You are viewing a plain text version of this content. The canonical link for it is here.
Posted to websh-dev@tcl.apache.org by Ronnie Brunner <ro...@netcetera.ch> on 2002/04/16 11:09:12 UTC
Re: web::config for mod_websh paths
> > Have an accessor in web::config that sets up these vars at startup
> > time for mod_websh and at request time for CGI (which is good enough
> > there).
>
> Accessing it through web::config isn't that hot, either, because it's
> not really a global config option.
Yes, but so are [web::config version] and [web::config copyright] ->
we don't mess up anything that's not messed up yet anyway.
> > -> a common accessor in cfg.c with specific implementations in
> > modwebsh_ap.c and modwebsh_cgi.c respectively should do the trick.
>
> Maybe we could do something sneaky with iPtr->scriptFile in Tcl
> itself? Too clever, probably...
I would access stuff like
[web::config mod_websh] : conf->scriptName
[web::config server_root] : ap_server_root
[web::config interpclass] : id (interpool.c:320) (only within a request)
[web::config doument_root] : ap_document_root(request_rec *r)
This seems the most straight forward approach to me.
------------------------------------------------------------------------
Ronnie Brunner ronnie.brunner@netcetera.ch
Netcetera AG, 8040 Zuerich phone +41 1 247 79 79 Fax: +41 1 247 70 75
---------------------------------------------------------------------
To unsubscribe, e-mail: websh-dev-unsubscribe@tcl.apache.org
For additional commands, e-mail: websh-dev-help@tcl.apache.org
Re: web::config for mod_websh paths
Posted by Ronnie Brunner <ro...@netcetera.ch>.
> I have .ws3 files, set up, of course, with mod_websh. What I would
> like to do is to automate the loading of websh.com.gen for everything
> in a certain directory. I don't want that file loaded for every .ws3
> file I request, though, because it might interfere with other things I
> am testing.
Your .ws3 file are either code or templates. (meaning the .ws3 file
needs be eval'd or web::putxfile'd). The setup is a little different
in both cases. The easier case is code files:
wep::interpmap {file} {
# what needs to be passed to websh.com.gen
if {[match_websh.com.gen $file]} {
return /path/to/websh.com.gen
}
# default (all other files use their own class)
return $file
}
Note: for this to work, match_websh.com.gen is a proc that finds out
whether it is a request that must include websh.com.gen using some
pattern match or regex or similar and the websh.com.gen script must do
the following (at the end of the code)
# all stuff in websh.com.gen
# ...
web::include [web::config script]
# EOF
The setup for template files would be
wep::interpmap {file} {
# what needs to be passed to websh.com.gen
if {[match_websh.com.gen $file]} {
return /path/to/websh.com.gen
}
# default (all other files use their own class)
return /path/to/defaulthandler
}
where match_websh.com.gen is the same as above, but the websh.com.gen
script must look something like
# all stuff in websh.com.gen
# ...
web::putxfile [web::config script]
# EOF
and the defaulthandler script looks something like
# this is the default handler
web::putxfile [web::config script]
# EOF
You can actually mix the approach for each interpreter class you
define, but this could be confusing. A better solution would be to add
a second handler, such as a handler for .whtml (templates) and a
handler for .ws3 (scripts) using the following generic interpmap setup:
wep::interpmap {file} {
# scripts
if {[string match *.ws3 $file]} {
# they all have their own interp class
return $file
}
# templates (i.e. all *.whtml files)
return /path/to/defaulthandler
}
where the defaulthandler is essentialy a [web::putxfile [web::config script]]
command, just as above.
Ronnie
------------------------------------------------------------------------
Ronnie Brunner ronnie.brunner@netcetera.ch
Netcetera AG, 8040 Zuerich phone +41 1 247 79 79 Fax: +41 1 247 70 75
---------------------------------------------------------------------
To unsubscribe, e-mail: websh-dev-unsubscribe@tcl.apache.org
For additional commands, e-mail: websh-dev-help@tcl.apache.org
Re: web::config for mod_websh paths
Posted by "David N. Welton" <da...@dedasys.com>.
Ronnie Brunner <ro...@netcetera.ch> writes:
> in the WebshConfig File, you still have to do this stuff manually:
> In your case, what you need to do is:
> Add a handler for websh (e.g. extension .ws3) and then you can
> (optionally) add a WebshConfig that overwrites web::interpmap.
> In you example: the WebshConfig file should throw an error, sionce
> web::config is unknown. Without that line, it'll generate a Server
> Error 500, since howdydoodee is not a valid interpreter class
> (i.e. no Websh3 file).
> What's the problem with interfereing with other stuff? Could it be
> that the current Websh handler is setup for *.html? Couldn't you add
> the handler in a <Directory> clause? This would route e.g. /websh/*
> files to mod_websh only.
> ... Sorry for not being very constructive here, I don't really know
> what you're trying to do ...
I have .ws3 files, set up, of course, with mod_websh. What I would
like to do is to automate the loading of websh.com.gen for everything
in a certain directory. I don't want that file loaded for every .ws3
file I request, though, because it might interfere with other things I
am testing.
Is that clearer? Sorry that the previous message wasn't.
--
David N. Welton
Consulting: http://www.dedasys.com/
Personal: http://www.dedasys.com/davidw/
Free Software: http://www.dedasys.com/freesoftware/
Apache Tcl: http://tcl.apache.org/
---------------------------------------------------------------------
To unsubscribe, e-mail: websh-dev-unsubscribe@tcl.apache.org
For additional commands, e-mail: websh-dev-help@tcl.apache.org
Re: web::config for mod_websh paths
Posted by Ronnie Brunner <ro...@netcetera.ch>.
> proc web::interpmap {file} {
> puts stderr [web::config script]
> return howdydoodee
> }
>
> doesn't seem to have any effect at all...
Well, I know! When I added the web::config stuff, I realized, that
in the main interpreter, we don't have the web::config command anyway
-> this stuff is mainly usefull to configure the actual applications
(since there you don't want to bother with changing absolute paths).
in the WebshConfig File, you still have to do this stuff manually: In
your case, what you need to do is:
Add a handler for websh (e.g. extension .ws3) and then you can
(optionally) add a WebshConfig that overwrites web::interpmap.
In you example: the WebshConfig file should throw an error, sionce
web::config is unknown. Without that line, it'll generate a Server
Error 500, since howdydoodee is not a valid interpreter class (i.e. no
Websh3 file).
...
What's the problem with interfereing with other stuff? Could it be
that the current Websh handler is setup for *.html? Couldn't you add
the handler in a <Directory> clause? This would route e.g. /websh/*
files to mod_websh only.
... Sorry for not being very constructive here, I don't really know what
you're trying to do ...
------------------------------------------------------------------------
Ronnie Brunner ronnie.brunner@netcetera.ch
Netcetera AG, 8040 Zuerich phone +41 1 247 79 79 Fax: +41 1 247 70 75
---------------------------------------------------------------------
To unsubscribe, e-mail: websh-dev-unsubscribe@tcl.apache.org
For additional commands, e-mail: websh-dev-help@tcl.apache.org
Re: web::config for mod_websh paths
Posted by "David N. Welton" <da...@dedasys.com>.
Ronnie Brunner <ro...@netcetera.ch> writes:
> Can't be: either we talk about the mainInterp (which is already
> there) or we talk about the new interp (for the pool) in which we
> don't eval any code until all the init stuff is called (otherwise we
> would have a serious chicken & egg problem ;-)
> -> should be straight forward except for the [web::config
> interpclass] which is not available in mainInterp (which also makes
> sense)
Hrm... well, it's not working. web::config script works, I tested it
out in a page. Maybe I'm just not getting it...
--
David N. Welton
Consulting: http://www.dedasys.com/
Personal: http://www.dedasys.com/davidw/
Free Software: http://www.dedasys.com/freesoftware/
Apache Tcl: http://tcl.apache.org/
---------------------------------------------------------------------
To unsubscribe, e-mail: websh-dev-unsubscribe@tcl.apache.org
For additional commands, e-mail: websh-dev-help@tcl.apache.org
Re: web::config for mod_websh paths
Posted by "David N. Welton" <da...@dedasys.com>.
Ronnie Brunner <ro...@netcetera.ch> writes:
> Can't be: either we talk about the mainInterp (which is already
> there) or we talk about the new interp (for the pool) in which we
> don't eval any code until all the init stuff is called (otherwise we
> would have a serious chicken & egg problem ;-)
> -> should be straight forward except for the [web::config
> interpclass] which is not available in mainInterp (which also makes
> sense)
So what is the magic sequence to have a locally hosted tcl site that
doesn't interfere with other stuff? Everything seems to be working on
www.apache.org, I just need to duplicate it locally in order to clean
it up.
I'm a bit stumped/frustrated...
proc web::interpmap {file} {
puts stderr [web::config script]
return howdydoodee
}
doesn't seem to have any effect at all...
--
David N. Welton
Consulting: http://www.dedasys.com/
Personal: http://www.dedasys.com/davidw/
Free Software: http://www.dedasys.com/freesoftware/
Apache Tcl: http://tcl.apache.org/
---------------------------------------------------------------------
To unsubscribe, e-mail: websh-dev-unsubscribe@tcl.apache.org
For additional commands, e-mail: websh-dev-help@tcl.apache.org
Re: web::config for mod_websh paths
Posted by Ronnie Brunner <ro...@netcetera.ch>.
Can't be: either we talk about the mainInterp (which is already there)
or we talk about the new interp (for the pool) in which we don't eval
any code until all the init stuff is called (otherwise we would have a
serious chicken & egg problem ;-)
-> should be straight forward except for the [web::config interpclass]
which is not available in mainInterp (which also makes sense)
> But wait...
>
> Let me see if I have this right.
>
> web::config is created in
>
> cfg_Init, which is called by Websh_Init
>
> Websh_Init is called by createWebInterp
>
> createWebInterp is called by poolGetWebInterp, but at the end of the
> file - after the various interpmap things have already been called.
>
> Doesn't this leave us basically in the same place as before? I mean,
> it's not so bad if you have one web site per apache, but the idea of
> doing per-site include files is very convenient. Your interpmap
> script solves that problem nicely.
>
> Or am I missing something?
>
------------------------------------------------------------------------
Ronnie Brunner ronnie.brunner@netcetera.ch
Netcetera AG, 8040 Zuerich phone +41 1 247 79 79 Fax: +41 1 247 70 75
---------------------------------------------------------------------
To unsubscribe, e-mail: websh-dev-unsubscribe@tcl.apache.org
For additional commands, e-mail: websh-dev-help@tcl.apache.org
Re: web::config for mod_websh paths
Posted by "David N. Welton" <da...@dedasys.com>.
Ronnie Brunner <ro...@netcetera.ch> writes:
> > > [web::config mod_websh] : conf->scriptName
> >
> > why such an odd name? How about web::config script to mirror 'info
> > script' in Tcl?
>
> sure: way better
>
> > > [web::config server_root] : ap_server_root
> > > [web::config interpclass] : id (interpool.c:320) (only within a request)
> > > [web::config doument_root] : ap_document_root(request_rec *r)
> >
> > > This seems the most straight forward approach to me.
> >
> > Ok, I will add these.
But wait...
Let me see if I have this right.
web::config is created in
cfg_Init, which is called by Websh_Init
Websh_Init is called by createWebInterp
createWebInterp is called by poolGetWebInterp, but at the end of the
file - after the various interpmap things have already been called.
Doesn't this leave us basically in the same place as before? I mean,
it's not so bad if you have one web site per apache, but the idea of
doing per-site include files is very convenient. Your interpmap
script solves that problem nicely.
Or am I missing something?
--
David N. Welton
Consulting: http://www.dedasys.com/
Personal: http://www.dedasys.com/davidw/
Free Software: http://www.dedasys.com/freesoftware/
Apache Tcl: http://tcl.apache.org/
---------------------------------------------------------------------
To unsubscribe, e-mail: websh-dev-unsubscribe@tcl.apache.org
For additional commands, e-mail: websh-dev-help@tcl.apache.org
Re: web::config for mod_websh paths
Posted by Ronnie Brunner <ro...@netcetera.ch>.
> > [web::config mod_websh] : conf->scriptName
>
> why such an odd name? How about web::config script to mirror 'info
> script' in Tcl?
sure: way better
> > [web::config server_root] : ap_server_root
> > [web::config interpclass] : id (interpool.c:320) (only within a request)
> > [web::config doument_root] : ap_document_root(request_rec *r)
>
> > This seems the most straight forward approach to me.
>
> Ok, I will add these.
------------------------------------------------------------------------
Ronnie Brunner ronnie.brunner@netcetera.ch
Netcetera AG, 8040 Zuerich phone +41 1 247 79 79 Fax: +41 1 247 70 75
---------------------------------------------------------------------
To unsubscribe, e-mail: websh-dev-unsubscribe@tcl.apache.org
For additional commands, e-mail: websh-dev-help@tcl.apache.org
Re: web::config for mod_websh paths
Posted by "David N. Welton" <da...@dedasys.com>.
Ronnie Brunner <ro...@netcetera.ch> writes:
> > > Have an accessor in web::config that sets up these vars at
> > > startup time for mod_websh and at request time for CGI (which is
> > > good enough there).
> > Accessing it through web::config isn't that hot, either, because
> > it's not really a global config option.
> Yes, but so are [web::config version] and [web::config copyright] ->
> we don't mess up anything that's not messed up yet anyway.
Ok.
> > > -> a common accessor in cfg.c with specific implementations in
> > > modwebsh_ap.c and modwebsh_cgi.c respectively should do the
> > > trick.
> > Maybe we could do something sneaky with iPtr->scriptFile in Tcl
> > itself? Too clever, probably...
> I would access stuff like
> [web::config mod_websh] : conf->scriptName
why such an odd name? How about web::config script to mirror 'info
script' in Tcl?
> [web::config server_root] : ap_server_root
> [web::config interpclass] : id (interpool.c:320) (only within a request)
> [web::config doument_root] : ap_document_root(request_rec *r)
> This seems the most straight forward approach to me.
Ok, I will add these.
--
David N. Welton
Consulting: http://www.dedasys.com/
Personal: http://www.dedasys.com/davidw/
Free Software: http://www.dedasys.com/freesoftware/
Apache Tcl: http://tcl.apache.org/
---------------------------------------------------------------------
To unsubscribe, e-mail: websh-dev-unsubscribe@tcl.apache.org
For additional commands, e-mail: websh-dev-help@tcl.apache.org