If i may. On Fri, Oct 19, 2007 at 09:31:04PM +0200, Oleg Verych wrote: [] > > >This change is not going to be major, because now it is possible to > > >multiply calls or using -a' || -o'. But going to be very good > > >optimization for such kinds of applications. > > _Optimization_, i.e. to do not call stat() on the same file on and on. > This is just silly.
This is called "userspace sucks" by kernel people, isn't it? Linux have so many changes every release, yet userspace still goes with all that UNIX zoo legacy burden. IMHO it's time to start think and implement new userspace around very good kernel ever. Even if POSIX doesn't require that, it doesn't prohibit that also. >From what i've seen so far, is very painful for me. Example. Early userspace people were living without pipefs for years. The pipe() is one of the most basic and important things in the shell. Yet perly debian developers, for instance, do not understand why shell execution environment on the pipe's end doesn't go to the current one: #!/bin/sh cat /proc/cmdline | read OP1 OP2 OP3 REST echo 'OP* do not go here, AAA! shell suxx!' # shend Not understanding basic system calls, like pipe(), so what system programming or OS development we are talking about? I'm sure, that those guys together with python crowd do not even understand, that script itself can be read from the possibly infinite pipe. How about that stuff, DDs? Broken userspace<->kernel link, no interest in both sides of system programming lead to such things with all that dirty hacks over and over again. Maybe i'm just a newbie ignorant, but my mind tells me, that examples are real, and this is not something, one can easily beat. ____ - To unsubscribe from this list: send the line "unsubscribe dash" in the body of a message to [EMAIL PROTECTED] More majordomo info at http://vger.kernel.org/majordomo-info.html
