On Monday 09 June 2008 05:17:19 James Keenan via RT wrote: > (We use File::Temp extensively in the testing, but I can't recall a > situation where I absolutely had to 'chmod' a tempfile. This may just > be an artifact of the previous way of doing things.)
I agree; it seems unlikely that File::Temp can create the file that it can't read and write. Setting global read and write permissions is very curious -- expecting that temp file to persist long enough for some other process to drop permissions and try to read it is just asking for race conditions and portability problems. +1 to remove the chmod altogether. -- c