delorie.com/archives/browse.cgi   search  
Mail Archives: cygwin-developers/2002/01/29/01:27:57

Mailing-List: contact cygwin-developers-help AT cygwin DOT com; run by ezmlm
List-Subscribe: <mailto:cygwin-developers-subscribe AT cygwin DOT com>
List-Archive: <http://sources.redhat.com/ml/cygwin-developers/>
List-Post: <mailto:cygwin-developers AT cygwin DOT com>
List-Help: <mailto:cygwin-developers-help AT cygwin DOT com>, <http://sources.redhat.com/ml/#faqs>
Sender: cygwin-developers-owner AT cygwin DOT com
Delivered-To: mailing list cygwin-developers AT cygwin DOT com
From: "Gary R. Van Sickle" <g DOT r DOT vansickle AT worldnet DOT att DOT net>
To: <cygwin-developers AT cygwin DOT com>
Subject: RE: Daemon reviewer
Date: Tue, 29 Jan 2002 00:27:43 -0600
Message-ID: <NCBBIHCHBLCMLBLOBONKEEMNCJAA.g.r.vansickle@worldnet.att.net>
MIME-Version: 1.0
X-Priority: 3 (Normal)
X-MSMail-Priority: Normal
X-Mailer: Microsoft Outlook IMO, Build 9.0.2416 (9.0.2911.0)
In-Reply-To: <01cd01c1a6dd$4619e550$0200a8c0@lifelesswks>
X-MimeOLE: Produced By Microsoft MimeOLE V6.00.2600.0000
Importance: Normal

> -----Original Message-----
> From: cygwin-developers-owner AT cygwin DOT com
> [mailto:cygwin-developers-owner AT cygwin DOT com]On Behalf Of Robert Collins
> Sent: Saturday, January 26, 2002 8:50 PM
> To: cygwin-developers AT cygwin DOT com
> Subject: Daemon reviewer
> Importance: High
>
>
> Gary, will you have to time make some sort of comment on the code/design
> quality in the next (say) fortnight?
> If not, please say so.
>

Code quality: Excellent.  I'd wager it'd come in dead-last in an obfuscated Perl
contest ;-).  I'm seeing a lot of FIXME's, but as you indicate none are perfect,
and it's optional for now anyway (and at least they're well-commented FIXMEs
;-)).

Design quality: After taking a look, I think I'm gonna have to recuse myself,
sorry.  I'm just not anywhere near familiar enough with cygwin1.dll's internals
to be able to make any sort of meaningful judgement on what I'm looking at in
but a fortnight.  I've been looking only at cygserver to date though;  I'll diff
the DLL code and see if I can make out anything there that might look like a
problem for the non-daemon case.

As for the soundness of the design in the most general of terms (i.e. your
list), FWIW I think a Cygwin 'Uberprocess' would probably make some things much
more straightforward, e.g. the "rm-queue" could be put in there and then never
cause a process to block.

I'm doing my level best as I type this to build the latest cygwin_daemon branch,
as I have been on and off for about a week or so now since you last updated it
(you know how it is with us CVS-challenged ;-))... and it just failed again for
the bajillionth time, great.[1]  Well, I'll get it built before the Rapture (I
hope), and once I do I'll use it exclusively here and at work, give it the
onceover twice with the daemon, and then lean on it without the daemon (but with
the daemonized DLL) until such time as:

- It becomes part of the regular distro.
- You drop it and nobody else picks it up.
- Problems reveal themselves.
- Injury or illness precludes me from carrying out my duties as tribal warlord.

[1] Could you maybe just email me the fricken binaries and/or put them up
somewhere so I don't have to hit the Cygwin CVS server another few hundred times
before I get slot A to fit into tab B here?  I'm just not getting the
cygwin_daemon branch to build now no matter what I do.  If it ain't a "multiple
definition of `tzset'" its a dozen other damn things....

--
Gary R. Van Sickle
Brewer.  Patriot.

- Raw text -


  webmaster     delorie software   privacy  
  Copyright © 2019   by DJ Delorie     Updated Jul 2019