On 2009-12-26, at 15:42, Rami Ojares / AMG wrote: > Yes, that was it! > The value may not be "" (empty string) when the type of attribute is > expression. > Have to keep this one in mind. > > A few follow up questions: > > What is the meaning of > compiler.swf9.execflex > in lps.properties.
Setting it one way or another causes the flex compiler jar to be invoked directly in the current process or to be invoked as a separate process. There was some concern early on that the flex compiler had a memory leak, by invoking it as a separate process, that leak would not contaminate the OL server. It is no longer necessary. > When I set it to true it seems to be invoking at least mxmlc in WEB-INF/bin > (and maybe some other programs also?) > But it could not compile one font saying that: > > Error: unable to build font 'NotesFont', in line: > [Embed(mimeType='application/x-font-truetype', > source='/opt/prg/tomcat/apache-tomcat-6.0.18/webapps/exodus/forms/resources/edosz.ttf', > fontName='NotesFont', > fontWeight='plain',fontStyle='plain',advancedAntiAliasing='true')] > index.lzx: 51: : Error: Unable to transcode > /opt/prg/tomcat/apache-tomcat-6.0.18/webapps/exodus/forms/resources/edosz.ttf, > in line: . The font may be a bold or italic style, try adding style="bold" > or "italic" attribute to font tag > > It doesn't help if I set font style to bold or italic (and the font should be > regular, although I don't know exactly what that means). > Anyway when compiler.swf9.execflex=false it gives me no problems. > > Is compiler.swf9.incremental effective in swf10? > And what about other properties with swf9 in their name? > The reason I ask is that the incremental compilation does not seem faster > than compilation without it. > I think it was faster when I used swf9 but I might be mistaken. Also it used > to crash Tomcat every third reload ... swf9 is a misnomer. We should have called it as3 or flex or something, had we known that swf9 and swf10 would be so closely related. I believe the only difference in swf9 and swf10 in OL is that it sets the target player when calling the flex, which enables the additional swf10 API's. Within the OL compiler, they are treated as synonyms. I'm believe incremental compilation is presently broken because the OL compiler writes all the intermediate files every time. The dependency tracking seems to have broken in the compiler. > - rami > > On 26.12.2009 21:05, P T Withington wrote: >> Google reveals this possible hint: >> >> http://www.mail-archive.com/[email protected]/msg04989.html >> >> Sorry there is no line number! Definitely a bug. >> >> On 2009-12-26, at 12:59, Rami Ojares / AMG wrote: >> >> >>> Hi, >>> >>> I encountered the following error after doing a lot of changes to my source >>> code. >>> >>> Error: Syntax error: the attribute value ended in mid-expression. >>> Look for an unclosed '(', '{', or '['. >>> >>> >>> Now I can't seem to find where that error is coming from. >>> Any ideas how to locate the source of error (since it does not give me any >>> clue about what file it originates from). >>> >>> - rami >>> >> >
