but the existing file can be removed by another thread in less time than 1 second. I understand why I have got downvoted, but the point is not on the method of creating the unique filename it is on how to avoid overwrinting the existing file.
--
tune | [reply] |
If someone else is creating and deleting files with your process ID and your timestamp, you've got more problems than can be solved within the scope of this discussion. For one, you must now do an atomic update, not just test in one operation and set in another.
Please specify your conditions when you provide a solution. I'm presuming standard Unix, where I can be reasonably assured that time().$$ is unique for a given process. If you've got some weirdo situation, please give it. I'm pretty sure you were downvoted (at least by me :) because you didn't specify your alternative context.
-- Randal L. Schwartz, Perl hacker
| [reply] |