Skip to main content
A
Anonymous
February 15, 2011
Answered

Best Coding?

  • February 15, 2011
  • 2 replies
  • 896 views

Okay I have a question about using CFC's.
I'm building an app and it contains several modules. Each module has it's own CFC.

Right now I have in the application.cfm the following line:

<cfset Application.clients  = createObject("component","modules.clients.cfcs.client")>

To call any function I can just use the Application.client."function"()

Now these cfc's have been getting bigger and will be getting bigger in the future.

My question is should I keep doing it this way or should I use the <cfinvoke> to call these functions.
I'm worried that the performance will take a hit. I have been looking into the MySQL procedures,

which should help out with the performance, but I'm not sure if keeping it in the application.cfm is the

right thing to do.

Anybody has some thoughts about this?

Thanks

    This topic has been closed for replies.
    Correct answer Owainnorth

    No, I never unload them - once they're there I leave them there until the Application is restarted.

    Putting stuff is not a problem as long as you have RAM - bear in mind however that even 3000 lines of code can compile down to a bare minimum of memory usage, possibly only 10's of kilobytes if that.

    I'd just monitor the RAM usage of CF as time goes on, but bear in mind having it in the App scope you only ever have one copy - if you were to be constantly unloading and reloading CFCs in people's sessions you'd actually end up using *more* memory and running slower.

    If the Application itself gets too big to be manageable, I'd consider splitting it up into separate Applications, rather than separate CFCs.

    As long as your database tables are properly set up and indexed, having years of data is no problem at all. There's no simple way of seeing how much memory a component is taking up, but you can get a rough idea by trying to locate your cfc in the CF cache directory (E:\CFusion\wwwroot\WEB-INF\cfclasses or thereabouts) - that's where the compiled templates are stored.

    O.

    2 replies

    A
    Anonymous
    February 17, 2011

    Okay I understand where everybody is coming from. And yes there are probably some improvements that can be made, but I have to agree with Owain. What is the difference between 10 300 lines of cfc to one 3000 line cfc.

    My question to everybody would be then for an application how many lines of code is reasonable in all the cfcs combined.

    To shed more light on  my app I have for instance two tables for the Housing Coorporations. One with the housing coorporations info and the other with  the objects belonging to this coorporation. These are some basic retrieve functions in the cfc.

    Now looking at the complaint part. I have a table with the complaint, one with actions to that complaint, one with additional info to that complaint, one with pictures for this complaint, one with documents for this complaint.

    Now I could move some of the info to the complaint table, but doing this would mean the minute someone for instance adds a picture I have to add lines of code to the cfc to actually add this info to the complaint table. So to me this basically mean move one part of code out of one cfc to the other cfc, which does not help to much with minimizing the code. I agree that less code is better.

    I have looked at my cfc and I see some parts that could be improved. But I wont be able to bring it back to 1500 lines of code.

    The only way I have seen so far to bring down the lines of code is to create procedures in MySQL. Then the query is in the procedure and not in the cfcs.

    Also to be realistic I have a lot of commenting going on in my cfc so that if I have to look back later I know what I did. (per function 1-2 lines of comments).

    Owainnorth
    Owainnorth
    Inspiring
    February 15, 2011

    Depends what they're doing.

    Personally, I keep all mine in the Application scope. That way when I need to use one it's already been read, parsed, compiled and stored in RAM so it's as fast as can be.

    The only thing I would ever change was if you had a CFC which is simply doing a massive amount of database work - that I'd move to a stored procedure, but would still call it from the Application scope.

    I actually have gone one step further and written myself a ComponentManager which sits in the Application scope and caches my CFCs as and when they're required, rather than having to initialise them all on Application startup. I kinda do:

    <cfset qCustomers = application.ComponentProxy.getManager('Customer').getAllAsQuery() />

    The ComponentProxy has a hashtable of components, when I call my getManager method it checks to see if it's already cached (and caches it if not) then returns a reference to it.

    Overkill, probably. But it works

    Just don't call anything "CLIENT", whatever you do...

    A
    Anonymous
    February 15, 2011

    Owain,

    Thanks for your reply.

    Let me explain more on what's going on. I'm making this app for  companies in europe that their customers can use to register complaints.

    My customers are construction companies, painting companies etc.

    Their customers are like housing coorporations, schools hoa's etc.

    Now for each group (hoa's, schools etc) i'm making a different module within the app.

    Each module has it's own cfc to retrieve information, complaints (and attached to the complaints, actions, documents, pictures etc), add delete and so on.

    Now that I'm done with the first module, I realized that one of the CFC's s like 3000 lines long. (and this is just the beginning)

    I have been looking at my queryies and trying to  optimize them. Still it's a lot of info that's being put in the application scope. If I look into the future their might be in total 7 more modules, and maybe other things added to the app. So it's not just staying with 2-3 cfc of 3000 lines.

    Also this app will hold the information from the customers for several years, which the dbase could get pretty big.

    As I was going over the app code again, I started to worry if I would get stuck at one point, with the way I was doing this and that the performance would get a hit. At the point that I'm at now I would rather change it than later on.

    I like your idea about the ComponentManager, but do you unload the cfc's when your done our do you leave them?

    I've been coding with ColdFusion for a while now but there is a lot I dont know yet.

    For instance how can you find out how much memory your cfc is using?

    Thanks

    ps. Sorry for the bad word choice "client". I used it as an example, not thinking how others would read it.

    Owainnorth
    OwainnorthCorrect answer
    Inspiring
    February 15, 2011

    No, I never unload them - once they're there I leave them there until the Application is restarted.

    Putting stuff is not a problem as long as you have RAM - bear in mind however that even 3000 lines of code can compile down to a bare minimum of memory usage, possibly only 10's of kilobytes if that.

    I'd just monitor the RAM usage of CF as time goes on, but bear in mind having it in the App scope you only ever have one copy - if you were to be constantly unloading and reloading CFCs in people's sessions you'd actually end up using *more* memory and running slower.

    If the Application itself gets too big to be manageable, I'd consider splitting it up into separate Applications, rather than separate CFCs.

    As long as your database tables are properly set up and indexed, having years of data is no problem at all. There's no simple way of seeing how much memory a component is taking up, but you can get a rough idea by trying to locate your cfc in the CF cache directory (E:\CFusion\wwwroot\WEB-INF\cfclasses or thereabouts) - that's where the compiled templates are stored.

    O.

    Terms & ConditionsAccessibility statement
    Using the community Terms of use Privacy Cookie preferences Do not sell or share my personal information ad choicesAdChoices