i.e. when there's actually a new indented list or paragraph:

{{{
* first level
* second level
- third level

First level
Second level
}}}

Note that a tab character is converted to 8 space characters
before computing the indentation level.

In addition, lists can now also be started using "-" (as this is commonly found
in changelogs as well).

There will be a follow-up changeset focused on the interaction between lists and quotes.

Adding sections to the wiki-tests file.

Added unit test for mixed lists with multi-line items.

Fixed a subtle unicode issue related to string expansion when using an `trac.util.markup.Element` argument.

The result of a string expansion (e.g. `"%s" % text`) is either a `str` object or an `unicode` object, depending on the arguments:
* if there's at least one `unicode` argument, the result will be `unicode`
* otherwise, it will be a `str` object, regardless of the actual type of
the argument __after__ the call to `__str__`.

So in this case, doing `"</p>%s<p>" % text`
when `text` is a `trac.util.markup.Element` object
will result in the `unicode` object returned by
`Fragment.__str__` to be coerced into a `str` object,
which in turn may raise the usual `UnicodeEncodeError` ...

Quick fix for the log TracLinks syntax involving hexa ids: a minimal length is now required if hexa numbers are used.

Added a `send_converted` helper method to `Mimeview`.

It can be used in modules that provide alternative download links.

Small fixes for the `IContentConverter`s in the ticket subsystem:

* ''Comma-delimited Text'' uses the `text/csv` MIME Type, as per rfc:4180 (actually, that was already done in r3309)
* ''Tab-delimited Text'' now uses the [http://www.iana.org/assignments/media-types/text/tab-separated-values text/tab-separated-values] MIME Type.

Use the wikified changeset log message in order to compute the ETag of a changeset view. Fixes #3116.

Small fixes for the `IContentConverter`s in the ticket subsystem:

* use `trac.ticket.Ticket` instead of `trac.ticket.model.Ticket`
* use `trac.ticket.Query` instead of `trac.ticket.query`
* use quality `8` instead of `9`, so that it's easily possible for
the user to force the usage of a custom converter, while staying
in the 0-9 range.
* make the conversion tests more platform independant (both modified
tests failed for me on Windows)