Blair,
My current theory on the unusual Access index naming I was experiencing is
that the offending tables are likely the ones that I have modified
programmatically (via ODBC) over time. It could be a past or current defect
in my SQL generation, or it could be a symptom of one of many differences
between various SQL dialects. To track it down, I have expanded one of my
tests, and started exercising the debugger in an effort to get the tests to
pass.
The noteworth part is that a breakpoint set inside of a couple of nested
blocks is not selected properly. In particular, the selection looks like
this:
s[elf halt.]
with the brackets representing the start and stop of the selection. In
short, the $s in self is not selected (of course it should be). The period
at the end of the expression _is_ selected - your call as to whether that is
a feature or a defect, though it is inconsistent with a quick test in a
workspace using Debug it:
1 to:10 do:[ :x |
10 to:20 do:[ :y |
self halt.
x + y.
].
].
In the above, I find that the period is not selected.
Mikael recently reported something similar, but IIRC that turned out to be a
bug in SmaCC. In this case, the offending code compiles. The image is at
PL4 and I _think_ the new compiler and VM are in use; I am fairly certain of
the new compiler being in use because I was reliably seeing block tempories
in the debugger.
I have a screen shot of the debugger just in case that would be helpful -
available on request, of course. I also have a backup of the image. The
problem persists at least under very minor modifications to the method
source.
Have a good one,
Bill
--
Wilhelm K. Schwab, Ph.D.
[hidden email]
Locked
