Perl 6
perl6 tasks: Revision 23
This page used to contain Rakudo-specific items. That content is now "here."<http://wiki.github.com/rakudo/rakudo/tasks>
Now this page contains Perl6-generic TODO items. That means, mostly, places where the established Synopsis could be made more explicit, and larger tasks involving the DRAFT synopsis.
^^ Established Synopsis (non-draft) TODO
* all spec: .iterator versus .Iterator
* S02 "radix literals"<http://perlcabal.org/syn/S02.html#line_2435> elaborate "default radix" behavior.
** is the :16(0b0) :16(0d0) peculiar to :16, or generic behavior if number is legal in the "default" radix e.g. :12(0b0)?
** does the statement about "default radix" apply only to :16(...) or also :16<...>?
** maybe worth elaborating what :N[0xdeadbeef] would do under "decimal" restriction
*** ...or relaxing that restriction
*** ...and perhaps making clear whether :N[4,$x] is legal and thus it's functional like :N(...)
* S03 zip and "parallel array traversal"<http://perlcabal.org/syn/S03.html#Traversing_arrays_in_parallel>
** see S07 section below WRT .eoi
** Probably no reason why it should not take single Iterators as well as lists and DWIM
*** that is zip((1,2,3);$iterator) should probably interleave (1,2,3) with $iterator.get results
*** but still treat zip((1,2,3); ($iterator,"foo","bar")) as (1,$iterator;2,"foo";3,"bar)
** The intent is probably that zip stops before calling .get on any Iterator for this pass when any .eoi != False
*** This needs to be clear in spec
*** This requires peekability of both .eoi == True and .eoi == Inf
*** Behavior when .eoi == undef ("may terminate, cannot tell") needs to be explicit
*** Or we could drop the specification that infinite lists are detected/aborted
**** we allow other things to be infinite, why not zip?
** Is the Iterator or the zip/operator code responsible for implementing (...,*) repeating last element?
* S04 "gather/take"<http://perlcabal.org/syn/S04.html#The_gather_statement_prefix>
** Define behavior of nested gather/takes
*** Is there caller.take and context(...).take "as per .leave"<http://perlcabal.org/syn/S04.html#Control_Exceptions>
*** Can you label gathers and use LABEL.take
* S09 "Compact Structures"<http://perlcabal.org/syn/S09.html#Compact_structs>
** Define packing behavior of class A is int8 { has $.b int8 }; my @a of A;
** Define packing behavior of class A is B is C ....
*** Define packing behavior in the "Diamond Problem"<http://en.wikipedia.org/wiki/Diamond_problem> case ("see also"<http://perlcabal.org/syn/S14.html#line_59>)
** Define packing/data storage behavior for extra attributes added with run-time class composition
* S12 Disambiguate second sentence in "line 69"<http://perlcabal.org/syn/S12.html#line_69>
* S12 Clear up "declaration's scope" in "line 82"<http://perlcabal.org/syn/S12.html#line_82> (e.g. not the has/our in decl)
* S12 Can := be used in attribute pseudo-assignment?
^^ DRAFT Synopsis
* (in progress) split up S32
* S32/* review prototypes esp WRT return types
** lots of Int/Str where maybe should be Any
** perhaps some missing return type variants e.g. List push/unshift (for feeds)
** needs attention from someone who has followed p6l closely
* S07 carefully decide whether generic item iterator "flattens" like the generic array iterator
* S07 Should probably implement .eoi to peek Iterator
* S07 Need way to test for Iterators known to never normally terminate (for zip)
** suggest .eoi be allowed to return Inf to indicate this
Now this page contains Perl6-generic TODO items. That means, mostly, places where the established Synopsis could be made more explicit, and larger tasks involving the DRAFT synopsis.
^^ Established Synopsis (non-draft) TODO
* all spec: .iterator versus .Iterator
* S02 "radix literals"<http://perlcabal.org/syn/S02.html#line_2435> elaborate "default radix" behavior.
** is the :16(0b0) :16(0d0) peculiar to :16, or generic behavior if number is legal in the "default" radix e.g. :12(0b0)?
** does the statement about "default radix" apply only to :16(...) or also :16<...>?
** maybe worth elaborating what :N[0xdeadbeef] would do under "decimal" restriction
*** ...or relaxing that restriction
*** ...and perhaps making clear whether :N[4,$x] is legal and thus it's functional like :N(...)
* S03 zip and "parallel array traversal"<http://perlcabal.org/syn/S03.html#Traversing_arrays_in_parallel>
** see S07 section below WRT .eoi
** Probably no reason why it should not take single Iterators as well as lists and DWIM
*** that is zip((1,2,3);$iterator) should probably interleave (1,2,3) with $iterator.get results
*** but still treat zip((1,2,3); ($iterator,"foo","bar")) as (1,$iterator;2,"foo";3,"bar)
** The intent is probably that zip stops before calling .get on any Iterator for this pass when any .eoi != False
*** This needs to be clear in spec
*** This requires peekability of both .eoi == True and .eoi == Inf
*** Behavior when .eoi == undef ("may terminate, cannot tell") needs to be explicit
*** Or we could drop the specification that infinite lists are detected/aborted
**** we allow other things to be infinite, why not zip?
** Is the Iterator or the zip/operator code responsible for implementing (...,*) repeating last element?
* S04 "gather/take"<http://perlcabal.org/syn/S04.html#The_gather_statement_prefix>
** Define behavior of nested gather/takes
*** Is there caller.take and context(...).take "as per .leave"<http://perlcabal.org/syn/S04.html#Control_Exceptions>
*** Can you label gathers and use LABEL.take
* S09 "Compact Structures"<http://perlcabal.org/syn/S09.html#Compact_structs>
** Define packing behavior of class A is int8 { has $.b int8 }; my @a of A;
** Define packing behavior of class A is B is C ....
*** Define packing behavior in the "Diamond Problem"<http://en.wikipedia.org/wiki/Diamond_problem> case ("see also"<http://perlcabal.org/syn/S14.html#line_59>)
** Define packing/data storage behavior for extra attributes added with run-time class composition
* S12 Disambiguate second sentence in "line 69"<http://perlcabal.org/syn/S12.html#line_69>
* S12 Clear up "declaration's scope" in "line 82"<http://perlcabal.org/syn/S12.html#line_82> (e.g. not the has/our in decl)
* S12 Can := be used in attribute pseudo-assignment?
^^ DRAFT Synopsis
* (in progress) split up S32
* S32/* review prototypes esp WRT return types
** lots of Int/Str where maybe should be Any
** perhaps some missing return type variants e.g. List push/unshift (for feeds)
** needs attention from someone who has followed p6l closely
* S07 carefully decide whether generic item iterator "flattens" like the generic array iterator
* S07 Should probably implement .eoi to peek Iterator
* S07 Need way to test for Iterators known to never normally terminate (for zip)
** suggest .eoi be allowed to return Inf to indicate this
