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: Curious behavior of CYGSERVER

Conrad Scott wrote:

"David A. Cobb" <> wrote:

I discussed above (

my problems attempting to do a bootstrap of gcc-3.2. One experiment
involved firing up CYGSERVER. As mentioned in the previous thread, the program went much much much further this way -- make check took about 13 hours!

I'm hoping that this is because `make check' did more work rather than
cygserver somehow slowing down the process . . .

In fact, the process goes remarkably *faster* with cygserver running. The make-check ran to completion (C, C++, Java, Obj-C). Ada didn't get built for other reasons.

In the process, I notice two unexpected behaviors. The CYGSERVER emits

dots (.) on the screen every little while - perhaps it emits one
everytime it gets called. This isn't bad -- it makes a handy pulse to
be sure the machine hasn't just frozen up. Three screenshots are
attached fro the make check run.

It's just being warm and cuddly, like Robert Collins mentioned: I've
made it less friendly, as Nicholas Wourms mentioned :-)

What is really bad is that somehow, having CYGSERVER involved defeats
the stdout redirection. For example, the make check not only ran for


hours, but it also gave me very little clue as to its success or


This I don't understand and cannot recreate any such problem.  That is,
with cygserver running, redirection of both stdout and stderr work fine
in my little tests.  Could you give more detail on exactly how stdout
redirection is "defeated"?

I have my configure, make-bootstrap, make-check bundled in a single "" script. If I do either:

$ source ./ >build.log 2>&1; or
$ ./ >build.log 2>&1

It goes merrily along but, when all is said and done, "build.log" don't exist. Not even somehow misplaced - I searched my whole darn machine.

I also did a $ find ..... >found-files.txt and had the same result (no output file)

BTW, I'm running on win2k just now: which platform are you using?
Perhaps the output of `cygcheck -s -v -r' would help matters here.

Win2000pro, everything up to latest release (no development stuff until I can get some good builds).

Also, as Nicholas mentioned, if you're using any version of cygserver
other than the development one on the cygwin_daemon branch, you'll need
to clean up its droppings after it's done otherwise everything runs very
slowly: that is, remove the /tmp/cygdaemo socket file if you're not
running cygserver anymore.

Thanks, I'll do that in future. I hadn't noticed. WAIT A MO', I cleaned up /tmp yesterday before one run and the Windows exception came back even with Cygserver active. Check my other thread about that.


// Conrad

David A. Cobb, Software Engineer, Public Access Advocate
"By God's Grace I am a Christian man, by my actions a great sinner." -- The Way of a Pilgrim; R. M. French, tr.
Life is too short to tolerate crappy software.

Attachment: cygcheck-2002-09-15.ZIP
Description: Zip compressed data

Unsubscribe info:
Bug reporting:

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