important filepro announcement

[Fairlight]

This public service announcement was brought to you by Jay Ashworth:
> On Thu, Jul 19, 2007 at 10:29:43AM -0400, Nancy Palmquist wrote:
> 
> > When I get results that are unexpected or have to look into a behavior 
> > that is unexplained, I do admit my frustration is sometimes high.
> 
> It seems worth analyzing why this is, because it happens to me as well,
> and it seems disproportionately high compared to some other environments.

The debugger is really not great for things that have a lot of processing.
If it has more than 30 lines, it's less than stellar.  Either you keep it
on and have to cycle through a lot, or you turn it on and off in processing
(thus having to alter processing to do that), and the debugging feels very
disjointed.

Worse still, the debugger won't give you the value of a variable or field
until the line -after- you've assigned it, which can get confusing besides
being annoying.  I've always disliked that.

Then there are the fine error codes and messages that don't actually
indicate the correct thing that's gone wrong and send you on a merry chase
for hours until you eventually get back to square one and figure it out
from scratch.

Personally, rather than relying on fP to tell me where the error is or what
the error is, I find it more expedient to note behaviourally how far it
must have gotten and start inserting writeline statements.  Some people use
LOGTEXT.  It's the equivalent of, "printf() is your friend".  You get all
the info in one place, and you get what you asked for, without the
cumbersome debugging interface.

Y'know, with two changes, they could make the debugger so much more
pleasant.  One, let it report the assigned/modified value on the same line
as the processing that does so, rather than the line after.  Two, put a
separate keystroke in to step forward rather than having to bother with
breaking out of expression checking to move forward, which I find clumsy
and awkward.  Those two things alone would make it more useful.  If I had
to add a third, I'd say give us the ability to set breakpoints when
debugging starts, so you don't have to go through and comment/uncomment a
bunch of debug on/off lines, you just specify the lines -in the debugger-
that you want to trigger on/off.

My  $0.02.

We won't even get into fpcgi debugging.  :-/

mark->
-- 
The latest synth mixdown...
http://media.fairlite.com/Isolation_Voiceless_Cry_Mix.mp3