Based on your lists of steps it looks like the slow process is booting to and from 
windows. Why don't we focus on fixing that instead of on other things. Lupus 
reccomends rewriting corcompare using the metadata API, that  would solve the major 
problem. Your solution solves the minor problem but not the major one.

Jackson



-----Original Message-----
From: [EMAIL PROTECTED]
[mailto:[EMAIL PROTECTED] Behalf Of Ben Maurer
Sent: Thursday, June 19, 2003 8:10 AM
To: [EMAIL PROTECTED]
Subject: Re: [Mono-list] How often is the class status page updated?


On Thu, 2003-06-19 at 03:18, Paolo Molaro wrote:
> *) the work is better than the current one (for many meanings of
> better: faster, more maintainable, more complaint to the spec...)
Yes, the work will be:
      * Faster: If the XML is in a good format, a bit of web design
        magic can actually make it faster than it is currently (right
        now we use a table heavy design, I am trying to fix that)
      * More maintainable: Even if we don't get our new corcompare tool,
        this will be easier for Miguel. Right now, to update the page he
        has to 1) boot to windows 2) run the current tool 3) transfer
        the output to his Linux box 4) run the current make file 5)
        transfer. With the transform on the web server he can 1) boot to
        windows 2) run tool 3) transfer. I have talked to him, and this
        is very important to him.
Remember, by "dynamic" I do *not* mean that the files are transformed on
each request. I am planning on caching the results. But since the
transform is taking place on the server, I don't want to take up gobs of
unnecessary CPU time.
_______________________________________________
Mono-list maillist  -  [EMAIL PROTECTED]
http://lists.ximian.com/mailman/listinfo/mono-list
_______________________________________________
Mono-list maillist  -  [EMAIL PROTECTED]
http://lists.ximian.com/mailman/listinfo/mono-list

Reply via email to