Hi,
One of the really good things about using a Message Based Schedular
is that you can build debug/trace into the Schedular and thereby
trace specific message and/or objects. This has cut my debugging time
in half! For instance:
You have 10 objects that receive a message "MSGX" and do something,
somewhere in the chain something goes wrong. I can put a trace "MSGX"
and watch the messages being processed, you can then see which of the
10 objects is causing the problem.
All the Best
Dave
On 27 Apr 2006, at 11:32, Francis Nugent Dixon wrote:
Hi from Paris,
To continue after Jaqueline's and Richard's comments.
Interesting subject. Maybe we could ask other people
if they use the debugger ?
Richard Gaskin wrote:
My first reaction is to pick my jaw up off the floor. How
can you not use the debugger much? I *live* in the
debugger! Who can program without a debugger?
You are making more work for yourself, I think.
I used HC every day for more than 10 years, and now
I use Revolution every day. I never used HC debugging,
and I haven't even thought about Revolution debugging !
I don't even know how it works.
I write my programs in bed (light switched off, no paper
and pencil), just do it in my head ! The next day, I code
in my scripts using (now) a very rigid naming convention
for fields, globals, locals, etc. And I'm up and running
in no time. I still draw little flow diagrams for sections
of my coding that may be complicated - that saves a
lot of hassle. Old habits die hard !
One thing I have found very useful. I declare globals
for every local, during testing, and then comment out
the global definitions after testing. They will be available
in the future, if I make any major mods to my scripts.
I use the "do" command often, but in the form :
put xxxxxxxxxxxx into DoIt -- DoIt is a global
do DoIt
Then I can see if I blew it !
I have a header card on my stacks (which I build
from pre-prepared models). I have displayglobals,
displayfields and displaybuttons functions available
in all my stacks, to see what I have defined. I often
code "exit xxx" in my scripts during testing, and comment
them out as I go. My only gripe is that the Variable Watcher
doesn't update in real-time. You have to close it and
open it again, each time. But then, I am just about to
create a button to do this ........... !
I keep all my functions and standard handlers in a
separate stack, and copy them into my new stacks
as required. Of course, they are clearly documented
with version numbers and dates, so I know what
changes/improvements I made, and when. I generate
quite a few scripts (like Export and Import), and move them
into my new stacks, etc, etc. .......
Debugger ? Could it be useful for untidy minds ?
Revolution is SO EASY to code, that it may offer
fertiliser for sloppy coding ......
And I was always against using the computer to find
my errors!
OK - just to be sure, I will invest in some time to
use debugger, just to see if it saves me more time.
I, for one, DON'T write perfect coding.
- Francis
"Nothing should ever be done for the first time !"
_______________________________________________
use-revolution mailing list
use-revolution@lists.runrev.com
Please visit this url to subscribe, unsubscribe and manage your
subscription preferences:
http://lists.runrev.com/mailman/listinfo/use-revolution
_______________________________________________
use-revolution mailing list
use-revolution@lists.runrev.com
Please visit this url to subscribe, unsubscribe and manage your subscription
preferences:
http://lists.runrev.com/mailman/listinfo/use-revolution
