Skip to content
This repository

HTTPS clone URL

Subversion checkout URL

You can clone with HTTPS or Subversion.

Download ZIP
tree: 8145c00d1f
Fetching contributors…

Cannot retrieve contributors at this time

file 113 lines (112 sloc) 6.055 kb
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 61 62 63 64 65 66 67 68 69 70 71 72 73 74 75 76 77 78 79 80 81 82 83 84 85 86 87 88 89 90 91 92 93 94 95 96 97 98 99 100 101 102 103 104 105 106 107 108 109 110 111 112 113
-*- mode: org -*-
* WIP Release 3.21: unofficial maintenance release of 3.20.
*** Branch "master"
*** DONE ANSI cleanups
*** WIP Complete manual
***** DONE All exported symbols documented.
***** WIP Cleanups
******* DONE Some docstrings talk of EXPRESSIONS, some of FORMS.
******* DONE "Yields" vs "returns" a value.
******* TODO Restricts vs constrains.
******* TODO Don't talk about noticers: they're an implementation detail.
******* TODO Simplify the language in the constraint section.
*** TODO Make iscream.el work under Slime.
*** Obvious opportunistic bugfixes (minimal behavioural changes)
* Future Development
*** Once 3.21 is out, move it onto "classic" branch, into deep maintenance mode.
*** New stuff appears for now on "wip" branch, which is rebased constantly.
    It is master + new code on top.
    Once 3.21 is out, wip becomes the new master.
*** New stuff is also on "dev" branch, which is not rebased -- for convenience.
    It gets master merged it and wip as cherries.
    Will be thrown away once 3.21 is out.
*** Release 4.0
***** DONE /=V allowing integer bounds to be constrained.
***** DONE Make BEST-VALUE make some sense.
***** Support at least some of the currently missing special forms
***** Add DEFINE-SCREAMER-FUNCTION or similar as an alias to SCREAMER::DEFUN.
***** Identify some performance issues and see if anything can be done.
***** Try to improve debuggability of Screamer generated code.
***** Change COUNT-FAILURES to make the number easier to access.
***** Consider adding a separate type for gaussian integers.
* Performance
*** TODO Rewrite dynamic-extent continuations using FLET
    At least SBCL cannot currently stack-allocate a let-bound lambda,
    but manages FLET and LABELS just fine.
*** TODO Dynamic-extent in CPS-CONVERT-TAGBODY.
* Maintenance
*** WIP Split the system into logical pieces.
    Most of the DEFUN-COMPILE-TIME functions should probably go
    to a file of their own, loaded before the main screamer code
    is compiled.
*** TODO Cleanups
***** TODO #:DUMMY-N gensyms used all over the place
      This makes macro-expansions harder to read than they need to be.
      Use differently named gensyms for different things.
***** TODO DEFMACRO-COMPILE-TIME
      Check if Allegro still needs this.
*** TODO primordial.lisp has some commented-out tests using LOOP
    Apparently LOCAL and LOOP interactions depend on the specific
    expansion of loop. :/ If nothing else, documentation should
    mention this. An alternative would be to provide our own
    implemntation of LOOP that is well-behaved when it comes to LOCAL.
*** TODO review the remaining limitations listed in the old source, and put the in the manual
     1. Does not handle SETF methods with multiple values for LOCAL
        SETF.
     2. If you do a (SETF (SYMBOL-FUNCTION 'FOO) ...) to a nondeterministic
function you will lose when you attempt to evaluate (FOO ...).
     3. If you do a (SETF (SYMBOL-FUNCTION 'FOO) ...) to a deterministic
function when FOO was previously defined as a nondeterministic
function you will lose when you attempt to evaluate (FOO ...).
     4. The function record table will not work if we ever support FLET and
LABELS and in particular, if we ever support FLET and LABELS of
nondeterministic functions.
     5. There is no way to force Screamer into compiling a deterministic
function as a nondeterministic one. A wizard might want to do this to
take advantage of the fact that a LOCAL SETF/SETQ in a nondeterministic
function does not cons up closures.
* Bugs
*** TODO LOCAL only reliable with explicit SETF/SETQ:
    If (INCF (FOO)) expands into a form that does SET-FOO instead of
    (SETF (FOO) ...) form, the code walker will not know about the
    side-effect.

    Options:
    - Implement SCREAMER:INCF, etc.
    - Wrap all non-SETF/SETQ forms in GLOBAL, so that those side-effects are never undone.
    - Leave it as-is.
* Old "Bugs to fix" list from screamer.lisp
  These haven't been verified or anything, just moved here.
*** TODO Deterministic/nondeterministic LOCAL SETF/SETQ undone out of order.
*** TODO Changing determinism status
    can cause code from a different file to be included causing wierd behavior.
*** TODO May loop when expressions contain circular lists.
*** TODO APPLY-NONDETERMINISTIC conses.
* Old "TTMTTD" list from screamer.lisp
  Things to Maybe Think Then Do? *shrug*
*** TODO Should have way of having a stream of values.
*** TODO Kashket's constraint additions Fall90-158.
*** TODO Compress trail after repeated LOCAL SETF/SETQ to same variable
*** TODO LOCAL SETF/SETQ on symbol can use special variable binding stack instead of unwind-protect.
*** TODO (F (G (H (EITHER X Y)))) where F, G and H are deterministic can
    CPS-CONVERT to (FUNCALL #'(LAMBDA (Z) (F (G (H Z)))) (EITHER X Y)).
*** TODO Should give warning when it recompiles functions due to changing determinism status.
*** TODO =V <V <=V >V and >=V should do congruence/monotone closure.
*** TODO =V should propagate domains.
*** TODO BEST-VALUE
*** TODO Should cache VARIABLE-LOWER-BOUND/VARIABLE-UPPER-BOUND for domain variables.
*** TODO Faster version of BIND! which doesn't cons.
*** TODO Get DIAGNOSIS and MONTAGUE to work.
*** TODO Get GROW-UP and NONLINEAR examples to work.
*** TODO FUNCALLV and APPLYV need to assert the domains of the variable that they return.
*** TODO Check that +V, -V, *V, /V, MINV and MAXV work and do the right thing with type propagation.
*** TODO Check that PROPAGATE does the right thing with bounds of integers.
*** TODO MEMBERV and derivatives should support vectors.
*** TODO Backtracking out of INTEGER-BETWEENV and the like will yield an
    unconstrained variable since the constraints are asserted locally.
* Refactoring
*** Depend on Alexandria
***** peal-off-documentation-string-and-declarations -> alexandria:parse-body
*** SETF and SETQ walking is suspiciously different
* Future Work
*** TODO See if it would make sense to make nondeterministic functions funcallable instances.
*** TODO Lock package on SBCL.
Something went wrong with that request. Please try again.