Changeset 337 - gearman - TracTimestamp: 09/16/07 09:49:26 (1 year ago) Author: hachi Message: This fixes the problem exposed in the test in the previous commit. Large jobs are somehow being split into multiple reads. I'm not sure exactly why this is since SSIZE_MAX is 2GB on the machine I'm testing on, but it seems that reads max out at 32kilobytes and so we have to run them multiple times to get all the data. Files:
{{#tags}}- {{label}}
{{/tags}}