This is the mail archive of the mailing list for the Cygwin project.

Index Nav: [Date Index] [Subject Index] [Author Index] [Thread Index]
Message Nav: [Date Prev] [Date Next] [Thread Prev] [Thread Next]
Other format: [Raw text]

Re: cygwin.bat

On Sun, May 11, 2003 at 04:15:05AM +0100, Sam Edge wrote:
>cgf wrote in <>
>in gmane.os.cygwin on Sat, 10 May 2003 22:18:35 -0400:
>> >> >If you use cygwin.bat, then under Windows NT/2k/XP you first have a
>> >> >CMD.EXE process created and then a bash.exe. The CMD.EXE sits around
>> >> >doing nothing until the bash.exe process exits.
>> >> > [snip]
>> >Negligible on a modern PC. Noticeable on some of the kit I've got
>> >here. ;-)
>> I *really* find that hard to believe.  I've run cygwin on some pretty
>> slow computers and I've never noticed a slowdown from running bash
>> in a command shell.  Even Windows should be smart enough to know that
>> it can swap out most of when bash.exe is running so I
>> don't know how this could be a resource thing.
>Oh yes. I've already said that the general system slow down due to one
>extra waiting process isn't going to be noticeable. (Although just
>because a process is swapped out doesn't mean it ceases consuming
>resources, especially on Windows!)

Of course.  It's consuming swap space on disk for one thing.  It's
unlikely that in-use handles or atoms or shared memory is going to have
much impact on active processes however.

>But the overhead in /launching/ an extra CMD.EXE to process the ".BAT"
>file that then starts up bash.exe can be noticed on old kit under NT.

So, once again, as I said:

Translation: Except for a neglible startup cost, it's probably his


Unsubscribe info:
Problem reports:

Index Nav: [Date Index] [Subject Index] [Author Index] [Thread Index]
Message Nav: [Date Prev] [Date Next] [Thread Prev] [Thread Next]