It's more complicated than I thought. I tried it in FB 4.6 (i.e., with the old compiler). If you have a one space inside the <String> tag without a CDATA section around it, then the value is null. But with CDATA around the space, the value is a string with one space character.
In my opinion, this behavior of the old compiler is inconsistent and should be considered a bug. So we have to decide how bug-compatible we want Falcon to be with the old compiler. It may be difficult to even achieve bug-compatibility in cases like this. Could we live with Falcon compiling it as a string with one space, regardless of whether there is CDATA or not? - Gordon -----Original Message----- From: Gordon Smith Sent: Monday, October 29, 2012 10:17 AM To: flex-dev@incubator.apache.org Subject: RE: Does Falcon need to take CDATA sections into account when parsing <fx:string/>? This is a bug. The tests in MXMLBooleanNode are more complete that for other primitive tags. I added a test to MXMLBooleanNode that shows CDATA works inside of <Boolean>. Each of the primitive tags needs this kind of test. - Gordon -----Original Message----- From: Roland Zwaga [mailto:rol...@stackandheap.com] Sent: Sunday, October 28, 2012 10:12 AM To: flex-dev@incubator.apache.org Subject: Does Falcon need to take CDATA sections into account when parsing <fx:string/>? Hi there, I've been taking my few first uncertain steps into grokking the Falcon sources. I was looking at the MXMLStringNode tests and wanted to see if the parser takes into account CDATA sections in them. So I added a test that set up an <fx:string/> node like this: String code = "<fx:String><![CDATA[ ]]></fx:String>"; //The CDATA section here contains a single space IMXMLStringNode node = getMXMLStringNode(code); assertThat("getValue", node.getValue(), is(" ")); Now, this test fails, because the node.getValue() returns null. IMHO, this is incorrect behaviour, since CDATA sections should be parsed literally without any assumptions or postprocessing. Does everyone agree on this? Thoughts please ;) -- regards, Roland -- Roland Zwaga Senior Consultant | Stack & Heap BVBA +32 (0)486 16 12 62 | rol...@stackandheap.com | +http://www.stackandheap.com http://zwaga.blogspot.com http://www.springactionscript.org http://www.as3commons.org