tigefa4u projects Trac

0.12.2rc1: updated ChangeLog and version tag.

0.12.2dev: add missing ''native'' eol-style

0.12.2dev: merged TracGuide updates from t.e.o in preparation for 0.12.2rc1.

0.12.2dev: `TimeoutError` shows original error on the same line (improve debugging of #9916)

0.12.2dev: in session handling, split purge transaction in two (#9705).

The previous change was probably not enough, as MySQL uses locks at the row level, not at table level.

Suppose that a purge transaction A first gets hold on a large number of session_attribute rows for deleting them, then tries to do the same for a large number of rows in the session session. At the same time, another transaction B could well have already locked one of these session rows, and then tries to access a session_attribute row already locked by transaction A. We again have a deadlock, at the row level.

0.12.2dev: in session handling, try to avoid table deadlocks (#9705)
- in `Session.save`, split `save_session` transaction in two (actually save the session, then eventually purge the expired sessions)
- always access tables in the same order, first `session`, then `session_attribute`

0.12.2dev: for rst rendering with docutils, don't store Trac env and
context in the directive and role handlers.

This is not thread-safe. Instead, we register the "trac" and
"code-block" directives and roles only once and the handlers retrieve
the Trac environment and rendering context from the input we feed to
the docutils API.

As a bonus, Trac plugins can now use the same mechanism and register
their own directives and role handlers. For retrieving contextual
information, they can do the same way as we do here, look up the
Inliner's `trac` attribute. The attribute will be an `(env, context)`
pair and will be set if and only if Trac rendered the rst content. The
Inliner is available directly as the `inliner` argument passed to role
handlers and from the `.inliner` attribute of the `state` argument
passed to the directive handler.

0.12.2dev: when producing a log link, the `get_repository_by_path` might trigger a !TracError. Catch this early, like we did in r10456 for changeset links.

0.12.2dev: when producing a changeset link, the `get_repository_by_path` might trigger a !TracError. Catch that to produce a 'missing' link instead of letting the error through.

SpamFilter: hopefully finally fix blogspam exclude