Custom Query (1030 matches)
Results (271 - 273 of 1030)
| Ticket | Resolution | Summary | Owner | Reporter |
|---|---|---|---|---|
| #412 | fixed | another case where simple arithmetic is beyond our grasp. | ||
| Description |
I'm too lazy to submit another bug report, but (on x8664): ? (defun foo ()
(let* ((x most-negative-fixnum)
(y 1))
(- x y)))
FOO
? (foo)
1152921504606846975
?
Hmm. I expected something more ... negative. (This is broken on 1.2 as well. I haven't checked ppc64, but I'd be suspicious.) |
|||
| #421 | fixed | Hemlock doesn't understand #| ... |# | ||
| Description |
For example: if the insertion point is before the form (defun foo (n) #| This isn't ready and contains an extra ) |# ) C-M-f should move to the end of the form, not to the commented-out close paren. |
|||
| #473 | fixed | ARGLIST-FROM-MAP confused by supplied-p variables | ||
| Description |
[I have a fix for this and will close this ticket as soon as I check it in; if I missed some case, we can re-open the ticket.] In the absence of better information, CCL:ARGLIST tries to construct a function's arglist from the function's symbol map (and lfun-bits.) Entries in the map generally appear in the reverse of the order in which variables come into scope, and supplied-p variables for optional args generally come into scope before subsequent optional args do; ARGLIST can't assume that ordered entries in the symbol map correspond to the names of arguments that can be passed by the caller unless we add some additional information or take some other steps to ensure this. (The "shape" of the argument list has been correct, but getting the names wrong can be confusing, especially since backtrace uses the same mechanism to identify arguments and other values.) |
|||
