Paulo Levi wrote:
Hi a bug, i've been seeing:
First a simplification:
Two different RandomAccessFile objects, inited in "r" mode and
pointing to the same file appear to be interfering on on their reads
(concurrent threads). They both manipulate the file pointer i guess,
but i expected nothing bad to happen, because they are in read mode.
The real story:
The library i'm using is more than a "RandomAccessFile", so i'm not
actually sure the error occurs here or in some static state somewhere.
I'd like to fix this so its transparent to the clients of the api i'm
building (me!) so i'd like to find the bug, or if this is a known
issue i guess i can make a factory + map of live RandomAccessFiles.
Asking if the situation described above is a known bug of RandomAccessFile?
This does not sound familiar. Are you 100% sure you don't have
concurrent threads seeking and reading with the same RandomAccessFile
instance? Do you have a test case that demonstrates the issue?
-Alan.
