Hi,
Sorry for this annoying little bug. You can fix it easily by adding
s: none
in the main context, just after the following comment :
;------ Data reading ------
BTW, it's funny to see how our personal coding style can evolve. While
looking at the piece of code you've exposed to show the problem, I was
thinking how would I wrote that today. In case, this can be useful for
someone :
Old code:
read-field: [
(null-flag: false)
read-length s: (either null-flag [field: none]
[field: sys-copy/part s len s: skip s len]) :s
]
New code (how I would write it today):
read-field: [
(null-flag: false)
read-length copy field len skip
(if null-flag [field: none])
]
It's untested, but looks much better, don't you think ? :-)
-DocKimbel.
Hallvard Ystad wrote:
> Dixit Ingo Hohmann (23.51 18.02.2004):
>
>
>>Hi Hallvard,
>>
>>I haven't seen this exact error before, but some that have been alike ...
>>my guess is that somehow s gets set somewhere else in [...]. Did you try to
>>
>> protect 's
>>
>>directly after setting it? If you still get the same error, there must be
>>a bug in now/time/precise
>
>
> Ah, you're right, of course! I didn't think that could be it, because the only thing
> I do in [...] is this:
> insert db ["ANALYZE TABLE reburls"]
> insert db ["ANALYZE TABLE checklist"]
> (with DocKimbel's mysql protocol). Now with protect 's, here's what I get:
> ** Script Error: Word s is protected, cannot modify
> ** Where: read-columns-headers
> ** Near: parse/all/case read-packet port [
> read-field (col/table: field)
> read-field (col/name: field)
> read-nbytes (col/length: len)
> read-nbytes (col/type: decode/type len)
> read-field (
> col/flags: decode/flags to integer! field/1
> col/decimals: to integer! field/2
> )
> ]
>
> And here's 'read-field:
> read-field: [
> (null-flag: false)
> read-length s: (either null-flag [field: none]
> [field: sys-copy/part s len s: skip s len]) :s
> ]
>
>
> I should have thought about this yesterday, when I was very puzzled to find out that
> the problem disappeared when I renamed 's to 'analyze!
>
> Problem solved. Or at least given to someone else (Doc?). Thanks, Ingo.
> HY
>
> Pr�tera censeo Carthaginem esse delendam
>
>
--
To unsubscribe from this list, just send an email to
[EMAIL PROTECTED] with unsubscribe as the subject.
