Fixed a bug in changeset id parsing The changeset id parsing logic can parse both, revision numbers and changeset id (hashes). It may happen, however, that a short id (12 chars) may consist of digits only (like "087537207973"). In that case it considered it being a revno which lead to an error. This is hard to disambiguate without querying the repo. However, it seems more appropriate to treat it as id when the maybe-revno is 12 digits and more in length. If it would be a revno, then the repo would contain at least 10^11 revisions.

Added support to add (move) bookmark when commiting a new changeset ...right from a commit dialog. This is usefull when one want to use feature-branch based approach when developing/fixing code managed in Mercurial.

Fixed UHE in browsers when Mercurial working is detected but not Mercurial is installed/configured The problem is that when a HGRevisionInfo is asked for a user or date, it tries to retrieve a changeset which, if Mercurial is not instaled, would result in an unhandled exception. This in turn makes browser unusable since it's asking for revision info virtually on each click. To solve this problem, when a there's an error retrieving changeset information, simply return a "fake" changeset with ID set but leave the rest of data empty.

Issue 79: Settings => mercurial+ => MessageNotUnderstood (part 4) Fixed DNU when clicking to 'Browser' button and no Mercurial is installed. Naturally, when there's no Mercurial, `HGCommand hgCommand` returns nil which does not undestand `asFilenane`. Care for this by checking for `nil` return value of `HGCommand hgCommand`.

update tags

#UI_ENHANCEMENT by cg class: HGSourceCodeManagementSettingsAppl changed: #windowSpec geometry

update tags

#DOCUMENTATION by cg class: SCMCommonPackageModelGroup comment/format in: #species

#REFACTORING by stefan class: stx_libscm_common removed: #initialize send super

*** empty log message ***