Stefano Mazzocchi wrote:
>
> Some suggestions for the CVS
>
> 1) stylebook is no longer used nor supported. We should remove it from
> our /lib directory
>
+1
> 2) having binary files in CVS is bad, but having 'compressed jars'
> inside the /lib directory is even worse because it both slows down CVS
> downloads (gzip doesn't work on compressed jars as it does on
> uncompressed ones) and slows Cocoon starting phase.
>
> The price to pay is some disk space, but given the price per MB of HD,
> this should be a resonable price.
>
> 3) since we'll be including a bunch of software under different
> licenses, we should come up with a /legal directory where we place all
> this stuff.
>
+1
> 4) we have three webapp directories 'webapp', webapp.site and
> webapp.tutorial, should we have only one?
>
I think we should add the tutorial to the 'webapp'. The difference
between webapp.site and webapp is that webapp contains the demos and
the site. So we could combine.
+1
> 5) the build file contains a number of properties, should we move them
> out into a property file which is flatter and easier to update?
>
+1
> 6) the code contains a number of things that I found.... well.... let's
> me look for a polite way of saying this.... hmmm, ehmmmm, errrr, .....
> fucked.
>
> if (this.getLogger().isDebugEnabled()) {
> incRequestCount();
> if (this.getLogger().isDebugEnabled()) {
> this.debug(environment, null, null);
> }
> }
>
> do you know where this code is? it's executed every single request
> (Cocoon.java:623). Ok, it doesn't decrease performance, but let's look
> into the code after we apply some automatic search/replace tools, ok?
>
> [well, I *hope* the above came after S&R or we have some serious
> problems]
>
> Moreover, there are classes (Main.java for example) where logging is
> still not wrapped by channel cheching code. This *does* decrease
> performance on command line operations.
>
> 7) still looking into Cocoon.java:748
>
> /**
> * Increment active request count for incoming requests, and save
> this
> * result if it's the maximum.
> */
> private static synchronized void incRequestCount() {
> if (++activeRequestCount > maxRequestCount) {
> maxRequestCount = activeRequestCount;
> }
> }
>
> /**
> * Decrement active request count.
> */
> private static synchronized void decRequestCount() {
> --activeRequestCount;
> }
>
> what are these for? are you aware of the fact that adding a synchronized
> methods to the front door might create a *huge* bottleneck on high
> load!?!?!
>
> Enough for now
>
> [Carsten, I think you can resurrect your directory-cleaning suggestions
> now]
Hmm, let's see: ah yes, no release planned for the next days! So I can
repost my suggestion soon. (Just a joke to survive this grey day...).
Carsten
495 mails to go...
>
> --
> Stefano Mazzocchi One must still have chaos in oneself to be
> able to give birth to a dancing star.
> <[EMAIL PROTECTED]> Friedrich Nietzsche
> --------------------------------------------------------------------
>
>
>
> ---------------------------------------------------------------------
> To unsubscribe, e-mail: [EMAIL PROTECTED]
> For additional commands, email: [EMAIL PROTECTED]
>
---------------------------------------------------------------------
To unsubscribe, e-mail: [EMAIL PROTECTED]
For additional commands, email: [EMAIL PROTECTED]
