Weblocks vs Webactions?

15 views
Skip to first unread message

petter

unread,
Sep 15, 2009, 1:04:01 PM9/15/09
to weblocks
I've been using Portable AllegroServe with Webactions for some time,
but it seems like it does not compile/work under the latest SBCL. I
managed to get it to compile, but it does not work.

So, before I spend time debugging it I was thinking maybe check out
some of the "new" and maybe more supported stuff like Weblocks. So how
does Weblocks compare to Webactions? Do you have any experience
migrating from Webaction based apps to Weblocks apps?

Leslie P. Polzer

unread,
Sep 15, 2009, 1:21:45 PM9/15/09
to webl...@googlegroups.com

petter wrote:

> So, before I spend time debugging it I was thinking maybe check out
> some of the "new" and maybe more supported stuff like Weblocks. So how
> does Weblocks compare to Webactions?

I took a glance at the manual. Some concepts are familiar but
a bunch of other things are handled differently in Weblocks.

In particular Webactions doesn't seem to have Weblocks' concept
of widgets and a session-scoped tree of them.

Maybe you're best off with a combination of Hunchentoot, the action
part of Weblocks (it's simple to use outside of Weblocks) and
HTML-TEMPLATE.

The migration effort depends to a large deal on your application
code.

Leslie

petter

unread,
Sep 15, 2009, 2:11:41 PM9/15/09
to weblocks
Thank you for your reply.

On Sep 15, 7:21 pm, "Leslie P. Polzer" <s...@viridian-project.de>
wrote:

> The migration effort depends to a large deal on your application
> code.

I guess I should just look at some of the examples and just try to
port a smaller portion of the app.

I'm currently storing CL-SQL objects in the websession variables and
then CLP to generate the dynamic HTML code, but I guess I could use
LSP instead. I'm also using the virtual host part of AllegroServe,
does Hunchentoot have something similar?

Leslie P. Polzer

unread,
Sep 16, 2009, 6:06:54 AM9/16/09
to webl...@googlegroups.com

petter wrote:

> I'm currently storing CL-SQL objects in the websession variables and
> then CLP to generate the dynamic HTML code, but I guess I could use
> LSP instead. I'm also using the virtual host part of AllegroServe,
> does Hunchentoot have something similar?

No, but it's easy to come up with. You just need a dispatcher
that uses the host:port combination to decide on the handler
to invoke.

Weblocks has similar functionality.

Leslie

Ian Eslick

unread,
Sep 16, 2009, 10:22:29 AM9/16/09
to webl...@googlegroups.com
Does the host:port dispatch happen before the webapp path dispatch?
(i.e. could you run a version of the same webapp at the same path
extension two different virtual hosts?)

Ian

Leslie P. Polzer

unread,
Sep 16, 2009, 10:52:40 AM9/16/09
to webl...@googlegroups.com

Ian Eslick wrote:
>
> Does the host:port dispatch happen before the webapp path dispatch?

Hostname trumps prefix. See SORT-WEBAPPS (src/application.lisp).

By the way the port isn't honored in the current code.


> (i.e. could you run a version of the same webapp at the same path
> extension two different virtual hosts?)

Some word is missing here, I think you meant to write "extension of two".
If the question is whether it is possible to have multiple hostnames
dispatch to the same webapp then the answer is yes. You can specify
a list of hostnames for a webapp.

Leslie

Ian Eslick

unread,
Sep 16, 2009, 12:10:21 PM9/16/09
to webl...@googlegroups.com
Awesome, that's new code since I last looked at that section. It's
hard to keep up with all the great things your team has been doing.
Is there a simple way, even if just by checkin titles, to see a good
summary of the enhancements that are made since docs can't hope to
keep up?

Ian

Leslie P. Polzer

unread,
Sep 16, 2009, 1:24:49 PM9/16/09
to webl...@googlegroups.com

Ian Eslick wrote:
>
> Awesome, that's new code since I last looked at that section. It's
> hard to keep up with all the great things your team has been doing.

Why thanks, to me it seems we're rather crawling along due to lack
of time and personnel...


> Is there a simple way, even if just by checkin titles, to see a good
> summary of the enhancements that are made since docs can't hope to
> keep up?

Heh. I think every new feature as well as major incompatible changes
made their way to the list as announcements. But of course that doesn't
help if you need to catch up; in that case try the release announcements.
0.8.3 and 0.8.2 had pretty detailed ones IIRC. They can be found on my
blog (http://blog.viridian-project.de/), on Lispforum and on c.l.l.

HTH,

Leslie

--
http://www.linkedin.com/in/polzer

petter

unread,
Sep 18, 2009, 8:27:46 PM9/18/09
to weblocks
I keep getting the following warning:

WEBLOCKS-DEMO> (start-weblocks-demo :port 8080)
WARNING:
An instance of WEBLOCKS-DEMO with name weblocks-demo is already
running, ignoring start request
NIL

According to netstat -l there is nobody listening on port 8080. I've
seen som comments on the net that this message can be ignored, but I
don't get any listeners so something is wrong...

Leslie P. Polzer

unread,
Sep 19, 2009, 2:31:52 AM9/19/09
to webl...@googlegroups.com

If there's a problem then it's likely to be unrelated to the message
above.

Are you using -stable or -dev?

petter

unread,
Sep 19, 2009, 3:50:46 AM9/19/09
to weblocks
> Are you using -stable or -dev?

I'm using stable. But Hunchentoot does not seem to work, so the
problem is elsewhere...

Leslie P. Polzer

unread,
Sep 19, 2009, 4:11:20 AM9/19/09
to webl...@googlegroups.com

Are you using all 1.0 versions of Chunga and Hunchentoot?

petter

unread,
Sep 19, 2009, 9:59:00 AM9/19/09
to weblocks


On Sep 19, 10:11 am, "Leslie P. Polzer" <s...@viridian-project.de>
wrote:

> Are you using all 1.0 versions of Chunga and Hunchentoot?

* (loop for sys in '(:hunchentoot :chunga) collect (list sys
(asdf:component-version (asdf:find-system sys))))

((:HUNCHENTOOT "1.0.0") (:CHUNGA "0.2.1"))


The Chunga was the one which was distributed with gentoo. I unmerged
it and downloaded it manually:

* (loop for sys in '(:hunchentoot :chunga) collect (list sys
(asdf:component-version (asdf:find-system sys))))

((:HUNCHENTOOT "1.0.0") (:CHUNGA "1.0.0"))

But still, the result is almost the same. Now I get a listen state on
the port, but no reply. I guess I'll have to check all the gentoo
installed lisp code as they tend to be little out of date.

Leslie P. Polzer

unread,
Sep 20, 2009, 5:30:50 AM9/20/09
to weblocks
On Sep 19, 3:59 pm, petter <petter.gus...@gmail.com> wrote:

> But still, the result is almost the same. Now I get a listen state on
> the port, but no reply. I guess I'll have to check all the gentoo
> installed lisp code as they tend to be little out of date.

I recommend using clbuild for Weblocks development.

petter

unread,
Sep 21, 2009, 9:57:10 AM9/21/09
to weblocks


On Sep 20, 11:30 am, "Leslie P. Polzer" <leslie.pol...@gmx.net> wrote:

> I recommend using clbuild for Weblocks development.

Looks like what gentoo portege is trying to do, but of course it's
difficult to qualify so many packages. It seems like there are some
missing dependencies and problems with clbuild as well:

./clbuild install weblocks cl-previalence
./clbuild lisp --eval "(asdf:oos 'asdf:load-op :weblocks-demo)"
component :TRIVIAL-TIMEOUT not found, required
by
./clbuild install trivial-timeout
./clbuild lisp --eval "(asdf:oos 'asdf:load-op :weblocks-demo)"
component :SALZA2 not found, required
by
./clbuild install salza2
./clbuild lisp --eval "(asdf:oos 'asdf:load-op :weblocks-demo)"
component :FARE-MATCHER not found, required by
./clbuild install fare-matcher
The following extra dependencies were found: fare-utils
include dependencies in update? (Y/n)
UPDATE wget fare-matcher
Finished getting.

gzip: stdin: unexpected end of file
tar: Child returned status 1
tar: Error exit delayed from previous errors
No changes!
Pushing to "/home/petter/lisp/clbuild-test/clbuild/source/fare-
matcher"...
No recorded local changes to push!
UPDATE wget fare-utils
Finished getting.

gzip: stdin: unexpected end of file
tar: Child returned status 1
tar: Error exit delayed from previous errors
No changes!


Petter
Reply all
Reply to author
Forward
0 new messages