| Profil von Uni EraThe Unified EraBlogListen |  Extras  Hilfe |
About ecmax.xobjA sub project, 'xobj', as part of project ecmax, is in development right now. Its goal
is to implement a simple
set of XmlCursor api
like Apache XmlBeans. The ecmax.xobj project is based on the Streaming API for Xml(StAX) and its open source RI from BEA. Switching to DOM from xmlBeansecmax's e4x impl is
replaced with a dom-based impl delivered by David
Caldwell. Now, the speed is at least 50% faster than before, the
size of release package is smaller than half of the old one which
contains even a simplified xmlBeans impl. Change log of ecmax: ----------------------------------------------- 1, 'like' expr. 2, big decimal instead of double. 3, defered xml and joint xml 4, xml + number redefined. 5, xml.idx_from_1 expr for xml. 6, xml.ns:name expr for xml. -7, change summary for xml. New Changes
The pervious
proposal of ecmax.change was disposed, and the idea about sequence
change(whole updated once its length changed) is also removed. To prepare to suggest the new proposal, one new promotion for e4x has been done in ecmax to support elements-accessing via 'xml.index_form_1' style. It's remarkable that the 'a.1' expression in ecmax is not equal to the same expression but the 'a[1]' expression in xpath or sdo. It's notable also that the new promotion has no side-effect upon the specifications about DecimalLiteral at chapter 7.8.3 of ecma-262. Finally, for example: x = <a><aa>bbb</aa></a>; print(x.1.toXMLString()); // x.1 is equivalent to x[0] print(x[0].toXMLString()); print(1.+.2); // equivalent to 1.0+0.2 ------------------------------------------------------------------ System.out - <a> <aa>bbb</aa> </a> System.out - <a> <aa>bbb</aa> </a> System.out - 1.2 Sophisticated Change Tracingx = <a
id="1" value="a" lock='true'> text in a. <b id='b1' value='b'>text in b</b> string in a. <c id='b1'>text in c</c> </a>; x = new SDO(x); x.@id = 2; delete x.@value; x.@name='A'; x.b.* = 'new text in b.'; x.b.c.@a = 'ca'; // cause element <c> created. x.b.@name='b'; x.b.@id='b2'; x.*[3] = 'new string in a.'; // cause length of x.* changed. x.*[3] = <bd>aaaa</bd>; x.*[3] = '<dd>eeeeeeee</dd>'; //delete x.*[3]; print(x); ------------------------------------------------------------------ logUpdate(N'.@id', 2, L'1') called. logDelete(N'.@value', L'a') called. logCreate(N'.@name', 'A') called. logUpdate(N'.b.*', 'new text in b.', L'text in b') called. logCreate(N'.b.c', L'') called. logCreate(N'.b.c.@a', 'ca') called. logCreate(N'.b.@name', 'b') called. logUpdate(N'.b.@id', 'b2', L'b1') called. logUpdate(N'.*', L{3}, L{4}) called. logCreate(N'.*.3', X'aaaa') called. logUpdate(N'.*.3', '<dd>eeeeeeee</dd>', X'aaaa') called. System.out - <a id="2" lock="true" name="A"> text in a. <b id="b2" value="b" name="b"> new text in b. <c a="ca"/> </b> string in a.new string in a. <dd>eeeeeeee</dd> </a> Note: in ecmax.sdo, if a sequence's length has been changed by some operation(but not append() operation, updated at 2007-4-16), it must be marked whole updated, instead of marking the changed element(s) within this sequence. Differences between .[] and .()Q: A new filter operator .[] has been introduced into E4X in ecmax to coexist with the old filter .(), but what's the difference between them? A: Generally in the two aspects: 1, when applied to common xml object, .[] behaves as the same as .(), and the extended syntax in .[] does no effect. 2, when applied to deferred xml object, .[] returns the same deferred xml object with internal changes, but .() may return a newer xml object or changes it 'instant' from defered. |
|
|
