Inveon: fix offsets for files larger than 2GB #1502
Conversation
| @@ -110,7 +110,7 @@ public boolean isThisType(RandomAccessInputStream stream) throws IOException { | |||
| { | |||
| FormatTools.checkPlaneParameters(this, no, buf.length, x, y, w, h); | |||
|
|
|||
| int planeSize = FormatTools.getPlaneSize(this); | |||
| long planeSize = (long) FormatTools.getPlaneSize(this); | |||
There was a problem hiding this comment.
Given that FormatTools.getPlaneSize() is returning an int and sqrt(2^31) => 46340 we're likely to overflow for big histopathology images in general. Should getPlaneSize (and the other helpers which use it) all be switched to use long, saving the need for the cast here? This will ensure we're not routinely overflowing for all readers using this and related helpers.
There was a problem hiding this comment.
That's up for debate I guess, but outside the scope of this PR. Either way the return value of getPlaneSize(this) would need to be cast - if it returns long we're just casting when a byte array is constructed instead.
There was a problem hiding this comment.
Yes, it's certainly outside the scope of this PR, just thought it worth mentioning if it's a more general limitation we might want to address as a follow-on task.
Inveon: fix offsets for files larger than 2GB
|
--no-rebase |
See
http://lists.openmicroscopy.org.uk/pipermail/ome-users/2014-December/004972.html
Builds should remain green with this change; the only thing to verify is that the offset passed to
dat.seekis calculated using long (not int) arithmetic.