You are viewing a plain text version of this content. The canonical link for it is here.
Posted to modperl@perl.apache.org by Kenneth Lee <ke...@alfacomtech.com> on 2000/07/20 09:10:15 UTC
Apache::ASP+
Hi all!
Any one started to make one? It seems a good idea. Of coz it's not simple
to build the whole complex platform (NGWS), but the concept of server-side
object and the event driven model is too interesting to me.
Kenneth
Re: Apache::ASP+
Posted by Joshua Chamas <jo...@chamas.com>.
Kenneth Lee wrote:
>
> See this link
>
> http://www.asptoday.com/articles/20000712.htm
>
> Paul Lindner wrote:
> >
> > Actually I think that Apache::ASP is coming right along with it's
> > support of XML and the underlying behaviors.
> >
Thanks Paul, but there is always room for improvement eh? :)
So, I took a good read of ASP+ at the link above, and was
less than impressed, but could see how some of the things
would be useful.
ASP+ looks like its supposed to be backwards compatible with
ASP, so adding its features to Apache::ASP would seem like
something that could/should be done, with the right configuration
options, so that the runtime penalties of some of the features
wouldn't be at everyone's detriment.
Below is a list of features, and where I see them with respect
to Apache::ASP. Those that "WILL BE DONE" just requires the
proper time & motivation for me to do. Those that "COULD BE DONE"
I think we should wait on until the ASP+ standard is out of beta
and we can sort out the good from the bad. I don't think we should
implement everything "just because", but thoughtfully take more
of a per need approach to these newer features.
--Joshua
ALREADY DONE
------------
- cookieless sessions, via SessionQueryParse config
- server farm session management, via StateDir NFS or other net mount
- sessions persist across server reboots, natural with StateDir
WILL BE DONE
------------
- auto fill HTML form field feature, probably via FormFill config
or $Response->{FormFill} member being set, will look at how others
do it first
- sessions stored in database, might implement via Apache::Session
- output caching, maybe with $Response->{Cache} member, need to
do more research here
COULD BE DONE
-------------
- new events like Application_BeginRequest
- Application_BeginRequest possibly same as Script_OnStart ?
- binding of form field names directly to variables in scripts
- new server side event model more like client side DOM scripting
- onserverclick event for form fields
- server side control binding is more HTMLish way of programming
- new <asp: ...> controls / tags
- support of SOAP and server side object access
- config.web configuration file, likely similar to .htaccess(?)
- global.asax instead of global.asa
- Application_OnEnd/OnStart occuring after global.asax
is modified
- state server properties
- per page configurable error page
- trace object, different method of debugging
Re: Apache::ASP+
Posted by Paul Lindner <pl...@redhat.com>.
On Thu, Jul 20, 2000 at 12:36:45PM -0700, Dmitry Beransky wrote:
> At 10:56 AM 7/20/00, you wrote:
> >See this link
> >
> >http://www.asptoday.com/articles/20000712.htm
>
> Very interesting (now I want to buy the enitre book and see what else is in
> there). There are some aspects of the new ASP I really like especially
> those targeted at improvements of the workflow (content/code
> separation). It looks to me, though, that this improvement comes at the
> cost of parsing HTML document into a DOM tree (although, I've always wished
> there were a cheap way to manipulate the DOM of a document on the
> server). I've also noticed that the markup used in the excerpt is a weird
> combination of HTML and XHTML (those <P /> tags). I wonder if ASP+ pages
> are [can be?] required to be written in XHTML to speed up the parsing time...
Yeah, I don't much care for the runat="server" way of specifying that
content should be handled by the server. I would think that they
could have used some enclosing tag to denote that the forms were to be
preprocessed...
Some of the stuff they talk about is already handled by Apache::ASP
(such as intelligently dealing with session state..) Script_OnStart hooks,
Debug Tracing, etc, etc..
BTW I find that the XMLSubsMatch included in the latest Apache::ASP
simpler and cleaner way of doing much of ASP+ (especially this
business of the special asp: namespace..)
I can say
XMLSubsMatch Apps:\w+
This converts all calls for tags like <Apps:foo /> or
<Apps:bar>...</Apps:bar> to equivalent functions in the Apps package.
Very, very useful.
It will only get more powerful when compile-time tags are added. This
will give a big performance boost for code that doesn't need to run on
every request, or allow for tags to insert code into the compiled
Apache::ASP page..
For example, you could define a tag named <Util:isloggedin>, and use
it like such:
<Util:isloggedin>
Welcome back
</Util:isloggedin>
You'd have this on the backend:
sub Util::isloggedin {
my ($args, $text);
return(qq( if ($Session->{isloggedin}) { ) .
qq( $Response->Write(' ) .
$text
qq( } );
);
}
almost closure like, if you ask me...
--
Paul Lindner
plindner@redhat.com
Red Hat Inc.
Re: Apache::ASP+
Posted by Joshua Chamas <jo...@chamas.com>.
Dmitry Beransky wrote:
>
> At 10:56 AM 7/20/00, you wrote:
> >See this link
> >
> >http://www.asptoday.com/articles/20000712.htm
>
> Very interesting (now I want to buy the enitre book and see what else is in
> there). There are some aspects of the new ASP I really like especially
> those targeted at improvements of the workflow (content/code
> separation). It looks to me, though, that this improvement comes at the
> cost of parsing HTML document into a DOM tree (although, I've always wished
> there were a cheap way to manipulate the DOM of a document on the
> server). I've also noticed that the markup used in the excerpt is a weird
> combination of HTML and XHTML (those <P /> tags). I wonder if ASP+ pages
> are [can be?] required to be written in XHTML to speed up the parsing time...
>
Whatever makes it into Apache::ASP, I assure you, I won't
be doing any DOM tree parsing. My approach tends to be
look just for what you need to look for in the HTML
and trigger off that while parsing, so when using XMLSubsMatch
or the <asp: ...> controls, the asp parser would only look
for <asp: ...> For automatic form filling, would just be
picking out <form>...</form>, etc.
These features can then be activated per need, thus a user
has control over what compile time performance their
environment will have.
Further, ASP scripts only get compiled once per process,
and only once per server with smart use of Apache::ASP->Loader()
so even complex compile time parsing won't be as expensive
as one might otherwise think.
As we move forward with integration of ASP+ features into
Apache::ASP, please keep me up on which features you would
find particularly useful to yourself.
-- Joshua
_________________________________________________________________
Joshua Chamas Chamas Enterprises Inc.
NodeWorks >> free web link monitoring Huntington Beach, CA USA
http://www.nodeworks.com 1-714-625-4051
Re: Apache::ASP+
Posted by Dmitry Beransky <db...@ucsd.edu>.
At 10:56 AM 7/20/00, you wrote:
>See this link
>
>http://www.asptoday.com/articles/20000712.htm
Very interesting (now I want to buy the enitre book and see what else is in
there). There are some aspects of the new ASP I really like especially
those targeted at improvements of the workflow (content/code
separation). It looks to me, though, that this improvement comes at the
cost of parsing HTML document into a DOM tree (although, I've always wished
there were a cheap way to manipulate the DOM of a document on the
server). I've also noticed that the markup used in the excerpt is a weird
combination of HTML and XHTML (those <P /> tags). I wonder if ASP+ pages
are [can be?] required to be written in XHTML to speed up the parsing time...
Dmitry
Re: Apache::ASP+
Posted by Perrin Harkins <pe...@primenet.com>.
On Fri, 21 Jul 2000, Kenneth Lee wrote:
> See this link
>
> http://www.asptoday.com/articles/20000712.htm
He sure makes a lot of fuss over adding the functionality of
HTML::FillInForms. The action => sub stuff doesn't look very difficult to
me, but it also doesn't look like a significant improvement over more
obvious ways to do this. Ultimately, it kind of looks like Microsoft is
struggling to add some of the basic features of WebObjects 1.0.
- Perrin
Re: Apache::ASP+
Posted by Kenneth Lee <ke...@alfacomtech.com>.
See this link
http://www.asptoday.com/articles/20000712.htm
Paul Lindner wrote:
>
> On Thu, Jul 20, 2000 at 03:10:15PM +0800, Kenneth Lee wrote:
> > Hi all!
> >
> > Any one started to make one? It seems a good idea. Of coz it's not simple
> > to build the whole complex platform (NGWS), but the concept of server-side
> > object and the event driven model is too interesting to me.
>
> Actually I think that Apache::ASP is coming right along with it's
> support of XML and the underlying behaviors.
>
> I'm not well versed in Microsoft's Next Generation Windows System
> (NGWS) but would appreciate hearing more about how it differs from
> some of the latest Apache::ASP enhancements.
>
> --
> Paul Lindner
> plindner@redhat.com
> Red Hat Inc.
Re: Apache::ASP+
Posted by Paul Lindner <pl...@redhat.com>.
On Thu, Jul 20, 2000 at 03:10:15PM +0800, Kenneth Lee wrote:
> Hi all!
>
> Any one started to make one? It seems a good idea. Of coz it's not simple
> to build the whole complex platform (NGWS), but the concept of server-side
> object and the event driven model is too interesting to me.
Actually I think that Apache::ASP is coming right along with it's
support of XML and the underlying behaviors.
I'm not well versed in Microsoft's Next Generation Windows System
(NGWS) but would appreciate hearing more about how it differs from
some of the latest Apache::ASP enhancements.
--
Paul Lindner
plindner@redhat.com
Red Hat Inc.