| commit | 7dfdf1224921ab0696e95702c1cdf08203c1bf78 | |
| author | Douglas Katzman <dougk@google.com> | |
| Wed, 28 May 2014 00:28:28 +0000 (27 20:28 -0400) | ||
| committer | Douglas Katzman <dougk@google.com> | |
| Wed, 28 May 2014 01:49:26 +0000 (27 21:49 -0400) | ||
| tree | f1e9d575532459a0579c8c020eb0ff5734746cf5 | treesnapshot (tar.gz zip) |
| parent | 62d5ccb791234b9b2f5b6b7da970c30eb194653a | commitdiff |
Don't return (AND ... (NOT SIMPLE-ARRAY)) from TYPE-OF.
Unparsing a CTYPE allows negation unless the result is for TYPE-OF.
Based on an examination of uses of TYPE-OF within the compiler
it did not seem important to have another mode in which TYPE-OF
would be allowed to return negations for internal reasons.
INVALID-ARRAY-INDEX-ERROR uses TYPE-OF on an array, and it's just
as well to avoid mentioning the non-simpleness for readability,
but could easily be restored with (type-specifier (ctype-of ...))
if so desired. INSPECT might want that behavior back as well.
Fixes lp#1317308.
Unparsing a CTYPE allows negation unless the result is for TYPE-OF.
Based on an examination of uses of TYPE-OF within the compiler
it did not seem important to have another mode in which TYPE-OF
would be allowed to return negations for internal reasons.
INVALID-ARRAY-INDEX-ERROR uses TYPE-OF on an array, and it's just
as well to avoid mentioning the non-simpleness for readability,
but could easily be restored with (type-specifier (ctype-of ...))
if so desired. INSPECT might want that behavior back as well.
Fixes lp#1317308.
| NEWS | diffblobblamehistory | |
| src/code/late-type.lisp | diffblobblamehistory | |
| src/code/pred.lisp | diffblobblamehistory | |
| tests/type.pure.lisp | diffblobblamehistory |