Re: cvs: php4 /ext/standard file.c

SS>> That is the problem I outlined earlier. There is no way to SS>> distinguish those two cases, AFAICT. Well, seems that standard guys shoot themselves in the foot once again. Not missing developer's collective foot in the way. Nice. So, there should be one solution or another. Like, why not to block that signal altogether? E.g., in activate() or somewhere around that? I see no other way of handling it gracefully - or we allow write() to be partial (thus moving burden of writing full_writes to application developers) or we ignore that signal. SS>> Do you see any rationale behind terminating Apache when a SS>> simple PHP script accesses a file which exceeds process SS>> limits? I think it makes more sense to fail gracefully and I was partially sarcastic here. I don't see any rationale in this behaviour at all, but supposedly Unix standard guys thought something before putting out such an insanity? They don't give us a way to know what caused partial write, and don't give us a way to reliably restart write without risk of process being killed. So, what are we going to do, having to deal with this insanity? -- Stanislav Malyshev, Zend Products Engineer stas@zend.com http://www.zend.com/ +972-3-6139665 ext.115