Excessive logging of data run length mismatches
| Project: | JNode FS |
| Component: | Code |
| Category: | bug report |
| Priority: | normal |
| Assigned: | Unassigned |
| Status: | closed |
Jump to:
Description
This one's my fault, but it seems that if there is an $ATTRIBUTE_LIST attribute we log INFO about it once in FileRecord#getAttribute, and then we get (related, but not directly related) ERROR messages about each individual chunk multiple times in NTFSNonResidentAttribute#readDataRuns.
Unfortunately #readDataRuns is called from the constructor so it's hard to set some kind of flag on the object to tell it not to log these errors, so we'll have to find some way to fix it properly and perhaps log the error about there not being enough data, at some other point in the object's lifecycle (e.g. in #readVCN.)
- Login to post comments
#1
Daniel, is that problem still existing? Can you give an update on that bug report?
#2
It does still appear to be there, from reading the code at least.
It still looks like making readVCN do this check is the way to go. It's just hard to find time to fix it since it's only a cosmetic issue and split attribute lists are (in our own data, anyway) relatively rare.
#3
Closing this as my work on related Issue 2817 will rid the code of these warnings. I'll check the log level of anything in the files I work on and ensure that everything is appropriate.