I've thought about this bug for a while, mainly pondering how other editors might overcome problem with incorrect time (NTP) settings on remote filesystem (the "document saved in future" problem). Simply quessing (I do not do desktop applications, so most probably I am more or less wrong due my zero knowledge of this topic):
- they calculate some offest: stores local timespamp of save action and immediately looks on remote timestamp, and since then takes its difference into account. Seems unlikely, because the save operation itself might take some time (?) - they fetch and compare full content on each check. Straightforward, but surely not suitable for large files. (but ... really?) One heretic question comes to my mind as well: - are we 100% sure that old versions of PSPad which did not make so many false positive alerts about changed content didn't potentially miss some real changes? It sounds scary, but it would be good to know. One simple workaround for our problem comes to my mind: in case of mismatching timestamps before raising alert do one extra full-content check. This would prevent all current false positives and even those when file have been edited by other application but the content of given file remained the same as last version seen by PSPad. It surely is not elegant, but it would heal our current ache. -- <http://forum.pspad.com/read.php?2,53148,58140> PSPad freeware editor http://www.pspad.com
