This is the mail archive of the cygwin 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: freopen irritates system-call/sh.exe

On 30 March 2006 16:10, Corinna Vinschen wrote:

> On Mar 30 14:56, Eric Blake wrote:
>> Jens Wilken <Jens.Wilken <at>> writes:
>>> How does freopen-ing stdout and stderr affect subsequent "system" calls?
>> It shouldn't have any effect on subsequent calls, other than that stderr
>> or stdout is mapped to a new underlying stream.  There are other changes,
>> unrelated to freopen, that do affect consoles popping up or not (and by
>> the way, snapshot 20060329 flashes a console when cygrunsrv starts an
> This has nothing to do with the snapshot.  It's the latest change to
> cygrunsrv which re-introduced the short flash.

  Perhaps so, but perhaps not - see my last post.  I'm fairly sure cygrunsrv
should not involved in invoking expect.exe when I write "make -k check" in a
bash shell in a newly-built gcc object dir!  I also haven't changed my
cygrunsrv; I can make the condition appear or disappear by using today's dll
or the one I built from the 17th.

> Either a console window flashes for a split-second, or you will never
> see a console window, but the console window is then unusable for output
> from the service application.  There's no way around this for now.
> To not lose the output capability, you'll have to live with the flashing
> window.

  Well, there is no service application in this case, and these ones pop up
for as long as the spawned expect.exe process lives.

Can't think of a witty .sigline today....

Unsubscribe info:
Problem reports:

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