a weird crash - any suggestions?

Peter Alcibiades palcibiades-first at yahoo.co.uk
Sun Mar 23 11:30:18 EDT 2008

Its Debian Lenny, 2.9 rc  2.

The script is very long, slow and amateurish, basically a brute force
solution to something I could not get to work any other way.  It goes
through and zeroes out the 30+ fields, then it puts the file into a
variable, and uses filter to get the right subset of the file.  It totals
the subset and puts totals into the fields.  It doesn't do anything other
than the most basic commands - add, filter, case, if.  I would not be
thanked for posting two or three pages of this stuff!  But it does work, and
it is accurate as compared to test runs against a spreadsheet using the same
input files.

But what is weird is the following.  You start this up in the ide, both
buttons get you to a report card with fields with white backgrounds, and the
print card script from the button on the report card works fine.  It also
works fine with development tools suspended.  

You launch the same stack from a splash stack, or from StackRunner, and in
both cases, a go to with update results in the report being done correctly,
but  the fields are all greyed out, and there is a crash from printing using
the print button with script given above.  Go there from go without update,
and the bottom leftmost field is blue, the others grey, and printing causes
a crash.  Why going there without update should produce a different result
from going there from a button that updates is mysterious.

Maybe this is connected to the difficulties with the Linux print dialog, so
maybe there is a way around it by disabling this and using Rev's.  But I
don't understand why it should work from ide and not from either stackrunner
or standalone.

I had the thought why not go back to 2.6.1 if there is no printing anyway. 
But you can't.  Save as legacy does put a file with the save as name, but
its not openable in 2.6.1.

Maybe I am the only one having so much trouble with Linux printing?  It is
not a question of lp being set wrongly.  Even after you set lp correctly and
confirm it, and can confirm it by putting the printer name in Rev, the print
dialogue still fails to recognise it.  I just filed this as a bug.

The problem is, you can work around it if printing a field by using the
shell and calling an awk script to process and then pipe the result to
kprinter.  But there seems to be no way to do anything like this with a
card.  And anyway, why a difficulty on just this particular card, not the
other one, with the identical script?

If anyone wants to see this horrible script I'll happily send it privately. 
But I don't see how it can be that if it runs fine from the ide.  Running
out of ideas.


Mark Schonewille-3 wrote:
> Hi Peter,
> Which operating system are you using, which version(s) of Revolution,  
> and what is the exact script of button A?

View this message in context: http://www.nabble.com/a-weird-crash---any-suggestions--tp16232253p16237083.html
Sent from the Revolution - User mailing list archive at Nabble.com.

More information about the Use-livecode mailing list