[ImageJ-bugs] [Bug 1036] version skew in scifio-ome-xml-0.12.1-SNAPSHOT.jar ?
bugzilla at fiji.sc
bugzilla at fiji.sc
Thu Mar 26 09:14:15 CDT 2015
http://fiji.sc/bugzilla/show_bug.cgi?id=1036
Mark Hiner <hinerm at gmail.com> changed:
What |Removed |Added
----------------------------------------------------------------------------
Status|NEW |RESOLVED
CC| |hinerm at gmail.com
Resolution|--- |WONTFIX
--- Comment #1 from Mark Hiner <hinerm at gmail.com> ---
Hi Aryeh,
The Bio-Formats update site enables Bio-Formats 5.1 and thus has to ship
updated components that use the new API (such as a modified scifio-ome-xml).
If any other update site ships its own version of Bio-Formats but doesn't also
override the components that have been updated to 5.1, there will be dependency
skew.
If any other update site ships its own version of the updated components
without also reverting Bio-Formats to 5.0.x, there will be dependency skew.
The NoClassDefFound is specifying a class (ome.units.quantity.Length) added in
Bio-Formats 5.1. Thus you must have another update site enabled that is
overwriting Bio-Formats 5.1 with 5.0 without similarly reverting
scifio-ome-xml. Turning on the SCIFIO-dev update site fixes the problem by
accident, by overwriting the scifio-ome-xml.jar with one compatible with
Bio-Formats 5.0.
Knowing this, I looked for what update site was providing the Bio-Formats
ome-xml.jar[1]. SCIFIO-dev actually provides its own copy of Bio-Formats 5.0 to
avoid skew like this. However, it looks like TrackMate-dev provides an
ome-xml.jar without also updating scifio-ome-xml.
So TrackMate-dev is incompatible with Bio-Formats update site. The resolution
is to disable one of the two until TrackMate-dev is updated to use Bio-Formats
5.1.
I also filed an issue for the updater to warn about conflicts[2] when update
sites provide the same artifacts.
[1]
http://imagej.net/Bug_reporting_best_practices#Disable_problematic_update_sites
[2] https://github.com/imagej/imagej-ui-swing/issues/36
--
You are receiving this mail because:
You are the assignee for the bug.
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://imagej.net/pipermail/imagej-bugs/attachments/20150326/1acaf95f/attachment-0001.html>
More information about the Imagej-bugs
mailing list