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: cygwin version 1.5.16-1 (exim 4.50-1, minires 1.00-1, perl 5.8.6-4) attempting to run spamassassin (spamc and spamd)


"Brian Dessent" wrote (before I subscribed to the list):

"Eric D. Williams" wrote:

Thanks for the advice and help.  I just reviewed your post in the list
archive (html). I will give the snapshot a try.  Is there any further
reference to the SYSTEMROOT issue you referred to, e.g., date?  Just so
I am clear you are suggesting a 'snapshot' of the -inst- not the
cygwin1, yes?

In this case the problem was in the DLL, so you can replace just that. Normally you can pretty much always test snapshots by just replacing the
cygwin1.dll, but you don't want to do that if the includes, libs etc.
are from a much older version. I.e. don't try to mix 1.5.16+ DLL with
1.5.12 includes and libs. And obviously if the cause of the problem is
in one of the utils, then replacing just the DLL won't work.

And that indeed fixed my problem, thank you very much. In fact the strace output seemed to indicate this was an issue by not indicating that the environment variable was present/set. In other words, if I recall correctly, it dropped after the first fork() (and as you may know, exim is a mother forker [it fork()'s a great deal] :)


Also, thanks for the explanation below, I have included with this post to increase the hits in the google.com cache. Thank you Brian et. al., your help saved me from another sleepless night :).

Eric


The SYSTEMROOT thing is explained in the archives.  It seems to come up
every so often.  The gist of it is that the windows socket functions
require that variable to be set, else they all fail.  When you have
applications like Apache that launch child processes it's often
customary in the unix world to strip their environment of all
nonessential variables, and if that happens with SYSTEMROOT then the
process can't use sockets.  Code was added in the DLL to try to always
put the variable back when launching a process if it was stripped, but
there was a recent regression that caused it to fail in certain
circumstances (which was fixed promptly by cgf.)

Brian

--
Unsubscribe info:      http://cygwin.com/ml/#unsubscribe-simple
Problem reports:       http://cygwin.com/problems.html
Documentation:         http://cygwin.com/docs.html
FAQ:                   http://cygwin.com/faq/



-- Eric Williams, Pres. Information Brokers, Inc. Phone: +1 202.889.4395 http://www.infobro.com/ Fax: +1 202.889.4396 eric_williams@infobro.com For More Info: info@infobro.com PGP Public Key http://new.infobro.com/KeyServ/EricDWilliams.asc Finger Print: 1055 8AED 9783 2378 73EF 7B19 0544 A590 FF65 B789 ---------------------------------------------------------------- The information in this message is confidential. It is intended solely for addressee(s). Access to this message by anyone else is unauthorized. If you are not the intended recipient, any disclosure, copying, distribution or any action taken or omitted to be taken in reliance on it, is prohibited and may be unlawful.

--
Unsubscribe info:      http://cygwin.com/ml/#unsubscribe-simple
Problem reports:       http://cygwin.com/problems.html
Documentation:         http://cygwin.com/docs.html
FAQ:                   http://cygwin.com/faq/


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