[kepler-dev] cache file name

Kevin Ruland kruland at ku.edu
Tue Nov 15 13:54:45 PST 2005

I think we've just pushed the problem off to making certain the
generated lsids are unique.

Chad Berkley wrote:

>ahh, nice fast computers break our software eh?  :)  In the new object 
>cache, the lsid is used as the serial file name so this shouldn't be a 
>Jing Tao wrote:
>>Hi devs:
>>Just now Dan and I found a bug for cache file name. Currently after 
>>download a doc, the cache manager will assign it a name. The name come 
>>from download time (by milli seconds) with file extension ".dat". It 
>>worked fine for a while.
>>But dan got a fast machine, then issues came out: it seems when he tried 
>>to open a workflow (there are 10 data resource actors there), two 
>>documents can be download at exactly same time and manager assigned same 
>>cache file name to them. One will overwrite another one. Two different 
>>data source will point to same chache file.
>>I think we should need another mechanism to assign cache file name.
>>Jing Tao
>>National Center for Ecological
>>Analysis and Synthesis (NCEAS)
>>735 State St. Suite 204
>>Santa Barbara, CA 93101
>>Kepler-dev mailing list
>>Kepler-dev at ecoinformatics.org
>Kepler-dev mailing list
>Kepler-dev at ecoinformatics.org

More information about the Kepler-dev mailing list