There's a several-year-old discussion of this issue here:
http://wiki.tcl.tk/19627
It looks like 'impedance mis-match' is an appropriate term for this
and that the tcl/sqlite type relationship is problematic, due to the
typelessness of tcl and the omission in the api of a way to indicate
the type of the substitution. I'm going to use something other than
tcl for what I'm doing.
On Sun, Jan 26, 2014 at 10:37 AM, Donald Allen <donaldcal...@gmail.com> wrote:
> This script
>
> #!/usr/bin/env tclsh
>
> package require sqlite3
>
> set x 1
>
> sqlite3 db /tmp/foo
>
> db eval {select (2 > :x) as foo} {
> puts "foo was $foo"
> }
>
> run on an up-to-date Arch Linux system produces
>
> foo was 0
>
> obviously incorrect. There seems to be an issue with variable
> substitution here. Changing the line in question to
>
> db eval {select (2 > 1) as foo} {
>
> and re-running produces
>
> foo was 1
>
> obviously correct.
>
> The sqlite documentation on the tcl api says that variable
> substitution can occur "in any position where it is legal to put a
> string or number literal". There is no information in the
> documentation about how to control this, so the api has to be deciding
> for itself whether a string or numeric literal is appropriate and I
> think it is getting it wrong here.
>
> db eval {select (2 > cast (:x as integer)) as foo} {
>
> works correctly.
>
> db eval {select (2 > '1') as foo} {
>
> produces
>
> foo was 0
>
> Looks like a bug to me, but perhaps I'm missing something. Comments?
>
> /Don Allen
_______________________________________________
sqlite-users mailing list
sqlite-users@sqlite.org
http://sqlite.org:8080/cgi-bin/mailman/listinfo/sqlite-users
