| Ticket UUID: | 1667660 | ||
| Title: | [lsort -integer] not updated for bignums | ||
| Type: | Bug | Created on: | 2007-02-24 05:49:16 |
| Submitter: | dgp | Assigned to: | dgp |
| Subsystem: | 17. Commands I-L | Severity: | |
| Priority: | 5 Medium | Last modified: | 2007-03-10 01:43:02 |
| Status: | Open | Closed by: | |
| Resolution: | None | Closed on: | |
| Version: | obsolete: 8.5a6 | ||
| Description: | ||||
% lsort -integer {1 2 1000000000000000000000000}
integer value too large to represent
| ||||
| User Comments: | ||||
afredd added on 2007-02-26 19:12:48:
Logged In: YES
user_id=1386588
Originator: NO
I don't see this as a bug, as the given string
is not a Tcl integer:
% info pat
8.5a4
% string is integer 1000000000000000000000000
0
...and the docs for lsort say:
# -integer
# Convert list elements to integers
# and use integer comparison.
Another argument for not changing -integer to
support bigint's (and wides > 2**32!), is that
this would slow it down considerably for all users,
most of whom never go near bigints.
However, i do think the example should be possible,
but with a new option that does perform numeric
comparison on *any* of Tcl's number types
eg. [lsort -numeric ...]
Also worth considering is the introduction of a
<=> *operator* that does a numeric comparison like:
proc ::tcl::mathop::<=> {a b} {
if {$a > $b} {
return 1
} elseif {$a < $b} {
return -1
} else {
return 0
}
}
This would be especially useful in writing lsort
comparison routines.
Just my $0.02.
dgp added on 2007-03-10 01:43:02:
Logged In: YES user_id=80530 Originator: YES yes this is all wrapped up in the TIP 297 issues. | ||||
