From kde-java Mon Apr 12 20:21:15 2004 From: Richard Dale Date: Mon, 12 Apr 2004 20:21:15 +0000 To: kde-java Subject: Re: [Kde-java] [PATCH] Make juic work again Message-Id: <200404122121.15132.Richard_Dale () tipitina ! demon ! co ! uk> X-MARC-Message: https://marc.info/?l=kde-java&m=108180144516706 On Monday 12 April 2004 13:15, Marco Ladermann wrote: > Am Sonntag, 11. April 2004 12:43 schrieb Dominique Devriese: > > Marco Ladermann writes: > > I'm not sure what's causing this, but I'm still getting uncompilable > > java files. The Layout variables are still not declared as instance > > members. > > This are my findings: There is a mode "declareInstances" where those nodes > are selected that have a property with a "name" except the parent is a > QLayoutWidget. For this I used the expression "../@class != > 'QLayoutWidget'", but I forgot (and my version of xsltproc didn't remember > me) that in XPath this is *not* equivalent to "not(../@class = > 'QLayoutWidget')". While the former is false for a parent node > without "class" attribute, the later evaluates to true! I changed that and > now it seems to produce the same code for saxon, xalan (which I also > upgraded to version 2.6.0) and xsltproc. Your example transformes, compiles > and runs with my jdk 1.4.2 and any of the above processors (juic -proc > (xalan|saxon) ...). This is good news that it works with all xslt things - the code generated looks very nice. RE: you KLineEdit with QApplication instead of KApplication top level. I don't think it is a good idea to trap signals - better to autogenerate code via juic or a KDevelop project template correctly! I think the jvm trap seg faults and turn them into java null pointer exceptions, so it would be difficult to distiguish between C++ seg faults (ie KLineEdit error), and java ones. -- Richard _______________________________________________ Kde-java mailing list Kde-java@kde.org https://mail.kde.org/mailman/listinfo/kde-java