Switching to CFMX

[Rob Rohan]

Hi everyone,

We are moving to cold fusion MX and are running into some snags. We use
fusebox on a rather large application, and are finding that quite a few
areas of our site throw errors like the following:

cfindex2ecfm1137626797 (Code of a method longer than 65535 bytes)

After doing some research it turns out that java can't have methods > 64k -
sounds like a familiar limitation - can anyone say DOS :) - but I am
somewhat confused as to how cfm gets made into .class files.

* Does cfinclude copy the contents of every included file then compile -
regardless of the switch statement?

* Does the compiled cfm make a whole class, is it only a method in some
master class, or is everything stuck into a single method? is there a way to
force other methods, like how cfcase only runs the necessary code?



We tried to split the index.cfm into index0.cfm and index1.cfm then include
based on another switch; however, that resulted in:

 (class: cfindex02ecfm1443745593, method: _factor1 signature:
(Ljavax/servlet/jsp/tagext/Tag;)Ljava/lang/Object;) Illegal target of jump
or branch
  ColdFusion cannot determine the line of the template that caused this
error. This is often caused by an error in the exception handling subsystem.

This seems like a serious problem with CFMX for a large application using
fusebox so I hope someone has experienced this and has a elegant workaround.

Any help / comments are appreciated


Cheers,
Rob Rohan
Senior Applications Engineer
Regweb, LLC

FAQ: http://www.thenetprofits.co.uk/coldfusion/faq
Archives: http://www.mail-archive.com/cf-talk@houseoffusion.com/
Unsubscribe: http://www.houseoffusion.com/index.cfm?sidebar=lists

RE: Switching to CFMX

[Benjamin S. Rogers]

> * Does cfinclude copy the contents of every included file 
> then compile - regardless of the switch statement?

I believe includes still happen dynamically. You can verify this by
deleting all the files in the following directory:

  \CFusionMX\wwwroot\WEB-INF\cfclasses

Restart the ColdFusion service and make a request to your page. It
should only create class files for the files that were needed to
complete the request.

BTW, I just checked and my developer version behaves this way.

> * Does the compiled cfm make a whole class, is it only a 
> method in some master class, or is everything stuck into a 
> single method? is there a way to force other methods, like 
> how cfcase only runs the necessary code?

It appears that each ColdFusion page (*.cfm), ColdFusion component
(*.cfc) and function gets its own class file. For a look at what's going
on behind the scenes, edit the web.xml file in the following directory:

  \CFusionMX\wwwroot\WEB-INF\

Change the "coldfusion.compiler.saveJava" parameter to "true." This will
create .java source files for each of the classes. You may be able to
get a better idea of what is going on from there. Again, you'll probably
have to restart the ColdFusion service.

> This seems like a serious problem with CFMX for a large 
> application using fusebox so I hope someone has experienced 
> this and has a elegant workaround.

I haven't run into any problems specific to porting Fusebox style
applications to ColdFusion MX, except the fact that the first request is
always interminably slow. I don't have any in production, However. I've
only ported several such applications to a Developer Version of
ColdFusion MX for testing purposes.

Benjamin S. Rogers
http://www.c4.net/
v.508.240.0051
f.508.240.0057

__
Get the mailserver that powers this list at http://www.coolfusion.com
FAQ: http://www.thenetprofits.co.uk/coldfusion/faq
Archives: http://www.mail-archive.com/cf-talk@houseoffusion.com/
Unsubscribe: http://www.houseoffusion.com/index.cfm?sidebar=lists

RE: Switching to CFMX

[Rob Rohan]

Thanks Benjamin,

I appreciate your response, and I'll check out the xml file - that's a great
tip.

You will run into this when doing large files or working on a large
application
http://webforums.macromedia.com/coldfusion/messageview.cfm?catid=3&threadid=
385775

The index file I am dealing with is > 2500 with a huge case statement that
includes / module calls a lot of files.

here is the problem:
when you write


balrg

...


that gets created into
public class whatever{
=>  public whatever(){
switch(barg){
case blarg:
}
...
=>  }
}

The parts marked between => can be no greater than 64k. One has no way of
knowing how large that is when writing a cfm page because the instruction
swapping is hidden from the programmer. It seems to happen at about 1500 or
so lines.

The 64k limit is a Java limitation.

The *only* work around I found is to break the large file into smaller files
(in my case index0.cfm, index1.cfm, index2.cfm, etc), and have a master
index with a fuse action list that include the proper index file.








...


It adds a whole layer of abstraction, but seems to work. Again this only
happen when doing really large files and the (evil) errors MX gives are the
following (for the next poor sap searching the archives):

 (class: cfindex12ecfm1447028664, method: _factor1 signature:
(Ljavax/servlet/jsp/tagext/Tag;)Ljava/lang/Object;) Illegal target of jump
or branch
  ColdFusion cannot determine the line of the template that caused this
error. This is often caused by an error in the exception handling subsystem.

OR

Code of a method longer than 65535 bytes

Thanks again
Rob


-Original Message-
From: Benjamin S. Rogers [mailto:[EMAIL PROTECTED]]
Sent: Monday, September 30, 2002 2:12 PM
To: CF-Talk
Subject: RE: Switching to CFMX


> * Does cfinclude copy the contents of every included file
> then compile - regardless of the switch statement?

I believe includes still happen dynamically. You can verify this by
deleting all the files in the following directory:

  \CFusionMX\wwwroot\WEB-INF\cfclasses

Restart the ColdFusion service and make a request to your page. It
should only create class files for the files that were needed to
complete the request.

BTW, I just checked and my developer version behaves this way.

> * Does the compiled cfm make a whole class, is it only a
> method in some master class, or is everything stuck into a
> single method? is there a way to force other methods, like
> how cfcase only runs the necessary code?

It appears that each ColdFusion page (*.cfm), ColdFusion component
(*.cfc) and function gets its own class file. For a look at what's going
on behind the scenes, edit the web.xml file in the following directory:

  \CFusionMX\wwwroot\WEB-INF\

Change the "coldfusion.compiler.saveJava" parameter to "true." This will
create .java source files for each of the classes. You may be able to
get a better idea of what is going on from there. Again, you'll probably
have to restart the ColdFusion service.

> This seems like a serious problem with CFMX for a large
> application using fusebox so I hope someone has experienced
> this and has a elegant workaround.

I haven't run into any problems specific to porting Fusebox style
applications to ColdFusion MX, except the fact that the first request is
always interminably slow. I don't have any in production, However. I've
only ported several such applications to a Developer Version of
ColdFusion MX for testing purposes.

Benjamin S. Rogers
http://www.c4.net/
v.508.240.0051
f.508.240.0057


__
Signup for the Fusion Authority news alert and keep up with the latest news in 
ColdFusion and related topics. http://www.fusionauthority.com/signup.cfm
FAQ: http://www.thenetprofits.co.uk/coldfusion/faq
Archives: http://www.mail-archive.com/cf-talk@houseoffusion.com/
Unsubscribe: http://www.houseoffusion.com/index.cfm?sidebar=lists

RE: Switching to CFMX

[Rob Rohan]

(a have to agree not my code)

-Original Message-
From: Benjamin S. Rogers [mailto:[EMAIL PROTECTED]]
Sent: Monday, September 30, 2002 3:32 PM
To: CF-Talk
Subject: RE: Switching to CFMX


Good to know. That explains why I haven't run into it as I tend to nest
fuses (I don't actually use Fusebox but something akin to it). I try to
keep individual switch statements to between 5 and 10 case statements
with no more than 20. Gets to hard to read otherwise.

Benjamin S. Rogers
http://www.c4.net/
v.508.240.0051
f.508.240.0057

-Original Message-
From: Rob Rohan [mailto:[EMAIL PROTECTED]]
Sent: Monday, September 30, 2002 5:45 PM
To: CF-Talk
Subject: RE: Switching to CFMX


Thanks Benjamin,

I appreciate your response, and I'll check out the xml file - that's a
great
tip.

You will run into this when doing large files or working on a large
application
http://webforums.macromedia.com/coldfusion/messageview.cfm?catid=3&threa
did=
385775

The index file I am dealing with is > 2500 with a huge case statement
that
includes / module calls a lot of files.

here is the problem:
when you write


balrg

...


that gets created into
public class whatever{
=>  public whatever(){
switch(barg){
case blarg:
}
...
=>  }
}

The parts marked between => can be no greater than 64k. One has no way
of
knowing how large that is when writing a cfm page because the
instruction
swapping is hidden from the programmer. It seems to happen at about 1500
or
so lines.

The 64k limit is a Java limitation.

The *only* work around I found is to break the large file into smaller
files
(in my case index0.cfm, index1.cfm, index2.cfm, etc), and have a master
index with a fuse action list that include the proper index file.








...


It adds a whole layer of abstraction, but seems to work. Again this only
happen when doing really large files and the (evil) errors MX gives are
the
following (for the next poor sap searching the archives):

 (class: cfindex12ecfm1447028664, method: _factor1 signature:
(Ljavax/servlet/jsp/tagext/Tag;)Ljava/lang/Object;) Illegal target of
jump
or branch
  ColdFusion cannot determine the line of the template that caused this
error. This is often caused by an error in the exception handling
subsystem.

OR

Code of a method longer than 65535 bytes

Thanks again
Rob


-Original Message-
From: Benjamin S. Rogers [mailto:[EMAIL PROTECTED]]
Sent: Monday, September 30, 2002 2:12 PM
To: CF-Talk
Subject: RE: Switching to CFMX


> * Does cfinclude copy the contents of every included file
> then compile - regardless of the switch statement?

I believe includes still happen dynamically. You can verify this by
deleting all the files in the following directory:

  \CFusionMX\wwwroot\WEB-INF\cfclasses

Restart the ColdFusion service and make a request to your page. It
should only create class files for the files that were needed to
complete the request.

BTW, I just checked and my developer version behaves this way.

> * Does the compiled cfm make a whole class, is it only a
> method in some master class, or is everything stuck into a
> single method? is there a way to force other methods, like
> how cfcase only runs the necessary code?

It appears that each ColdFusion page (*.cfm), ColdFusion component
(*.cfc) and function gets its own class file. For a look at what's going
on behind the scenes, edit the web.xml file in the following directory:

  \CFusionMX\wwwroot\WEB-INF\

Change the "coldfusion.compiler.saveJava" parameter to "true." This will
create .java source files for each of the classes. You may be able to
get a better idea of what is going on from there. Again, you'll probably
have to restart the ColdFusion service.

> This seems like a serious problem with CFMX for a large
> application using fusebox so I hope someone has experienced
> this and has a elegant workaround.

I haven't run into any problems specific to porting Fusebox style
applications to ColdFusion MX, except the fact that the first request is
always interminably slow. I don't have any in production, However. I've
only ported several such applications to a Developer Version of
ColdFusion MX for testing purposes.

Benjamin S. Rogers
http://www.c4.net/
v.508.240.0051
f.508.240.0057




__
This list and all House of Fusion resources hosted by CFHosting.com. The place for 
dependable ColdFusion Hosting.
FAQ: http://www.thenetprofits.co.uk/coldfusion/faq
Archives: http://www.mail-archive.com/cf-talk@houseoffusion.com/
Unsubscribe: http://www.houseoffusion.com/index.cfm?sidebar=lists

RE: Switching to CFMX

[Benjamin S. Rogers]

Good to know. That explains why I haven't run into it as I tend to nest
fuses (I don't actually use Fusebox but something akin to it). I try to
keep individual switch statements to between 5 and 10 case statements
with no more than 20. Gets to hard to read otherwise.

Benjamin S. Rogers
http://www.c4.net/
v.508.240.0051
f.508.240.0057

-Original Message-
From: Rob Rohan [mailto:[EMAIL PROTECTED]] 
Sent: Monday, September 30, 2002 5:45 PM
To: CF-Talk
Subject: RE: Switching to CFMX


Thanks Benjamin,

I appreciate your response, and I'll check out the xml file - that's a
great
tip.

You will run into this when doing large files or working on a large
application
http://webforums.macromedia.com/coldfusion/messageview.cfm?catid=3&threa
did=
385775

The index file I am dealing with is > 2500 with a huge case statement
that
includes / module calls a lot of files.

here is the problem:
when you write


balrg

...


that gets created into
public class whatever{
=>  public whatever(){
switch(barg){
case blarg:
}
...
=>  }
}

The parts marked between => can be no greater than 64k. One has no way
of
knowing how large that is when writing a cfm page because the
instruction
swapping is hidden from the programmer. It seems to happen at about 1500
or
so lines.

The 64k limit is a Java limitation.

The *only* work around I found is to break the large file into smaller
files
(in my case index0.cfm, index1.cfm, index2.cfm, etc), and have a master
index with a fuse action list that include the proper index file.








...


It adds a whole layer of abstraction, but seems to work. Again this only
happen when doing really large files and the (evil) errors MX gives are
the
following (for the next poor sap searching the archives):

 (class: cfindex12ecfm1447028664, method: _factor1 signature:
(Ljavax/servlet/jsp/tagext/Tag;)Ljava/lang/Object;) Illegal target of
jump
or branch
  ColdFusion cannot determine the line of the template that caused this
error. This is often caused by an error in the exception handling
subsystem.

OR

Code of a method longer than 65535 bytes

Thanks again
Rob


-Original Message-
From: Benjamin S. Rogers [mailto:[EMAIL PROTECTED]]
Sent: Monday, September 30, 2002 2:12 PM
To: CF-Talk
Subject: RE: Switching to CFMX


> * Does cfinclude copy the contents of every included file
> then compile - regardless of the switch statement?

I believe includes still happen dynamically. You can verify this by
deleting all the files in the following directory:

  \CFusionMX\wwwroot\WEB-INF\cfclasses

Restart the ColdFusion service and make a request to your page. It
should only create class files for the files that were needed to
complete the request.

BTW, I just checked and my developer version behaves this way.

> * Does the compiled cfm make a whole class, is it only a
> method in some master class, or is everything stuck into a
> single method? is there a way to force other methods, like
> how cfcase only runs the necessary code?

It appears that each ColdFusion page (*.cfm), ColdFusion component
(*.cfc) and function gets its own class file. For a look at what's going
on behind the scenes, edit the web.xml file in the following directory:

  \CFusionMX\wwwroot\WEB-INF\

Change the "coldfusion.compiler.saveJava" parameter to "true." This will
create .java source files for each of the classes. You may be able to
get a better idea of what is going on from there. Again, you'll probably
have to restart the ColdFusion service.

> This seems like a serious problem with CFMX for a large
> application using fusebox so I hope someone has experienced
> this and has a elegant workaround.

I haven't run into any problems specific to porting Fusebox style
applications to ColdFusion MX, except the fact that the first request is
always interminably slow. I don't have any in production, However. I've
only ported several such applications to a Developer Version of
ColdFusion MX for testing purposes.

Benjamin S. Rogers
http://www.c4.net/
v.508.240.0051
f.508.240.0057



__
Structure your ColdFusion code with Fusebox. Get the official book at 
http://www.fusionauthority.com/bkinfo.cfm
FAQ: http://www.thenetprofits.co.uk/coldfusion/faq
Archives: http://www.mail-archive.com/cf-talk@houseoffusion.com/
Unsubscribe: http://www.houseoffusion.com/index.cfm?sidebar=lists

Re: Switching to CFMX

[Sean A Corfield]

I've been trying to resist responding to this thread but...

On Monday, Sep 30, 2002, at 14:44 US/Pacific, Rob Rohan wrote:
> 
>delimiters=",">
>   
>   
>   
>   
>   
>   ...
> 

Let me say, first of all, that I'm glad you've found a workaround.

Secondly, this is exactly the sort of issue that makes me very nervous 
of Fusebox. It tries hard to hide complexity from programmers but this 
shows just how much code it's really hiding and just how much machinery 
lurks behind the framework. Of course, this isn't really FB's fault: 
old-style CF just doesn't provide the expressive machinery to support 
the sort of near-OO encapsulation that FB attempts. I think that FB4 
has the potential to be much cleaner and idiomatic, based on the OO 
features that MX introduced to CF. I'm looking forward to seeing it.

> It adds a whole layer of abstraction, but seems to work. Again this 
> only
> happen when doing really large files and the (evil) errors MX gives 
> are the
> following

This is good to know for other Fuseboxers. Thanx for following up with 
a workaround without 'bashing' CF!

An Architect's View -- http://www.corfield.org/blog/

Macromedia DevCon 2002, October 27-30, Orlando, Florida
Architecting a New Internet Experience
Register today at http://www.macromedia.com/go/devcon2002

__
Get the mailserver that powers this list at http://www.coolfusion.com
FAQ: http://www.thenetprofits.co.uk/coldfusion/faq
Archives: http://www.mail-archive.com/cf-talk@houseoffusion.com/
Unsubscribe: http://www.houseoffusion.com/index.cfm?sidebar=lists

RE: Switching to CFMX

[Nat Papovich]

Unfortunately though, Rob's really long index doesn't appear to use Fusebox
3's "core file", which I assume, is the "complex machinery" you mention,
Sean. In pre-FB3, there is very little to no machinery used, so long
cfswitch statements are the result of the poor design of whoever wrote
unfortunate Rob's code, not some implied Fusebox architectural path. But
that's not to say there aren't certain aspects of FB3's "machinery" that I'm
looking forward to optimizing towards OO/CFCs/Java :)

I have encountered this same limitation in CFMX/Java in a really long custom
tag written by (surprise) someone else. Sean, do you know if this (64k)
limitation is something that can be addressed by Macromedia? Or is it
inherent to Java? Or... ?

NAT

> -Original Message-
> From: Sean A Corfield [mailto:[EMAIL PROTECTED]]
> Sent: Monday, September 30, 2002 4:07 PM
> To: CF-Talk
> Subject: Re: Switching to CFMX
>
>
> I've been trying to resist responding to this thread but...
>
> On Monday, Sep 30, 2002, at 14:44 US/Pacific, Rob Rohan wrote:
> > 
> >  > delimiters=",">
> > 
> > 
> > 
> > 
> > 
> > ...
> > 
>
> Let me say, first of all, that I'm glad you've found a workaround.
>
> Secondly, this is exactly the sort of issue that makes me very nervous
> of Fusebox. It tries hard to hide complexity from programmers but this
> shows just how much code it's really hiding and just how much machinery
> lurks behind the framework. Of course, this isn't really FB's fault:
> old-style CF just doesn't provide the expressive machinery to support
> the sort of near-OO encapsulation that FB attempts. I think that FB4
> has the potential to be much cleaner and idiomatic, based on the OO
> features that MX introduced to CF. I'm looking forward to seeing it.
>
> > It adds a whole layer of abstraction, but seems to work. Again this
> > only
> > happen when doing really large files and the (evil) errors MX gives
> > are the
> > following
>
> This is good to know for other Fuseboxers. Thanx for following up with
> a workaround without 'bashing' CF!
>

__
This list and all House of Fusion resources hosted by CFHosting.com. The place for 
dependable ColdFusion Hosting.
FAQ: http://www.thenetprofits.co.uk/coldfusion/faq
Archives: http://www.mail-archive.com/cf-talk@houseoffusion.com/
Unsubscribe: http://www.houseoffusion.com/index.cfm?sidebar=lists

RE: Switching to CFMX

[Benjamin S. Rogers]

> Secondly, this is exactly the sort of issue that
> makes me very nervous of Fusebox. It tries hard
> to hide complexity from programmers but this 
> shows just how much code it's really hiding and
> just how much machinery lurks behind the
> framework.

This is the exact same argument that gets made against ColdFusion. I'm
sure those early Word Perfect assembly programmers said the same thing
about programmers writing word processing programs in languages like C
and *gasp* VisualBasic.

> Of course, this isn't really FB's fault: 

Of course this isn't really
(C's|C++'s|Java's|VisualBasic's|ColdFusion's) fault. :)

Benjamin S. Rogers
http://www.c4.net/
v.508.240.0051
f.508.240.0057

__
This list and all House of Fusion resources hosted by CFHosting.com. The place for 
dependable ColdFusion Hosting.
FAQ: http://www.thenetprofits.co.uk/coldfusion/faq
Archives: http://www.mail-archive.com/cf-talk@houseoffusion.com/
Unsubscribe: http://www.houseoffusion.com/index.cfm?sidebar=lists

Re: Switching to CFMX

[Sean A Corfield]

On Monday, Sep 30, 2002, at 17:55 US/Pacific, Nat Papovich wrote:
> Unfortunately though, Rob's really long index doesn't appear to use 
> Fusebox
> 3's "core file", which I assume, is the "complex machinery" you 
> mention,
> Sean. In pre-FB3, there is very little to no machinery used, so long
> cfswitch statements are the result of the poor design of whoever wrote
> unfortunate Rob's code, not some implied Fusebox architectural path.

Thanx for clarifying. I need to go back and read that FB3 book again! :)

I was thinking more along the lines of the impact all the included 
files had on the code size - the "core files" *are* machinery. However, 
if you're reassuring me that with FB3, the 64k limit is unlikely to be 
hit then that is even better to know - after all, it would not be very 
encouraging if "large Fusebox applications" (whatever they may be - 
we've had discussions before about 'what is a large application?') 
simply failed on CFMX due to code size.

> I have encountered this same limitation in CFMX/Java in a really long 
> custom
> tag written by (surprise) someone else. Sean, do you know if this (64k)
> limitation is something that can be addressed by Macromedia? Or is it
> inherent to Java? Or... ?

It is inherent in Java's virtual machine - the offset on a jump is 16 
bits and that causes a 64k limit. The CFMX compiler tries to work 
around this in some cases by figuring out that it can rearrange code, 
but not all files are amenable to this.

An Architect's View -- http://www.corfield.org/blog/

Macromedia DevCon 2002, October 27-30, Orlando, Florida
Architecting a New Internet Experience
Register today at http://www.macromedia.com/go/devcon2002

__
Get the mailserver that powers this list at http://www.coolfusion.com
FAQ: http://www.thenetprofits.co.uk/coldfusion/faq
Archives: http://www.mail-archive.com/cf-talk@houseoffusion.com/
Unsubscribe: http://www.houseoffusion.com/index.cfm?sidebar=lists

Re: Switching to CFMX

[Sean A Corfield]

On Tuesday, Oct 1, 2002, at 06:49 US/Pacific, Benjamin S. Rogers wrote:
> This is the exact same argument that gets made against ColdFusion.

Not by me.

> I'm
> sure those early Word Perfect assembly programmers said the same thing
> about programmers writing word processing programs in languages like C
> and *gasp* VisualBasic.

Not in my experience. I wrote my first word processing program (a 
template-driven automobile insurance letter editor) in assembly 
language in '81 and it was a pain. I wrote my second word processing 
program (a dual-language, English-Cyrillic, general WP) in C in '84. I, 
like many other assembly programmers, were quite glad of the additional 
expressive power in C - and of course we reassured ourselves that it 
often generated code that was close to what we could do by hand (not 
always of course, but compiler optimizations improved rapidly in the 
80's).

>> Of course, this isn't really FB's fault:
> Of course this isn't really
> (C's|C++'s|Java's|VisualBasic's|ColdFusion's) fault. :)

Well, actually this hidden complexity *is* ColdFusion's fault. Or at 
least, pre-MX ColdFusion's fault. Like C, it was not designed to 
support OO-style frameworks (see the mess that is early X-Windows 
source code). Now we're seeing a problem that *is* Java's fault - it 
was designed with inherent limitations.

An Architect's View -- http://www.corfield.org/blog/

Macromedia DevCon 2002, October 27-30, Orlando, Florida
Architecting a New Internet Experience
Register today at http://www.macromedia.com/go/devcon2002

__
Signup for the Fusion Authority news alert and keep up with the latest news in 
ColdFusion and related topics. http://www.fusionauthority.com/signup.cfm
FAQ: http://www.thenetprofits.co.uk/coldfusion/faq
Archives: http://www.mail-archive.com/cf-talk@houseoffusion.com/
Unsubscribe: http://www.houseoffusion.com/index.cfm?sidebar=lists

Fusebox and CFMX was RE: Switching to CFMX

[Mike Brunt]

Sean, thanks for all your informative and very helpful posts.  I started
using CF in 1996 and Fusebox opened up all sorts of great things to pre CFMX
ColdFusion.  I know there are many luminaries who dislike Fusebox no doubt
for good reason within their own well-structured company's/worlds, people
for whom I have the utmost respect.  For our operation Fusebox made
ColdFusion very understandable fairly quickly for those whom had not used CF
previously.  It also added a level of legitimacy to ColdFusion in many
hallowed halls (not that CF needed legitimacy but perception is powerful).
Like you I look forward to Fuseboxes  next iteration with great
anticipation, I know the great minds behind Fusebox are carving out even
greater things thanks to CFMX.  Things just keep getting more exciting and
bountiful.

Mike Brunt - CTO
Webapper Services LLC
http://www.webapper.com
Downey CA Office
562.243.6255

"Making the NET Work"

-Original Message-
From: Sean A Corfield [mailto:[EMAIL PROTECTED]]
Sent: Monday, September 30, 2002 4:07 PM
To: CF-Talk
Subject: Re: Switching to CFMX


I've been trying to resist responding to this thread but...

On Monday, Sep 30, 2002, at 14:44 US/Pacific, Rob Rohan wrote:
> 
>delimiters=",">
>   
>   
>   
>   
>   
>   ...
> 

Let me say, first of all, that I'm glad you've found a workaround.

Secondly, this is exactly the sort of issue that makes me very nervous
of Fusebox. It tries hard to hide complexity from programmers but this
shows just how much code it's really hiding and just how much machinery
lurks behind the framework. Of course, this isn't really FB's fault:
old-style CF just doesn't provide the expressive machinery to support
the sort of near-OO encapsulation that FB attempts. I think that FB4
has the potential to be much cleaner and idiomatic, based on the OO
features that MX introduced to CF. I'm looking forward to seeing it.

> It adds a whole layer of abstraction, but seems to work. Again this
> only
> happen when doing really large files and the (evil) errors MX gives
> are the
> following

This is good to know for other Fuseboxers. Thanx for following up with
a workaround without 'bashing' CF!

An Architect's View -- http://www.corfield.org/blog/

Macromedia DevCon 2002, October 27-30, Orlando, Florida
Architecting a New Internet Experience
Register today at http://www.macromedia.com/go/devcon2002


__
Get the mailserver that powers this list at http://www.coolfusion.com
FAQ: http://www.thenetprofits.co.uk/coldfusion/faq
Archives: http://www.mail-archive.com/cf-talk@houseoffusion.com/
Unsubscribe: http://www.houseoffusion.com/index.cfm?sidebar=lists

RE: Fusebox and CFMX was RE: Switching to CFMX

[Rob Rohan]

I hope I did not give the impression that I dislike fusebox - because I like
fusebox. I am by default a Java programmer, and fusebox added a tasty OO
flavor to CF. It does appear the MX will now natively do most of what
fusebox set out to do - which is the main reason for our switch.

In conclusion,

Fusebox = good
Java 64k method limit = bad

(JSP and Cocoon suffer from this as well)

:)
Rob


-Original Message-
From: Mike Brunt [mailto:[EMAIL PROTECTED]]
Sent: Monday, September 30, 2002 4:53 PM
To: CF-Talk
Subject: Fusebox and CFMX was RE: Switching to CFMX


Sean, thanks for all your informative and very helpful posts.  I started
using CF in 1996 and Fusebox opened up all sorts of great things to pre CFMX
ColdFusion.  I know there are many luminaries who dislike Fusebox no doubt
for good reason within their own well-structured company's/worlds, people
for whom I have the utmost respect.  For our operation Fusebox made
ColdFusion very understandable fairly quickly for those whom had not used CF
previously.  It also added a level of legitimacy to ColdFusion in many
hallowed halls (not that CF needed legitimacy but perception is powerful).
Like you I look forward to Fuseboxes  next iteration with great
anticipation, I know the great minds behind Fusebox are carving out even
greater things thanks to CFMX.  Things just keep getting more exciting and
bountiful.

Mike Brunt - CTO
Webapper Services LLC
http://www.webapper.com
Downey CA Office
562.243.6255

"Making the NET Work"

-Original Message-
From: Sean A Corfield [mailto:[EMAIL PROTECTED]]
Sent: Monday, September 30, 2002 4:07 PM
To: CF-Talk
Subject: Re: Switching to CFMX


I've been trying to resist responding to this thread but...

On Monday, Sep 30, 2002, at 14:44 US/Pacific, Rob Rohan wrote:
> 
>delimiters=",">
>   
>   
>   
>   
>   
>   ...
> 

Let me say, first of all, that I'm glad you've found a workaround.

Secondly, this is exactly the sort of issue that makes me very nervous
of Fusebox. It tries hard to hide complexity from programmers but this
shows just how much code it's really hiding and just how much machinery
lurks behind the framework. Of course, this isn't really FB's fault:
old-style CF just doesn't provide the expressive machinery to support
the sort of near-OO encapsulation that FB attempts. I think that FB4
has the potential to be much cleaner and idiomatic, based on the OO
features that MX introduced to CF. I'm looking forward to seeing it.

> It adds a whole layer of abstraction, but seems to work. Again this
> only
> happen when doing really large files and the (evil) errors MX gives
> are the
> following

This is good to know for other Fuseboxers. Thanx for following up with
a workaround without 'bashing' CF!

An Architect's View -- http://www.corfield.org/blog/

Macromedia DevCon 2002, October 27-30, Orlando, Florida
Architecting a New Internet Experience
Register today at http://www.macromedia.com/go/devcon2002



__
Get the mailserver that powers this list at http://www.coolfusion.com
FAQ: http://www.thenetprofits.co.uk/coldfusion/faq
Archives: http://www.mail-archive.com/cf-talk@houseoffusion.com/
Unsubscribe: http://www.houseoffusion.com/index.cfm?sidebar=lists