Fix commit with class renames for Mercurial 4.8 and later Starting with Mercurial 4.8 one has to use `-f` option to force rename of file over a file that is already tracked, otherwuse `hg mv` fails. This is tested by `#test_commit_17h`. This commit fixes this issue.

Issue #239: Fix all Smalltak/X source files to be in unicode (UTF8 without BOM) and prefixed by "{ Encoding: utf8 }" when any unicode character is present - All source *.st files are now Unicode UTF8 without BOM Files are in two groups (fileOut works this way in Smalltalk/X): - containing a unicode character have "{ Encoding: utf8 }" at the header - ASCII only are without the header

Fix `HGRevisonInfo >> #date` to return correct date Oops, due to a wrong format specification, the returned string was completely bogus! This commit fixes this and makes sure the date can be then parsed to an instance of a `Date` using `Date class >> readFrom:`

Added directory `objmingw` to default `.hgignore` ...generated upon first commit (or when there's no `.hgignore`)

Partial fix for "null" `HGChangeset` "null" changeset is currently kept as a singleton instance, i.e., `HGChangeset null` always return the same object. However, some queries (such as `#isObsolete` or `#successors`) require repository to be set - obviously, for "null" changeset there's no repository. Therefore these queries resulted into DNU. To paetially fix this, short-circuit these query methods to answer trasonable values for "null" changeset. Better solution would be to make "null" changeset a normal changeset within a repository.

Merge

Fix parsing og `hg merge` output for Mercurial 4.6 and newer In Mercurial 4.6 the wording of last line of the output has changed. This commit updates the parser so it can handle both wordings.

Fix bad error messages in `HGChangesetId` parser The error was signalled correctly but its message was wrong / misleading (copy-paste error)

update tags

#REFACTORING by cg class: HGRevisionAnnotation removed: #annotatesClass: #annotatesMethod: