dammit

[Richard Gaskin]

xbury.cs at clearstream.com wrote:
 > On 08.06.2004 07:27:18 metacard-bounces wrote:
 >
 >>...I can't get a day's work done with the current state of
 >> the error messages coming from the v2.6 engine.
...
 >>Am I the only one who finds this a show-stopper?
 >>Have y'all worked out a solution and forgot to share? :)
 >>
 >>How does the Rev IDE cope with it?
 >
 > what is the problem?

I was hoping to be able to provide a URL to the Bugzilla report that has 
all the details, but alas I can't figure out how to get my old reports 
there.  I suppose the upside is that it may mean it's fixed, but without 
being able to track my reports I can't know.

I brief, if you look at the raw data passed with the errorDialog message 
you'll find a great many erroneous entries, usually with error offsets 
of 0 and no errorObject.

However, simply deleting all such references will not provide a reliable 
workaround, as there are times when valid error will have that pattern.

The symptom is simple enough:  something causes a script execution error 
but the error message as displayed in the Execution Error dialog has no 
possible relationship to the actual issue (usually either "Expected a 
Boolean" or "Not an integer" in cases where no Booleans or integers 
should come into play).

This can be pinned down to the engine with a simple test:  generate a 
script error in MC IDE 2.6b with the v2.6 engine, and repeat the same 
error with the same IDE using the v2.5.1 engine.  In v2.5.1 the error 
message makes sense, in v2.6 it does not.

Adding to the mystery, in most cases where a mismatched error message is 
presented, all you need to do to get the correct error message to 
display is to repeat the actions that led to the error -- the second 
time you should get a reasonable message.

I've compared the error message text strings in the MC IDE 2.6b and RR 
2.2 and they are identical, so we can rule out a repeat of the debacle 
from years ago in which the error offsets changed between builds.

Given all this, one of two things is true:

- All Rev developers using v2.2 are similarly affected,
   but are too shy to complain.

- RunRev has implemented a workaround for the bug in the
   v2.2 IDE, so it does not affect users of that IDE.

I just did a simple test in Rev 2.2:
1. Create a new stack
2. Add a button
3. Set the script of the button to:
     on mouseup
        dfgsdgf
     end mouseup
4. Click the button

Result:
the error message reports: "value is not a boolean (true or false)"
It should of course be "can't find handler".

So it appears to be equally show-stopping for all Transcript users.

I'll take a poll on the use-rev list to see how many there have noticed 
that they aren't able to know what's happening with their scripts.....

-- 
  Richard Gaskin
  Fourth World Media Corporation
  ___________________________________________________________
  Ambassador at FourthWorld.com       http://www.FourthWorld.com