This Bugzilla instance is a read-only archive of historic NetBeans bug reports. To report a bug in NetBeans please follow the project's instructions for reporting issues.
Probably the culprit is "-clean-jnlp-userdir" target. 1) Create suite called e.g. suite1 2) Suite Properties -> Basic Branding -> check the branding checkbox 3) mkdir ~/.suite1 4) Shift-F11 ~/.suite1 is moved to ~/.suite1.bak Please don't try to call your suite vim, kde, gnome, gnome2, java, fluxbox, mplayer, emacs, xine, xmms, gconf.... The result (probably better word is _feedback_) could be *very* surprising ;)
Do you have any better idea how to solve issue #64249?
Could perhaps change default name of userdir for both JNLP and regular app to e.g. "~/.nb-app-$suitename"?
> Could perhaps change default name of userdir for both JNLP and regular > app to e.g. "~/.nb-app-$suitename"? Sounds good and also solves my original problem: I have an suite e.g. called "jvim". I run the application, exit, S-F11, run again, .jvim is "backuped". So that should definitelly solve the problem also for non-OS-conflicting names.
Rare (5.0-beta-waiver) to be dangerous (P1)
committed Up-To-Date 1.19 apisupport/harness/release/jnlp.xml
verified
> Could perhaps change default name of userdir for both JNLP and regular > app to e.g. "~/.nb-app-$suitename"? It seems that only jnlp part was fixed. Please fix or evaluate also zipped distribution case. (lowering priority --> imediate deleting has gone)
Probably just about fixing app.conf? RCS file: /cvs/apisupport/harness/release/etc/app.conf,v . . @@ -1,5 +1,5 @@ # ${HOME} will be replaced by JVM user.home system property -default_userdir="${HOME}/.${APPNAME}/dev" +default_userdir="${HOME}/.nbapp-${APPNAME}/dev"
Intentionally. 1. Zipped apps are not run directly by any Ant targets. 2. The default userdir already includes the /dev suffix so it is unlikely to interfere even with ~/.gnome or whatever. 3. Unlike in the case of JNLP, the clean target does not remove the ZIP app userdir - that, I think, being the original complaint in this bug.
Yup, clear. Wanted to clarify dev@openide thread. Thanks. v/c