Estimated hours taken: 8
Branches: main
library/*.m:
Bring these modules up to date with our current style guidelines.
Use predmode declarations where appropriate. Use state variable syntax
where appropriate. Reorder arguments where this makes it possible to
to use state variable syntax. Standardize format of predicate
description comments. Standardize indentation.
Estimated hours taken: 10
Branches: main
library/*.m:
Bring these modules up to date with our current style guidelines.
Use predmode declarations where appropriate. Use state variable syntax
where appropriate. Reorder arguments where this makes it possible to
to use state variable syntax.
library/io.m:
Export some predicates that are duplicated in term_io.m and/or
parser.m, and give them more expressive names.
library/parser.m:
library/term_io.m:
Delete the now unnecessary copies of those predicates in these modules.
Estimated hours taken: 1
Branches: main
More code cleanups for the integer module. This does not change
any algorithms.
library/integer.m:
Use distinct variable names. e.g X and Y rather than X1 and X2
Fix indentation so it adheres more closely to what is specified
in the coding standards.
While converting a list of digits to string call error/1 if something
goes wrong rather than appending "Woops" to the string.
Replace calls to int__to_float/2 with calls to float__float/1.
Fix an incorrect reference in an error message.
Estimated hours taken: 1.5
Branches: main
Add some functions to the integer and rational modules.
Clean up the code in these modules so it adheres more closely to
our current coding standards.
NEWS:
Mention these changes.
library/integer.m:
Shift the constant functions integer.zero/0 and integer.one/0
back into the interface. According to the log they were deleted
because they were not sufficiently useful. In my view they
the symbolic names are preferable to writing them as integer(1)
and integer(0).
Add a function version of integer.pow/3.
Shift a large comment that concerns implementation details and
possible improvements to the module from the interface into the
implementation.
(This avoids it being included in the library reference manual).
Various code cleanups, in particular fix spacing and
indentation throughout.
library/rational.m:
Add function rational/1 that converts an int to a rational and
function from_integer/1 that does the same with an integer.
For consistency with the above add a function from_integers/2 and
deprecate the version rational_from_integers/2.
Remove the functions izero and ione from the implementation and
just call integer.zero/0 and integer.one/0 now that they are
exported from the integer module again.
Various code cleanups.
Estimated hours taken: 0.2
Branches: main, release
library/mer_std.m:
library/rational.m:
library/integer.m:
These modules were unintentionally licensed under the
GPL rather than the LGPL.
Estimated hours taken: 0.1
library/int.m:
library/integer.m:
Remove functions `int:^/2' and `integer:^/2', since
the operator `^' is now used for record syntax.
compiler/code_util.m:
Remove `int:^/2' from the list of builtins.
NEWS:
Document record syntax and the removal of the above functions.
Estimated hours taken: 2
(+ unknown time from Dan Hazel)
Incorporate Dan Hazel's rewrite of library/integer.m into the standard
library, with the following changes:
- I added determinism declarations for a couple of functions,
so that it compiles with `--no-infer-all'.
- I deleted the functions `integer__zero', `integer__one', ...,
`integer__ten' and `integer__intmax' from the interface,
since I don't think they are sufficiently useful.
- I changed the messages in the calls to error/1 to be
a bit more consistent
The log message below is Dan Hazel's original description of his
changes revised a little by me.
library/integer.m:
integer.m was a bit sleepy to be of effective use so I've butchered
it. Sincere apologies to author aet (I'm not sure who you are.)
I've followed some of the suggested "Possible improvements:" in the
preface and rewritten all of the code using a different data
structure.
i) Previously stored as a "sign" and list of integers of base 10000
ordered least significant to most.
My problems with this are:
It doesn't sort into an intuitive order using mercury's compare
(for solutions output for example).
It makes div/mod too slow.
The benefits -- simpler algorithms for `+' `*' -- are not cost
benefits (which I think should be the priority in a library)
as aet is walking the full list to normalise anyway.
** Now stored as a "sign" which also stores the list length and
a list of all +ve or all -ve digits of base 2^14 ordered most
significant to least.
The conversion to base 10 string is more complex of course.
It's no longer quite arbitrary precision - something like
(2^14)^maxint
ii) Normalisation was being done as a separate pass and too often.
This was the most expensive part if I remember.
** Almost all normalisation is now done on the fly (you're down
there anyway).
iii) The new data structure reduces comparison to mercury term
comparison which seems much faster.
iv) I've added
/\ \/ << >> ^ \
v) I've added:
:- func integer__float(integer) = float.
which will clumsily overflow I imagine (fix if you care); and
:- func integer__int(integer) = int.
which checks for overflow and calls error/1 in that case.
vi) Normalising can now be done using machine arithmetic.
vii) There's a new pow algorithm which works by shifting the exponent.
viii)
The procedure/function names should be recognisable but
much of the code is completely rewritten. Sorry if it's
clumsy and ugly.
Estimated hours taken: 0.5
library/integer.m:
Fix a bug in the definition of `big_plus': it wasn't handling the
cases where the sign was zero correctly. I added some additional
code to handle those cases.
Also a few cosmetic changes to make the layout more consistent.
Estimated hours taken: 5
[This change was by Ralph Becket.
I'm just the person who reviewed it and committed it. -fjh.]
Add functions for the single output det predicates in a number
of modules in the standard library. Basically, for each
:- pred f(in, ..., in, out) is det.
I have added the declaration
:- func f(in, ..., in) = out.
and definition
f(X1, ..., Xn) = Y :-
f(X1, ..., Xn, Y).
library/char.m:
library/dir.m:
library/map.m:
library/string.m:
library/list.m:
library/set.m:
Make the changes described above.
library/array.m:
As above, except array input modes are all array_ui or
array_di as appropriate and array output modes are array_uo.
library/int.m:
Added forward versions of +/2, */2 and -/2 as plus/2, times/2
and minus/2 respectively, to make it easier to pass these
as arguments to higher-order predicates.
Also added func constants for max_int, min_int and bits_per_int.
library/integer.m:
Replaced local functions for list head, tail and length with
calls to equivalent functions now defined in list.m.
library/io.m:
Added func for error_message/2.
library/list.m:
Add functions det_head/1 and det_tail/1 which abort on null lists.
library/set.m:
Add functions map/2, filter_map/2 and fold/3.
library/std_util.m:
Added utility function to construct a pair object from its
arguments and general purpose higher order functions for
partial functions and for function composition, exponentiation
and exchanging the arguments of a binary function.
Estimated hours taken: 13
library/integer.m:
Implementation of an arbitrary precision integer type and
operations on it.
library/rational.m:
Implementation of arbitrary precision rational numbers.
library/library.m:
Necessary changes for adding the above two modules.
NEWS:
Noted the addition of the above two modules.
