For long time were mishandling nested function. The patch
below modifies the compiler to signal error when nested
functions are encountered. With the patch the compiler
discovered a few cases where we used nested functions.
Generated Lisp code looks bogus (AXIOMsys should crash when
executing it). Strange thing is that ATM I am not able to
produce a testcase: the code seem to work as expected.
I really do not know _how_ current code works: attempts
to trace current code either gave no effect or infinite
recursion. OTOH the code below gives expected trace.
For testing I used the following example:
Tr := Tree(Integer)
mytree := tree(1, [tree(2, [])$Tr, tree(3, [])$Tr])$Tr
member?(1, mytree)
member?(3, mytree)
member?(4, mytree)
The patch (including removing nested functions from algebra)
is below:
Index: src/algebra/aggcat.spad.pamphlet
===================================================================
--- src/algebra/aggcat.spad.pamphlet (revision 548)
+++ src/algebra/aggcat.spad.pamphlet (working copy)
@@ -526,13 +526,13 @@
-- x
s = t ==
- eq?(s,t) => true
- #s ~= #t => false
--- I'd rather write
--- every?(member?(#1, t), parts s)
--- but the compiler complains
- every?(mem?, parts s) where
- mem?(x: S): Boolean == member?(x, t)
+ eq?(s,t) => true
+ #s ~= #t => false
+ --- I'd rather write
+ --- every?(member?(#1, t), parts s)
+ --- but the compiler complains
+ mem? : S->Boolean := member?(#1, t)
+ every?(mem?, parts s)
remove_!(f:S->Boolean, t:%) ==
for m in parts t repeat if f m then remove_!(m, t)
@@ -693,13 +693,7 @@
cardinality s == #s
construct l == (s := set(); for x in l repeat insert_!(x,s); s)
count(x:S, s:%) == (member?(x, s) => 1; 0)
--- I'd rather write
--- subset?(s, t) == #s < #t and every?(x +-> member?(x, t), parts s)
--- but the compiler complains
- subset?(s, t) ==
- #s >= #t => false
- every?(mem?, parts s) where
- mem?(x: S): Boolean == member?(x, t)
+ subset?(s, t) == #s <= #t and s = intersect(s,t)
coerce(s:%):OutputForm ==
Index: src/algebra/tree.spad.pamphlet
===================================================================
--- src/algebra/tree.spad.pamphlet (revision 548)
+++ src/algebra/tree.spad.pamphlet (working copy)
@@ -117,8 +117,9 @@
node?(t1, t2) ==
t1 = t2 => true
t case empty => false
- any?(inode?, children t2) where
- inode?(t: %): Boolean == node?(t1, t)
+ inode?:% -> Boolean := node?(t1, #1)
+ any?(inode?, children t2)
+
leaf? t ==
t case empty => false
empty? children t
@@ -146,8 +147,9 @@
true
member?(n, t) ==
t case empty => false
- n = value t or any?(mem?, children t) where
- mem?(c: %): Boolean == member?(n, c)
+ mem?: % -> Boolean := member?(n, #1)
+ n = value t or any?(mem?, children t)
+
members t == parts t
parts t == --buggy?
t case empty => empty()
Index: src/interp/define.boot
===================================================================
--- src/interp/define.boot (revision 548)
+++ src/interp/define.boot (working copy)
@@ -47,6 +47,8 @@
(sig:= getSignatureFromMode(lhs,e)) =>
-- here signature of lhs is determined by a previous declaration
compDefine1(['DEF,lhs,[first sig,:rest signature],specialCases,rhs],m,e)
+ $insideCapsuleFunctionIfTrue =>
+ stackAndThrow ['"Nested functions unsupported:", form]
if signature.target=$Category then $insideCategoryIfTrue:= true
--?? following 3 lines seem bogus, BMT 6/23/93
--? if signature.target is ['Mapping,:map] then
--
Waldek Hebisch
[email protected]
--~--~---------~--~----~------------~-------~--~----~
You received this message because you are subscribed to the Google Groups
"FriCAS - computer algebra system" group.
To post to this group, send email to [email protected]
To unsubscribe from this group, send email to
[email protected]
For more options, visit this group at
http://groups.google.com/group/fricas-devel?hl=en
-~----------~----~----~----~------~----~------~--~---
