add isl_union_set_is_params

Signed-off-by: Sven Verdoolaege <skimo@kotnet.org>

isl_point_get_coordinate: check validity of "pos" argument

Signed-off-by: Sven Verdoolaege <skimo@kotnet.org>

interface/extract_interface.cc: adjust to rename of Diagnostic in clang

Reported-by: Tobias Grosser <tobias@grosser.es>
Signed-off-by: Sven Verdoolaege <skimo@kotnet.org>

rename isl_pw_aff_add to isl_pw_aff_union_add

As pointed out by Andreas Kloeckner, isl_pw_aff_add behaves in a way
that is inconsistent with other functions defined over pairs of
isl_pw_affs in that it produces an isl_pw_aff that is defined on
the union of the definition domains of its arguments.
We therefore rename isl_pw_aff_add to isl_pw_aff_union_add and
add a new isl_pw_aff_add that produces a result that is only
defined on the intersection of the definition domains of its arguments.

Signed-off-by: Sven Verdoolaege <skimo@kotnet.org>

reimplement pw_aff_mul on top of a generic isl_pw_*_on_shared_domain

This new function will be useful for implementing other similar functionality.

Signed-off-by: Sven Verdoolaege <skimo@kotnet.org>

isl_map_intersect: require arguments to live in the same space

Before, we would allow one of the arguments to be a parameter domain.
As pointed out by Andreas Kloeckner, this breaks some invariants
in other parts of the code, notably isl_pw_*_intersect_domain.
Since we have an isl_map_intersect_params now, we no longer need
this special case, so we remove it.

Signed-off-by: Sven Verdoolaege <skimo@kotnet.org>

isl_input.c: read_tuple: use isl_map_intersect_params where needed

Signed-off-by: Sven Verdoolaege <skimo@kotnet.org>

Merge branch 'maint'

Insert missing error check into isl_pw_*_intersect_domain

Signed-off-by: Andreas Kloeckner <inform@tiker.net>
Signed-off-by: Sven Verdoolaege <skimo@kotnet.org>

Run isl_basic_map_finalize after isl_{basic_,}map_eliminate

Signed-off-by: Andreas Kloeckner <inform@tiker.net>
Signed-off-by: Sven Verdoolaege <skimo@kotnet.org>

isl_stream_read_map: handle nested parenthesized expressions

Signed-off-by: Sven Verdoolaege <skimo@kotnet.org>

isl_basic_{set,map}_partial_lex{min,max}_pw_multi_aff: normalize isl_affs

Signed-off-by: Sven Verdoolaege <skimo@kotnet.org>

add isl_aff_normalize

Signed-off-by: Sven Verdoolaege <skimo@kotnet.org>

add extract_interface for constructing interfaces to other languages

Currently, we only support the generation of a Python interface.
In particular, a ctypes based interface.

Signed-off-by: Sven Verdoolaege <skimo@kotnet.org>

mark some data structures and functions for export to language bindings

We prefer to only include function in interfaces to other languages
that have been explicitly marked for export.
In particular, we do not want to add functions that are obsolescent
and are only kept for backward compatibility, or functions that
only make sense for a C interface, or functions that are only meant
for other libraries such as barvinok that are tightly coupled to isl.

Signed-off-by: Sven Verdoolaege <skimo@kotnet.org>

add isl_*_to_str

Signed-off-by: Sven Verdoolaege <skimo@kotnet.org>

add isl_union_pw_qpolynomial_read_from_str

Signed-off-by: Sven Verdoolaege <skimo@kotnet.org>

add isl_multi_aff_lift

Signed-off-by: Sven Verdoolaege <skimo@kotnet.org>

add isl_local_space_lifting

Signed-off-by: Sven Verdoolaege <skimo@kotnet.org>

add isl_local_space_lift

Signed-off-by: Sven Verdoolaege <skimo@kotnet.org>

add isl_local_space_intersect

Signed-off-by: Sven Verdoolaege <skimo@kotnet.org>

add isl_multi_aff_read_from_str

Signed-off-by: Sven Verdoolaege <skimo@kotnet.org>

add isl_pw_aff_read_from_str

Signed-off-by: Sven Verdoolaege <skimo@kotnet.org>

add isl_aff_read_from_str

Signed-off-by: Sven Verdoolaege <skimo@kotnet.org>

add isl_pw_multi_aff_read_from_str

Signed-off-by: Sven Verdoolaege <skimo@kotnet.org>

add isl_pw_multi_aff_from_map

Signed-off-by: Sven Verdoolaege <skimo@kotnet.org>

add isl_basic_map_lexmin_pw_multi_aff

Signed-off-by: Sven Verdoolaege <skimo@kotnet.org>

pip: optionally use isl_basic_set_partial_lexmin_pw_multi_aff

Signed-off-by: Sven Verdoolaege <skimo@kotnet.org>

add isl_basic_{set,map}_partial_lex{min,max}_pw_multi_aff

Signed-off-by: Sven Verdoolaege <skimo@kotnet.org>

add isl_pw_multi_aff_substitute

Signed-off-by: Sven Verdoolaege <skimo@kotnet.org>

add isl_set_substitute

Signed-off-by: Sven Verdoolaege <skimo@kotnet.org>

add isl_local_space_substitute

Signed-off-by: Sven Verdoolaege <skimo@kotnet.org>

isl_printer_print_pw_multi_aff: partial support for C output

Signed-off-by: Sven Verdoolaege <skimo@kotnet.org>

add isl_pw_multi_aff

Signed-off-by: Sven Verdoolaege <skimo@kotnet.org>

add isl_pw_*_get_dim_id

Signed-off-by: Sven Verdoolaege <skimo@kotnet.org>

add isl_pw_*_get_dim_name

Signed-off-by: Sven Verdoolaege <skimo@kotnet.org>

add isl_pw_*_has_tuple_id

Signed-off-by: Sven Verdoolaege <skimo@kotnet.org>

add isl_pw_*_set_dim_id

Signed-off-by: Sven Verdoolaege <skimo@kotnet.org>

add isl_pw_*_get_tuple_id

Signed-off-by: Sven Verdoolaege <skimo@kotnet.org>

add isl_pw_*_get_tuple_name

Signed-off-by: Sven Verdoolaege <skimo@kotnet.org>

add isl_basic_map_from_multi_aff

Signed-off-by: Sven Verdoolaege <skimo@kotnet.org>

add isl_multi_aff

An isl_multi_aff object represents a sequence of
zero or more affine expressions, all defined on the same domain space.
As such, they are different from isl_aff_lists as the elements in a list
may have different spaces and no space information is available at all
about empty lists.

Signed-off-by: Sven Verdoolaege <skimo@kotnet.org>

isl_pw_templ.c: rename align_params_pw_pw_and to include PW

This allows isl_pw_templ.c to be included twice from the same file.

Signed-off-by: Sven Verdoolaege <skimo@kotnet.org>

add isl_map_dim_max

Signed-off-by: Sven Verdoolaege <skimo@kotnet.org>

privately export isl_basic_map_foreach_lexopt

Signed-off-by: Sven Verdoolaege <skimo@kotnet.org>

add isl_basic_map_eliminate

Signed-off-by: Sven Verdoolaege <skimo@kotnet.org>

add isl_basic_map_has_dim_id

Signed-off-by: Sven Verdoolaege <skimo@kotnet.org>

Merge branch 'maint'

isl_basic_map_eliminate_vars: gauss result

Signed-off-by: Sven Verdoolaege <skimo@kotnet.org>

add isl_aff_set_dim_id

Signed-off-by: Sven Verdoolaege <skimo@kotnet.org>

add isl_local_space_set_dim_id

Signed-off-by: Sven Verdoolaege <skimo@kotnet.org>

add isl_map_find_dim_by_name

Signed-off-by: Sven Verdoolaege <skimo@kotnet.org>

add isl_space_find_dim_by_name

Signed-off-by: Sven Verdoolaege <skimo@kotnet.org>

export isl_pw_aff_intersect_domain

Signed-off-by: Sven Verdoolaege <skimo@kotnet.org>

export isl_pw_aff_plain_is_equal

Signed-off-by: Sven Verdoolaege <skimo@kotnet.org>

add generic isl_pw_*_set_tuple_id

For consistency, this means we need to change the prototype of
isl_pw_aff_set_tuple_id.

Signed-off-by: Sven Verdoolaege <skimo@kotnet.org>

isl_set_intersect: properly handle intersection with parameter domains

isl_set_intersect currently allows one of its arguments to be a parameter
domain while the other is not.  This will probably change in future, but
while we still support this situation, we should make sure that the result
lives in the same space as that of the argument that is not a parameter domain.

Signed-off-by: Sven Verdoolaege <skimo@kotnet.org>

isl_access_info_compute_flow: align parameters of input may access relations

This part was missing from 13124d3 (isl_access_info_compute_flow: align
parameters of input access relations, Thu Aug 11 14:03:55 2011 +0200).

Signed-off-by: Sven Verdoolaege <skimo@kotnet.org>

Merge branch 'maint'

fix isl_set_involves_dims

It would only consider the set to involve a dim if _all_ its subsets
would involve the dim, while it should consider the set to involve a
dim if _any_ of its subsets do.

Signed-off-by: Sven Verdoolaege <skimo@kotnet.org>

add isl_set_plain_is_fixed

Signed-off-by: Sven Verdoolaege <skimo@kotnet.org>

add isl_set_from_params

Signed-off-by: Sven Verdoolaege <skimo@kotnet.org>

add isl_space_set_from_params

Signed-off-by: Sven Verdoolaege <skimo@kotnet.org>

Merge branch 'maint'

isl_qpolynomial_morph: properly handle denominators in morph

Before, the denominator would get ignored during the morphing of the divs.

Signed-off-by: Sven Verdoolaege <skimo@kotnet.org>

add isl_mat_diag

Signed-off-by: Sven Verdoolaege <skimo@kotnet.org>

isl_tab_pip.c: sol_for_add: correctly set denominator

Signed-off-by: Sven Verdoolaege <skimo@kotnet.org>

isl_tab_pip.c: sol_map_add: fix double free on error path

Signed-off-by: Sven Verdoolaege <skimo@kotnet.org>

isl_stream: accept "==" token

Signed-off-by: Sven Verdoolaege <skimo@kotnet.org>

add isl_union_pw_*_plain_is_equal

Signed-off-by: Sven Verdoolaege <skimo@kotnet.org>

add isl_pw_*_plain_is_equal

Signed-off-by: Sven Verdoolaege <skimo@kotnet.org>

add isl_set_plain_cmp

Signed-off-by: Sven Verdoolaege <skimo@kotnet.org>

add isl_map_params

Signed-off-by: Sven Verdoolaege <skimo@kotnet.org>

isl_stream_next_token: treat "-0" as two tokens '-' and '0'

Usually, a '-' followed by a sequence of digits is treated as a single
(negative) number.  When parsing affine expressions, negative numbers
are allowed without a preceding '+'.  However, if "-0" is parsed as '0'
then there is no way of distinguishing between "-0" and "0" and so "-0"
was not allowed in affine expressions.
We now treat "-0" as two tokens, so that it can be used in affine expressions.
The flip side is that "-0" can no longer be used where a single integer
is expected.

Signed-off-by: Sven Verdoolaege <skimo@kotnet.org>

drop "nparam" argument from isl_{set,map}_read_from_{file,str}

The "nparam" argument is removed because it is inconsistent with
the rest of the API.
The "nparam" argument was only ever useful for reading sets
represented in the original PolyLib format.
For reading maps, it was of limited use because there was no way
of specifying the number of input dimensions.
In both the isl format and the extended PolyLib format, the number
of parameters can be read off from the input.

Signed-off-by: Sven Verdoolaege <skimo@kotnet.org>

doc: document *_move_dims functions

Signed-off-by: Sven Verdoolaege <skimo@kotnet.org>

drop isl_div abstraction

The isl_div abstraction has been superseded by the isl_aff abstraction,
so use that instead.

Signed-off-by: Sven Verdoolaege <skimo@kotnet.org>

isl_{in,}equality_alloc: take isl_local_space intead of isl_space

This makes it easier to construct constraints involving existentially
quantified variables.

Signed-off-by: Sven Verdoolaege <skimo@kotnet.org>

doc: update example code to change in return type of isl_constraint_set_*

The change happened way back in e2a2317 (reimplement isl_constraint in terms
of isl_aff, Fri Jul 1 17:20:52 2011 +0200).

Signed-off-by: Sven Verdoolaege <skimo@kotnet.org>

Merge branch 'maint'

provide full prototypes for some functions with no arguments

Reported-by: Zbigniew Chamski <zbigniew.chamski@gmail.com>
Signed-off-by: Sven Verdoolaege <skimo@kotnet.org>

doc: fix typo

Signed-off-by: Sven Verdoolaege <skimo@kotnet.org>

add isl_set_gist_params

Signed-off-by: Sven Verdoolaege <skimo@kotnet.org>

make isl_pw_* object live in a map space

Before, the space of an isl_pw_* object was that of the cells.
Conceptually, however, they map elements from these cells to something
and so their space should be that of a map.
This is especially useful for isl_pw_multi_aff objects that will
be introduced later.  Since we want to be able to convert back and forth
between such objects and isl_maps, we need to keep track of the
complete map space.

isl_aff, isl_qpolynomial and isl_qpolynomial_fold objects now
also live in a map space, although internally, they still keep
track of the domain space.

Signed-off-by: Sven Verdoolaege <skimo@kotnet.org>

add isl_space_extend_domain_with_range

Signed-off-by: Sven Verdoolaege <skimo@kotnet.org>

add isl_space_is_domain

Signed-off-by: Sven Verdoolaege <skimo@kotnet.org>

add isl_local_space_is_set

Signed-off-by: Sven Verdoolaege <skimo@kotnet.org>

add isl_local_space_domain

Signed-off-by: Sven Verdoolaege <skimo@kotnet.org>

implement isl_constraint on top of isl_vec

In e2a2317 (reimplement isl_constraint in terms of isl_aff,
Fri Jul 1 17:20:52 2011 +0200), were reimplemented on top of isl_aff
because both objects share many features.
There are also some difference, however, and it is therefore better
to implement isl_constraint directly on top of isl_vec.
In particular, an isl_aff has a denominator, whereas an isl_constraint
does not.  Although the denominator of an isl_aff is usually 1, it may
not always be and could therefore lead to subtle bugs.

Moreover, the space of an isl_constraint may be either that of
a set or a map, whereas, in principle, the domain of isl_aff is always
a set.  We allowed the domain of an isl_aff to be a map when it was
used to implement an isl_constraint, but this was not very clean.
This would become even more dramatic when we change the space of an isl_aff
to be that of a map.

Signed-off-by: Sven Verdoolaege <skimo@kotnet.org>

add isl_vec_neg

Signed-off-by: Sven Verdoolaege <skimo@kotnet.org>

isl_aff.c: extract isl_local_space_get_active

This function will also be useful for our new implementation of isl_constraints.

Signed-off-by: Sven Verdoolaege <skimo@kotnet.org>

isl_output.c: drop redundant "set" argument

We used to have to keep track of whether a given isl_space was that
of a map or a set, but now the isl_space itself contains that information
so we no longer need this extra argument.

Signed-off-by: Sven Verdoolaege <skimo@kotnet.org>

isl_union_*_alloc: project space onto parameters

Before, we were expecting the user to provide a parameter space,
but we might as well project whatever the user gives us onto
its parameter space.

Signed-off-by: Sven Verdoolaege <skimo@kotnet.org>

explicitly differentiate between spaces of maps, sets and parameter sets

Before, no distinction was made internally between spaces of maps,
sets and parameter spaces.  In particular, an isl_space with zero
input and output dimensions could have been any of those.
When printing isl_sets and isl_maps, we could keep track of whether
the original object as a set or a map, but parametric domains and
zero-dimensional domains were still printed identically.

Now we explicitly keep track of whether a space is supposed to be
that of a map, set or parametric set.  This means for example that
"[N] -> { [] : N >= 0 }" and "[N] -> { : N >= 0 }" are no longer considered
to be identical.
This may break some code that was using one space in place of another,
especially since until recently there was no official way of creating
a parameter space.

The main advantage is that we can now recognize parameter spaces
and treat them appropriately.  In particular, the domains of affine
expressions can now be correctly identified as being either (possibly
zero-dimensional) sets or parameter domains.

Signed-off-by: Sven Verdoolaege <skimo@kotnet.org>

allow isl_ids with negative reference counts

This will be useful for static isl_ids that are not supposed to change.

Signed-off-by: Sven Verdoolaege <skimo@kotnet.org>

proper implementation of isl_union_map_from_range

The original implementation would just return the input
because isl_sets and isl_maps have the same internal representation.
This will change in the near future, however, and so we need
to actually compute the range of all the elements in the input.

Signed-off-by: Sven Verdoolaege <skimo@kotnet.org>

isl_map_from_domain_and_range: drop explicit conversion to isl_map

Signed-off-by: Sven Verdoolaege <skimo@kotnet.org>

isl_set_apply_pw_qpolynomial_fold: drop explicit conversion to isl_map

Signed-off-by: Sven Verdoolaege <skimo@kotnet.org>

isl_tab_basic_set_non_trivial_lexmin: drop explicit conversion to isl_basic_map

And similarly for isl_tab_basic_set_non_neg_lexmin.

In future, the space of an isl_basic_map will be different from that
of an isl_basic_set.  Calling isl_basic_map_from_range may therefore
result in possibly costly changes.

Signed-off-by: Sven Verdoolaege <skimo@kotnet.org>

isl_tab_pip.c: fix whitespace issue

Signed-off-by: Sven Verdoolaege <skimo@kotnet.org>