From 4.6.4 to 4.6.5 ==================== (06.11.2006) Improved handling of wildcards. In particular, changed the grammar translation such that no rule is added for the non-terminal for arbitrary trees, ("."). Instead transitions always consider that every node can be derived from ".". Improved the online algorithm for arbitrary queries such that traversal of subtrees is accelerated as soon as they can not contain matches. From 4.6.3 to 4.6.4 ==================== (23.05.2006) Removed bug in the online algorithm for arbitrary queries From 4.6.2 to 4.6.3 ==================== (20.07.2005) Added web-interface to fxgrep From 4.6.1 to 4.6.2 ==================== (05.07.2005) Removed a bug in the translation of patterns into grammars, occuring when the secondary node is was a text node which is last in a pattern. From 4.6 to 4.6.1 ==================== (17.01.2005) Removed a bug from the one-pass algortihm for non right-ignoring queries, occuring when the right content of some ancestor had to be visited. Optimized the code for the online matching From 4.4.2 to 4.6 ==================== (27.12.2004) Improved the algorithm for one-pass matching along the parsing such that matches are detected (up to conjunction and negations of content models) as soon as possible From 4.4.1 to 4.4.2 ==================== (27.10.2004) Added regular expressions for attribute names From 4.4 to 4.4.1 ==================== (20.10.2004) Regular expressions for element names match any substring rather than the whole name From 4.3 to 4.4 ==================== (15.10.2004) Removed a bug from grammar analysis in preAlg.sml From 4.2 to 4.3 ==================== (16.06.2004) Improved online matching for arbitrary queries with conjunctions (18.03.2004) Online matching for arbitrary unary queries From 4.1 to 4.2 ==================== (09.10.2003) Distribution includes now a graphical user interface From 4.0 to 4.1 ==================== (05.06.2003) Rather than sets of variables matches (both primaries and secondaries) can be specified via boolean formulas over variables. Reduced code size for matching. Changes in the interface to fxt. Updated documentation. From 2.1 to 4.0 ==================== (17.02.2003) Allowing negations and conjunctions of paths in patterns. A pattern can be now a conjunction of possibly negated paths followed by the target node. Changes in the interface to fxt. Updated documentation. From 2.0 to 2.1 ==================== (18.07.2002) Changed the default leading and trailing forest in forest patterns (a[b] <--> a[_ b _]) Changed interface to fxt to allow for binary patterns in fxt Changed the reference to a secondary from & to % (15.04.2002) Added regular expressions for element names From 1.4.5 to 2.0 ==================== (11.04.2002) Added binary patterns (20.02.2002) Modified the Makefiles to avoid unnecessary recompilations (21.01.2002) Removed a bug in down transitions in the second pass Improved side transitions in the second pass From 1.4.4 to 1.4.5 ==================== (17.09.2001) Reduced code size Updated documentation Removed a bug in parsing escape sequences in patterns Removed a bug in pattern matching of procesing instructions From 1.4.3 to 1.4.4 ==================== (13.09.2001) Removed unrelevant debug messages From 1.4.2 to 1.4.3 ==================== (28.08.2001) Internal changes to give to fxt information needed for one-pass transforming From 1.4.1 to 1.4.2 ==================== (6.08.2001) Extension of the pattern language of fxgrep: - regular conditions on the paths may now be specified as for example: (a/)+b - matches the b-s that have only a ancestors - regular contextual conditions are now allowed as for example: a[(b#)+_]/b - matches every second b of a document with type a if this contains only b sons Internal optimization for the one-pass matching (both for the while- and after-parsing matchings): - no-hooks: if a grammar is content ignorant (i.e. there is no need of inspecting the content of an element in order to tell whether it is a match or not) and all matches are to be reported then matches are reported directly as they are detected after seeing their tag name and their attributes. - hooks: the extra administrative task of maintaining an optional partial document tree, for the possible matches is reduced, as the decision whether to construct the subtree is delayed until after the element name and the attributes are processed. From 1.4 to 1.4.1 ==================== Internal changes that permit the interfacing of fxgrep with fxt.