webmacro-devel — Discuss development issues related to the core of WebMacro

Re: [Webmacro-devel] 2.0: Use DelegatingTemplateProvider?

[Keats <ke...@ea...>]

----- Original Message -----
> I'd like to propose using Sebastian's DelegatingTemplateProvider as the
> default template provider in WM 2.0.

We agreed to do this a while ago.  I think all we need are some docs on how
to configure it ... hint, hint.

Keats

[Webmacro-devel] Velocity info on Wiki altered

[Marc P. <ma...@an...>]

I've removed the velocity info (keeping links). Example text:

Velocity is a WebMacro clone without the brokering and pluggable directives 
that WebMacro provides.

For information on Velocity, see http://jakarta.apache.org

:)

-- 
Marc Palmer
Contract Java Consultant/Developer

http://www.anyware.co.uk/marc/
http://www.wangjammers.org

[Webmacro-devel] 2.0: Use DelegatingTemplateProvider?

[Marc P. <ma...@an...>]

Hi all,

I'd like to propose using Sebastian's DelegatingTemplateProvider as the 
default template provider in WM 2.0.

This is a far more flexible mechanism than the default template provider, 
and it's proved excellent in the Ignition webapp.

The only thing it's missing (as far as I am concerned) is changes to the 
paths at runtime, but that could be implemented quite easily.

Any objections to this? If not I'll add it to NextRelease.

-- 
Marc Palmer
Contract Java Consultant/Developer

http://www.anyware.co.uk/marc/
http://www.wangjammers.org

Re: [Webmacro-devel] FINALLY! Ignition alpha 1 for download...

[Marc P. <ma...@an...>]

On Fri, 13 Jun 2003 01:19:15 +0100, Marc Palmer <ma...@an...> wrote:

>
> Hi all,
>
> Once again it took longer than I hoped, but that's life!
>
> Finally I am pleased to announce that you can download (WebMacro?) 
> Ignition 1.0 Alpha 1 from:
>
> http://www.wangjammers.org/ignition
>
> I fully realise that it is now Friday 13th. I'm sure this will not go 
> smoothly...

Well there you go. Found a "bug". Missing a "get" method on Visitor 
variable, so the Visitor demo does not display the existing colour you have 
stored when you return to the site. Just a template var issue.
 
Will fix it today, ready for the next release.
 
Marc
 
-- 
Marc Palmer
Contract Java Consultant/Developer
 
http://www.anyware.co.uk/marc/
http://www.wangjammers.org
 

Re: [Webmacro-devel] A running Ignition install

[Marc P. <ma...@an...>]

On Fri, 13 Jun 2003 11:00:32 +0100, Marc Palmer <ma...@an...> wrote:

>
> If you wish to see Ignition demos and/or read the docs, I will be running 
> it on my desktop PC today.
>
> The URI is: http://212.135.236.36:8080/ignition
>
> You may find this more convenient than downloading and installing it, but 
> of course you cannot try writing your own templates.

PS. Before you try the Email or SQL demos, you need to know:
 
1. The Email to/from address must be a valid address in my domain. So, use 
a To:/From: address of ma...@an... for it to work when sending an 
email. The emailing of form params should work anyway.
 
2. Enter a numeric digit or two as the "key" when using the SQL demo. I 
just have some dummy test data there, indexed on a numeric field. It's just 
the default demo templates showing a trivial example.
 
 
-- 
Marc Palmer
Contract Java Consultant/Developer
 
http://www.anyware.co.uk/marc/
http://www.wangjammers.org
 

[Webmacro-devel] A running Ignition install

[Marc P. <ma...@an...>]

If you wish to see Ignition demos and/or read the docs, I will be running 
it on my desktop PC today.

The URI is: http://212.135.236.36:8080/ignition

You may find this more convenient than downloading and installing it, but 
of course you cannot try writing your own templates.

-- 
Marc Palmer
Contract Java Consultant/Developer

http://www.anyware.co.uk/marc/
http://www.wangjammers.org

[Webmacro-devel] FINALLY! Ignition alpha 1 for download...

[Marc P. <ma...@an...>]

Hi all,

Once again it took longer than I hoped, but that's life!

Finally I am pleased to announce that you can download (WebMacro?) Ignition 
1.0 Alpha 1 from:

http://www.wangjammers.org/ignition

I fully realise that it is now Friday 13th. I'm sure this will not go 
smoothly...

This is a preview release for you guys to see if you think this should be 
taken in under the WM project umbrella. There is a source download too, but 
it doesn't include the whole build environment yet. That's a job for 
another day.

Currently, apart from the framework itself, it provides:

* Visitor tracking (using cookie, no login required)
* Visitor-specific value storage (persistent on server)
* Multi-page data "remembering" + commit (pigeonholes)
* XML processing
* Template-based send email (email template has access to all 
helpers+plugins a page can see)
* Template-based send form params by email
* SQL queries
* Date + File manipulation helpers
* RSS news feeds
* Click-through tracking
* Active/Peak session counting
* Redirects

This list will be expanded significantly in the future we hope with your 
help - we already have a long list of components to write, which should be 
easy now the framework is done.

These currently serve as demonstrations of the framework, with full source 
in the bin+src download. Full javadocs are built and deployed with the WAR, 
ready to browse.

The docs are not as good as we'd like right now, particularly for the demo 
plugins, but the core webapp is fully javadoc'd. There is user-level 
documentation for the entire system too, served by the application - 
although detailed info on using the demo components is only available 
through the online help system built in, and will be refined further in 
future.

There are some "innovations" in there we'd like to get feedback on too. The 
help system and type information in particular - it's a big part of the 
concept. No more "hey, what properties are there on variable X?" questions 
from your web designers. Also the notion of "WMScript types" I have sort of 
introduced, to hide some of the Java-ness (i.e. "number" instead of 
java.langInteger/Long, "list" instead of array/Iterator/List/Enumeration) 
when things are semantically the same.

Please give it a spin and let me know what you think. Bear in mind that 
this is quite a large project and has been squeezed out with the "bare 
minimum" required to demo it well to you guys at a quality we personally 
find acceptable.

Next on the agenda are dynamic reloading of config, Transactions, 
Authentication, and field Validation (as opposed to parameter validation 
which we have already). We have some nice plans for these.

We're pretty excited about this. We hope you find it useful, or at the very 
least interesting. We think it has massive potential for bringing WM to the 
masses, not just us Java nerds.

Thanks to Lane for help with Visitor and VLH stuff, and for enthusiasm :)

Good night!
The Wangjammers Team -- Marc Palmer
Contract Java Consultant/Developer

http://www.anyware.co.uk/marc/
http://www.wangjammers.org

Re: [Webmacro-devel] Another parser exception issue

[Eric B. R. <eb...@tc...>]

On Thursday, June 12, 2003, at 12:41  PM, Keats wrote:

> Eric,
>
> I just tested your patch.  The error message is great except that the 
> line
> number is still the EOF line number instead of the start-of-block.

Oh crap.  I'll look into this tonight.  I think to fix this the parser 
is needs to track the "last block opening" position, but that's not a 
big deal.

eric

Re: [Webmacro-devel] Another parser exception issue

[Keats <ke...@ea...>]

Eric,

I just tested your patch.  The error message is great except that the line
number is still the EOF line number instead of the start-of-block.

Keats

Re: [Webmacro-devel] Another parser exception issue

[Keats <ke...@ea...>]

> Do we need some javacc work to do this?
> -Lane

I believe Eric has already implemented the first part -- changing the error
message.  The second part, changing the syntax, would definitely require
javacc work.

I don't know the details, but basically when the parser mode is "looking for
a block" it should transition to "in a block" when it finds a "{" or a
newline.  Any other non-whitespace character should be a parse error (IMHO).

Keats


Kea...@di... wrote:

When you specify an invalid directive option before a block using curly
braces, you see a parse exception like the following:
org.webmacro.parser.ParseException: Encountered EOF, expecting #end at 76.9
This would happen, for example, if I mistakenly type "is" instead of "at" in
the following:
#setblock $b is macro { "$a" }
This is because when the parser sees an unrecognized token, "is" in this
case, it assumes an implicit #begin and starts looking for a matching #end.
It would be more clear if the message read something like:
org.webmacro.parser.ParseException: Unclosed block beginning at line 44.
Check for invalid directive options or missing a #end.
Is this doable?
It might be better to avoid this issue completely by only assuming an
implicit #begin after a newline. So,
#setblock $b $a #end
would produce an error like:
ParseException: Encountered $a, expecting { or newline.
The alternatives:
#setblock $b {$a}
or
#setblock $b
$a
#end
would be still work fine.
Personally I think this would be a big improvement in the WMScript syntax.
Keats

Re: [Webmacro-devel] Documentation should be a high priority for v2..

[Lane S. <la...@op...>]

Hi Kelvin,

This is excellent feedback and what we need. I good kick in the ass to 
upgrade our docs. First, we need to nail down the release set for 2.0 
and then someone needs to complete a wiki to pdf gateway. Where the 
source wiki pages would be really first rate and grouped as you have 
suggested.

-Lane

Kelvin Westlake wrote:

> Hi All,
>  
> I work a Hosting company and get many requests from customers asking 
> for help developing with many different frameworks, fortunately for me 
> its company policy not to offer support for 3rd party tools,  so 90% 
> of the time I get off easily.  Why am I telling you this? simple,  if 
> documentation isn't thorough or doesn't cater for new users, then 
> they'll scratch they're heads for a few minutes and decide the 
> learning curve is to great and look for an alternative.
>  
> I've recently been charged with creating a portal for internal use, so 
> over the past few weeks I've been testing lots of frameworks (for 
> different uses including db, orm and templating) and I can honestly 
> tell you that all of them have very bad-boring-uninteresting 
> documentation.  Some of them read like a Phd thesis, while others 
> simply fail to realise what documentation should consist of.
>  
> In my eyes documentation primarily consist of 3 parts - Tutorial, 
> Reference, API documentation.. and please don't take this personally, 
> but webmacro's fails miserably on the first 2 points. For example you 
> have a "Quick Start" on the site,  this should contain simple 
> instructions for downloading, setting up and creating a first 
> example.. Instead it contains DownloadWebMacro (which should be 
> seperate) and AllDirectives and NoServlet, what is this supposed to 
> mean to a new user?  then where is the reference documentation giving 
> information on different classes and how they can best be used? there 
> is some information available, but its needs to be hunted down (such 
> information *needs* to be easily accessible).
>  
> If a user can't quickly understand WebMacros or see how different 
> aspects of it are meant to work then they will quickly grow frustrated 
> and go elsewhere, please don't take any of this personally, this is 
> just me playing devils advocate.  Fortunately I detest tag libraries 
> and struts, so I chose Webmacros for the project :) (also Ibatis for 
> the DB stuff, which has some fantastic documentation 
> http://heanet.dl.sourceforge.net/sourceforge/ibatisdb/ibatis_db_guide-1-2-5_beta3.pdf).
>  

Re: [Webmacro-devel] Another parser exception issue

[Lane S. <la...@op...>]

Do we need some javacc work to do this?

-Lane

Kea...@di... wrote:

>
> When you specify an invalid directive option before a block using 
> curly braces, you see a parse exception like the following:
>
> org.webmacro.parser.ParseException: Encountered EOF, expecting #end at 
> 76.9
>
> This would happen, for example, if I mistakenly type "is" instead of 
> "at" in the following:
>
> #setblock $b is macro { "$a" }
>
> This is because when the parser sees an unrecognized token, "is" in 
> this case, it assumes an implicit #begin and starts looking for a 
> matching #end.
>
> It would be more clear if the message read something like:
>
> org.webmacro.parser.ParseException: Unclosed block beginning at line 
> 44.  Check for invalid directive options or missing a #end.
>
> Is this doable?
>
> It might be better to avoid this issue completely by only assuming an 
> implicit #begin after a newline.  So,
>
> #setblock $b $a #end
>
> would produce an error like:
>
> ParseException: Encountered $a, expecting { or newline.
>
> The alternatives:
>
> #setblock $b {$a}
>
> or
>
> #setblock $b
> $a
> #end
>
> would be still work fine.
>
> Personally I think this would be a big improvement in the WMScript 
> syntax.
>
> Keats

Re: [Webmacro-devel] Assigning tasks to people

[Lane S. <la...@op...>]

Marc Palmer wrote:

>
> OK, for us to get organised enough to get a 2.0 release together with 
> the new features we agree, we'll need to:
>
> 1. We need to collate requests from the wm-user list and decide/vote 
> what we will achieve for 2.0
>
> 2. We need to put all these things into an SF tracker and assign them 
> to people who are willing to commit to doing the work.
>
> Sound good? 

I've been making this request now for about 30 days. C'mon you reticent 
WM devvies. Time to push the rock of hard knocks another turn.

-Lane

>
>
>

Re: [Webmacro-devel] Documentation should be a high priority for v2..

[Marc P. <ma...@an...>]

On Thu, 12 Jun 2003 01:04:27 +0100, Kelvin Westlake <ke...@lu...> 
wrote:

> Hi All,
>
> I work a Hosting company and get many requests from customers asking for 
> help developing with many different frameworks, fortunately for me its 
> company policy not to offer support for 3rd party tools,  so 90% of the 
> time I get off easily.  Why am I telling you this? simple,  if 
> documentation isn't thorough or doesn't cater for new users, then they'll 
> scratch they're heads for a few minutes and decide the learning curve is 
> to great and look for an alternative.
>
> I've recently been charged with creating a portal for internal use, so 
> over the past few weeks I've been testing lots of frameworks (for 
> different uses including db, orm and templating) and I can honestly tell 
> you that all of them have very bad-boring-uninteresting documentation.  
> Some of them read like a Phd thesis, while others simply fail to realise 
> what documentation should consist of.

I really hear you on this Kelvin!

This is of course mainly because:

1. Documenting software is boring
2. Documenting software well is extremely time-consuming (detracting from 
programming-hours)
3. Documenting software well requires constant maintenance, difficult when 
other programmers are making changes to open-source code and don't all have 
the skills to write good docs.

[snip]
> If a user can't quickly understand WebMacros or see how different aspects 
> of it are meant to work then they will quickly grow frustrated and go 
> elsewhere, please don't take any of this personally, this is just me 
> playing devils advocate.  Fortunately I detest tag libraries and struts, 
> so I chose Webmacros for the project :) (also Ibatis for the DB stuff, 
> which has some fantastic documentation 
> http://heanet.dl.sourceforge.net/sourceforge/ibatisdb/ibatis_db_guide-1- 
> 2-5_beta3.pdf).

I completely agree. The Wiki site is good because it means any of us can 
get up there and document stuff quickly, but it means that the site is 
somewhat disorganised, gets out of date easily, and in many places lacking 
in detail and "authority" - where somebody kindly wrote some docs without 
necessarily understanding all the details, or the target audience for the 
docs.

This is the key problem with WM - there are two audiences. Template writers 
and programmers. Sometimes the reader is both, but that's not what WM is 
supposed to be about! "Let your artists design... Let your programmers 
code" is the mantra on the site but sadly the docs don't reflect this in my 
view.

On frameworks (and docs) - I am desperately trying to get my WM 
application/framework released today but am bogged-down fixing javadoc 
errors and proof-reading the end-user documentation. I have split the 
documentation into that for designers, and that for programmers.

What I've done so far is pretty good I think, but it's a long way from as 
good as it could be - due to time constraints. I will be interested to see 
what you think of it, as you have experience of docs for other frameworks. 
I for example, could not find out in 5 minutes what Struts actually really 
does. That's plain ridiculous! Sure the answers are there but you have to 
read volumes to get anywhere.

We definitely need to totally overhaul the WM Wiki docs for 2.0 - with 
sections for template writers, and sections for programmers. Template 
writers should never be referred to API javadocs on the site, and there 
should be concise plain english descriptions and examples for everything 
(i.e. the context tools etc.)

Best wishes,
Marc

-- 
Marc Palmer
Contract Java Consultant/Developer
Currently available for hire!
http://www.anyware.co.uk/marc/
http://www.wangjammers.org

Re: [Webmacro-devel] Another small 2.0 thing

[Eric B. R. <eb...@tc...>]

On Friday, June 6, 2003, at 02:40  PM, Eric B. Ridge wrote:

> Brian is absolutely correct.  All of this mess is pure religion.
>
> Since I get to rule the world, The Church of Eric hereby decrees that 
> all WebMacro source code must be formatted according the style 
> guidelines set forth by the Apache Jakarta group.  See 
> http://jakarta.apache.org/turbine/common/code-standards.html for the 
> guidelines.
>
> For what it's worth, this is NOT my personal style, so I feel your 
> pain.
>
> I'll try to get all the code updated in CVS before we start moving 
> packages around... hopefully Sunday.

Sunday, Wednesday, what's the difference?

The changes have been made, and imports have been "optimized" using 
IDEA's default "import optimization" settings.  Enjoy!

eric

Re: [Webmacro-devel] Changelog?

[Eric B. R. <eb...@tc...>]

On Wednesday, June 11, 2003, at 09:25  AM, Marc Palmer wrote:

>
> Guys,
>
> Is it possible to use SF CVS/RSS or something to get an automated 
> changelog page going on www.webmacro.org? We really need this to keep 
> people up to date with what's happening.

subscribe to the -cvs mailing list.  :)

cvs2cl.pl works great, but I don't want to keep a changelog in cvs (and 
expect all the committers to use it).

And I really don't want to open up the Wiki code right now... I just 
don't have the time.

eric

Re: [Webmacro-devel] Another parser exception issue

[Eric B. R. <eb...@tc...>]

On Wednesday, June 11, 2003, at 07:34  PM, Kea...@di...=20
wrote:
> It would be more clear if the message read something like:
>
> org.webmacro.parser.ParseException: Unclosed block beginning at line=20=

> 44. =A0Check for invalid directive options or missing a #end.

Okay, this is committed into CVS.  I'm about ready to screw around with=20=

changing the code style too...

Also, do you use MS OutlookWeb for email?  You've got some funky=20
unicode (or something) characters between "line 44."  and "Check". =20
Looks like spaces, but they're actually:  \xc2\xa0

I've seen this happen a lot with people that use OutlookWeb, that's why=20=

I asked.  I have no idea if there's a fix for it, other than to=20
convince it you want to only send text/plain messages... if OWA even=20
supports that option.

eric=

[Webmacro-devel] Documentation should be a high priority for v2..

[Kelvin W. <ke...@lu...>]

Hi All,

I work a Hosting company and get many requests from customers asking for =
help developing with many different frameworks, fortunately for me its =
company policy not to offer support for 3rd party tools,  so 90% of the =
time I get off easily.  Why am I telling you this? simple,  if =
documentation isn't thorough or doesn't cater for new users, then =
they'll scratch they're heads for a few minutes and decide the learning =
curve is to great and look for an alternative.

I've recently been charged with creating a portal for internal use, so =
over the past few weeks I've been testing lots of frameworks (for =
different uses including db, orm and templating) and I can honestly tell =
you that all of them have very bad-boring-uninteresting documentation.  =
Some of them read like a Phd thesis, while others simply fail to realise =
what documentation should consist of.=20

In my eyes documentation primarily consist of 3 parts - Tutorial, =
Reference, API documentation.. and please don't take this personally, =
but webmacro's fails miserably on the first 2 points. For example you =
have a "Quick Start" on the site,  this should contain simple =
instructions for downloading, setting up and creating a first example.. =
Instead it contains DownloadWebMacro (which should be seperate) and =
AllDirectives and NoServlet, what is this supposed to mean to a new =
user?  then where is the reference documentation giving information on =
different classes and how they can best be used? there is some =
information available, but its needs to be hunted down (such information =
*needs* to be easily accessible).

If a user can't quickly understand WebMacros or see how different =
aspects of it are meant to work then they will quickly grow frustrated =
and go elsewhere, please don't take any of this personally, this is just =
me playing devils advocate.  Fortunately I detest tag libraries and =
struts, so I chose Webmacros for the project :) (also Ibatis for the DB =
stuff, which has some fantastic documentation =
http://heanet.dl.sourceforge.net/sourceforge/ibatisdb/ibatis_db_guide-1-2=
-5_beta3.pdf).

Re: [Webmacro-devel] Another parser exception issue

[Eric B. R. <eb...@tc...>]

On Wednesday, June 11, 2003, at 07:34  PM, Kea...@di...=20
wrote:

> <snip>

> It would be more clear if the message read something like:
>
> org.webmacro.parser.ParseException: Unclosed block beginning at line=20=

> 44. =A0Check for invalid directive options or missing a #end.
>
> Is this doable?

absolutely!  I'm the one that added this message in the first place=20
(was as the result of some bug report about the template actually=20
parsing "correctly").

I'll take care of this one.

> It might be better to avoid this issue completely by only assuming an=20=

> implicit #begin after a newline. =A0So,

I don't know if this is doable or not.  It is a good idea tho.  I'll=20
look into it.  No promises.  :)

eric

>
> #setblock $b $a #end
>
> would produce an error like:
>
> ParseException: Encountered $a, expecting { or newline.
>
> The alternatives:
>
> #setblock $b {$a}
>
> or
>
> #setblock $b
> $a
> #end
>
> would be still work fine.
>
> Personally I think this would be a big improvement in the WMScript=20
> syntax.
>
> Keats

[Webmacro-devel] Another parser exception issue

[<Kea...@di...>]

When you specify an invalid directive option before a block using curly 
braces, you see a parse exception like the following:

org.webmacro.parser.ParseException: Encountered EOF, expecting #end at 
76.9 

This would happen, for example, if I mistakenly type "is" instead of "at" 
in the following:

#setblock $b is macro { "$a" }

This is because when the parser sees an unrecognized token, "is" in this 
case, it assumes an implicit #begin and starts looking for a matching 
#end.

It would be more clear if the message read something like:

org.webmacro.parser.ParseException: Unclosed block beginning at line 44. 
Check for invalid directive options or missing a #end.

Is this doable?

It might be better to avoid this issue completely by only assuming an 
implicit #begin after a newline.  So,

#setblock $b $a #end

would produce an error like: 

ParseException: Encountered $a, expecting { or newline.

The alternatives:

#setblock $b {$a}

or 

#setblock $b
$a
#end

would be still work fine.

Personally I think this would be a big improvement in the WMScript syntax.

Keats

[Webmacro-devel] ContextTool changes

[<Kea...@di...>]

Here are the diffs, in case anyone's interested.  I cleaned up 
internalGet() a bit by adding doFunctionCall() and getTool().

Keats 

Index: Context.java
===================================================================
RCS file: /cvsroot/webmacro/webmacro/src/org/webmacro/Context.java,v
retrieving revision 1.60
diff -r1.60 Context.java
95,96c95,96
<       if (loadTools)
<          loadTools("ContextTools");
---
> //      if (loadTools)
> //         loadTools("ContextTools");
244a245,261
> 
>    private Object doFunctionCall(FunctionCall fc) throws 
PropertyException {
>       Object ret = null;
>       String fname = fc.getName();
>       MethodWrapper func = (MethodWrapper)_funcs.get(fname);
>       if (func == null){
>          func = _broker.getFunction(fname);
>       }
>       if (func != null){
>          Object[] args = fc.getArguments(this);
>          ret = func.invoke(args);
>       }
>       else {
>          _log.error("Function " + fname + " was not loaded!");
>       }
>       return ret;
>    }
245a263,277
>    private Object getTool(String name){
>       ContextTool ct = (ContextTool) _tools.get(name); 
>       Object ret = null;
>       if (ct != null) {
>          try {
>             ret = ct.init(this);
>             put(name,ret);
>             _initializedTools.put(ct,ret);
>          }
>          catch (PropertyException e) {
>             _log.error("Unable to initialize ContextTool: " + name, e);
>          }
>       }
>       return ret; 
>    }
257,282c289,290
<          Object tool = _tools.get(name);
<          if (tool != null) {
<             try {
<                ContextTool ct = (ContextTool) tool;
<                ret = ct.init(this);
<                put(name,ret);
<                _initializedTools.put(ct,ret);
<             }
<                               catch (PropertyException e) {
<                _log.error("Unable to initialize ContextTool: " + name, 
e);
<             }
<          }
<          else if (name instanceof FunctionCall){
<             FunctionCall fc = (FunctionCall)name;
<             String fname = fc.getName();
<             MethodWrapper func = (MethodWrapper)_funcs.get(fname);
<             if (func == null){
<                func = _broker.getFunction(fname);
<             }
<             if (func != null){
<                Object[] args = fc.getArguments(this);
<                ret = func.invoke(args);
<             }
<             else {
<                _log.error("Function " + fname + " was not loaded!");
<             }
---
>          if (name instanceof FunctionCall){
>              return doFunctionCall((FunctionCall)name);
285,289c293,300
<                               // changed by Keats 30-Nov-01
<             return UNDEF;
<                       }
<          //throw new
<          // PropertyException.NoSuchVariableException(name.toString());
---
>              ret = getTool((String)name); 
>              if (ret == null && _tools.isEmpty()){ 
>                 // if tools haven't been loaded, load the tools and try 
again
>                 loadTools("ContextTools");
>                 ret = getTool((String)name); 
>                 if (ret == null) ret = UNDEF;
>              }
>          }






Kea...@di...
Sent by: web...@li...
06/11/2003 02:24 PM

 
        To:     Brian Goetz <br...@qu...>
        cc:     Marc Palmer <ma...@an...>, web...@li...
        Subject:        Re: [Webmacro-devel] If anyone has a few spare minutes...



This sounds like a good idea to me, and quite simple to implement.   I can 
do this if you like.  (In fact I've already made the changes locally. Just 
need to test.) 

There are some other ContextTool improvements I'd like to make, some of 
which I've discussed before. 

- generic ContextTool wrappers to allow classes to be configured as tools 
w/o conforming to the interface.   
- static context tools that don't need to be instantiated per request. 
- configuration of context tools outside of WebMacro.properties, possibly 
in a template, that would allow adding tools without downing the 
container. 

Keats 





Brian Goetz <br...@qu...> 
Sent by: web...@li... 
06/11/2003 12:38 PM 
        
        To:        Marc Palmer <ma...@an...> 
        cc:        web...@li... 
        Subject:        Re: [Webmacro-devel] If anyone has a few spare 
minutes...



> Hmmm, I thought that was the whole point of ContextTool? What does it do 
at 
> the moment that is "lazy"?

There's some per-request overhead (I think its a map.clone()) to put
the context tools in the context.  I'd rather see this happen when,
say, the first failed context resolution occurs.





-------------------------------------------------------
This SF.NET email is sponsored by: eBay
Great deals on office technology -- on eBay now! Click here:
http://adfarm.mediaplex.com/ad/ck/711-11697-6916-5
_______________________________________________
Webmacro-devel mailing list
Web...@li...
https://lists.sourceforge.net/lists/listinfo/webmacro-devel

Re: [Webmacro-devel] If anyone has a few spare minutes...

[<Kea...@di...>]

This sounds like a good idea to me, and quite simple to implement.   I can 
do this if you like.  (In fact I've already made the changes locally. Just 
need to test.)

There are some other ContextTool improvements I'd like to make, some of 
which I've discussed before.

- generic ContextTool wrappers to allow classes to be configured as tools 
w/o conforming to the interface. 
- static context tools that don't need to be instantiated per request.
- configuration of context tools outside of WebMacro.properties, possibly 
in a template, that would allow adding tools without downing the 
container.

Keats 






Brian Goetz <br...@qu...>
Sent by: web...@li...
06/11/2003 12:38 PM

 
        To:     Marc Palmer <ma...@an...>
        cc:     web...@li...
        Subject:        Re: [Webmacro-devel] If anyone has a few spare minutes...


> Hmmm, I thought that was the whole point of ContextTool? What does it do 
at 
> the moment that is "lazy"?

There's some per-request overhead (I think its a map.clone()) to put
the context tools in the context.  I'd rather see this happen when,
say, the first failed context resolution occurs.





-------------------------------------------------------
This SF.NET email is sponsored by: eBay
Great deals on office technology -- on eBay now! Click here:
http://adfarm.mediaplex.com/ad/ck/711-11697-6916-5
_______________________________________________
Webmacro-devel mailing list
Web...@li...
https://lists.sourceforge.net/lists/listinfo/webmacro-devel

Re: [Webmacro-devel] If anyone has a few spare minutes...

[Marc P. <ma...@an...>]

On Wed, 11 Jun 2003 09:38:02 -0700, Brian Goetz <br...@qu...> wrote:

>> Hmmm, I thought that was the whole point of ContextTool? What does it do 
>> at the moment that is "lazy"?
>
> There's some per-request overhead (I think its a map.clone()) to put
> the context tools in the context.  I'd rather see this happen when,
> say, the first failed context resolution occurs.

Yes, that sounds really good to me. I'll put it on the NextRelease page.


-- 
Marc Palmer
Contract Java Consultant/Developer
Currently available for hire!
http://www.anyware.co.uk/marc/
http://www.wangjammers.org

Re: [Webmacro-devel] If anyone has a few spare minutes...

[Brian G. <br...@qu...>]

> Hmmm, I thought that was the whole point of ContextTool? What does it do at 
> the moment that is "lazy"?

There's some per-request overhead (I think its a map.clone()) to put
the context tools in the context.  I'd rather see this happen when,
say, the first failed context resolution occurs.

Re: [Webmacro-devel] If anyone has a few spare minutes...

[Marc P. <ma...@an...>]

On Wed, 11 Jun 2003 09:08:34 -0700, Brian Goetz <br...@qu...> wrote:

>> It would be great if someone could provide some actual user-land info 
>> about the various ContextTools we supply with WM.
>
> I'd also like to see more lazy loading of Context tools, as cloning
> the context tools map is an expensive operation and part of every
> request, whether the user is going to use them or not.

Hmmm, I thought that was the whole point of ContextTool? What does it do at 
the moment that is "lazy"?

"The first time the ContextTool is referenced on a template its init() 
method will be called. Whatever object is returned by init() will be placed 
into the Context exactly as if the servlet programmer had done a 
Context.put(ContextTool.init(context)) prior to invoking the template."
 - http://www.webmacro.org/ContextTool


-- 
Marc Palmer
Contract Java Consultant/Developer
Currently available for hire!
http://www.anyware.co.uk/marc/
http://www.wangjammers.org