Skip to content
Projects
Groups
Snippets
Help
Loading...
Help
Support
Keyboard shortcuts
?
Submit feedback
Contribute to GitLab
Sign in / Register
Toggle navigation
C
cython
Project overview
Project overview
Details
Activity
Releases
Repository
Repository
Files
Commits
Branches
Tags
Contributors
Graph
Compare
Labels
Merge Requests
0
Merge Requests
0
Analytics
Analytics
Repository
Value Stream
Snippets
Snippets
Members
Members
Collapse sidebar
Close sidebar
Activity
Graph
Commits
Open sidebar
nexedi
cython
Commits
21205bb0
Commit
21205bb0
authored
Mar 04, 2010
by
Stefan Behnel
Browse files
Options
Browse Files
Download
Email Patches
Plain Diff
moved code for flattening parallel assignments from parser into PostParse transform
parent
c4f5a501
Changes
3
Show whitespace changes
Inline
Side-by-side
Showing
3 changed files
with
147 additions
and
122 deletions
+147
-122
Cython/Compiler/ParseTreeTransforms.py
Cython/Compiler/ParseTreeTransforms.py
+136
-0
Cython/Compiler/Parsing.pxd
Cython/Compiler/Parsing.pxd
+0
-2
Cython/Compiler/Parsing.py
Cython/Compiler/Parsing.py
+11
-120
No files found.
Cython/Compiler/ParseTreeTransforms.py
View file @
21205bb0
...
@@ -242,6 +242,142 @@ class PostParse(CythonTransform):
...
@@ -242,6 +242,142 @@ class PostParse(CythonTransform):
self
.
context
.
nonfatal_error
(
e
)
self
.
context
.
nonfatal_error
(
e
)
return
None
return
None
def
visit_ParallelAssignmentNode
(
self
,
node
):
"""Flatten parallel assignments into separate single
assignments or cascaded assignments.
"""
self
.
visitchildren
(
node
)
expr_list
=
[]
for
assign_node
in
node
.
stats
:
if
isinstance
(
assign_node
,
Nodes
.
CascadedAssignmentNode
):
expr_list
.
extend
(
assign_node
.
lhs_list
)
else
:
expr_list
.
append
(
assign_node
.
lhs
)
expr_list
.
append
(
assign_node
.
rhs
)
expr_list_list
=
[]
flatten_parallel_assignments
(
expr_list
,
expr_list_list
)
nodes
=
[]
for
expr_list
in
expr_list_list
:
lhs_list
=
expr_list
[:
-
1
]
rhs
=
expr_list
[
-
1
]
if
len
(
lhs_list
)
==
1
:
node
=
Nodes
.
SingleAssignmentNode
(
rhs
.
pos
,
lhs
=
lhs_list
[
0
],
rhs
=
rhs
)
else
:
node
=
Nodes
.
CascadedAssignmentNode
(
rhs
.
pos
,
lhs_list
=
lhs_list
,
rhs
=
rhs
)
nodes
.
append
(
node
)
if
len
(
nodes
)
==
1
:
return
nodes
[
0
]
else
:
return
Nodes
.
ParallelAssignmentNode
(
nodes
[
0
].
pos
,
stats
=
nodes
)
def
flatten_parallel_assignments
(
input
,
output
):
# The input is a list of expression nodes, representing the LHSs
# and RHS of one (possibly cascaded) assignment statement. For
# sequence constructors, rearranges the matching parts of both
# sides into a list of equivalent assignments between the
# individual elements. This transformation is applied
# recursively, so that nested structures get matched as well.
rhs
=
input
[
-
1
]
if
not
rhs
.
is_sequence_constructor
or
not
sum
([
lhs
.
is_sequence_constructor
for
lhs
in
input
[:
-
1
]]):
output
.
append
(
input
)
return
complete_assignments
=
[]
rhs_size
=
len
(
rhs
.
args
)
lhs_targets
=
[
[]
for
_
in
range
(
rhs_size
)
]
starred_assignments
=
[]
for
lhs
in
input
[:
-
1
]:
if
not
lhs
.
is_sequence_constructor
:
if
lhs
.
is_starred
:
error
(
lhs
.
pos
,
"starred assignment target must be in a list or tuple"
)
complete_assignments
.
append
(
lhs
)
continue
lhs_size
=
len
(
lhs
.
args
)
starred_targets
=
sum
([
1
for
expr
in
lhs
.
args
if
expr
.
is_starred
])
if
starred_targets
:
if
starred_targets
>
1
:
error
(
lhs
.
pos
,
"more than 1 starred expression in assignment"
)
output
.
append
([
lhs
,
rhs
])
continue
elif
lhs_size
-
starred_targets
>
rhs_size
:
error
(
lhs
.
pos
,
"need more than %d value%s to unpack"
%
(
rhs_size
,
(
rhs_size
!=
1
)
and
's'
or
''
))
output
.
append
([
lhs
,
rhs
])
continue
map_starred_assignment
(
lhs_targets
,
starred_assignments
,
lhs
.
args
,
rhs
.
args
)
else
:
if
lhs_size
>
rhs_size
:
error
(
lhs
.
pos
,
"need more than %d value%s to unpack"
%
(
rhs_size
,
(
rhs_size
!=
1
)
and
's'
or
''
))
output
.
append
([
lhs
,
rhs
])
continue
elif
lhs_size
<
rhs_size
:
error
(
lhs
.
pos
,
"too many values to unpack (expected %d, got %d)"
%
(
lhs_size
,
rhs_size
))
output
.
append
([
lhs
,
rhs
])
continue
else
:
for
targets
,
expr
in
zip
(
lhs_targets
,
lhs
.
args
):
targets
.
append
(
expr
)
if
complete_assignments
:
complete_assignments
.
append
(
rhs
)
output
.
append
(
complete_assignments
)
# recursively flatten partial assignments
for
cascade
,
rhs
in
zip
(
lhs_targets
,
rhs
.
args
):
if
cascade
:
cascade
.
append
(
rhs
)
flatten_parallel_assignments
(
cascade
,
output
)
# recursively flatten starred assignments
for
cascade
in
starred_assignments
:
if
cascade
[
0
].
is_sequence_constructor
:
flatten_parallel_assignments
(
cascade
,
output
)
else
:
output
.
append
(
cascade
)
def
map_starred_assignment
(
lhs_targets
,
starred_assignments
,
lhs_args
,
rhs_args
):
# Appends the fixed-position LHS targets to the target list that
# appear left and right of the starred argument.
#
# The starred_assignments list receives a new tuple
# (lhs_target, rhs_values_list) that maps the remaining arguments
# (those that match the starred target) to a list.
# left side of the starred target
for
i
,
(
targets
,
expr
)
in
enumerate
(
zip
(
lhs_targets
,
lhs_args
)):
if
expr
.
is_starred
:
starred
=
i
lhs_remaining
=
len
(
lhs_args
)
-
i
-
1
break
targets
.
append
(
expr
)
else
:
raise
InternalError
(
"no starred arg found when splitting starred assignment"
)
# right side of the starred target
for
i
,
(
targets
,
expr
)
in
enumerate
(
zip
(
lhs_targets
[
-
lhs_remaining
:],
lhs_args
[
-
lhs_remaining
:])):
targets
.
append
(
expr
)
# the starred target itself, must be assigned a (potentially empty) list
target
=
lhs_args
[
starred
].
target
# unpack starred node
starred_rhs
=
rhs_args
[
starred
:]
if
lhs_remaining
:
starred_rhs
=
starred_rhs
[:
-
lhs_remaining
]
if
starred_rhs
:
pos
=
starred_rhs
[
0
].
pos
else
:
pos
=
target
.
pos
starred_assignments
.
append
([
target
,
ExprNodes
.
ListNode
(
pos
=
pos
,
args
=
starred_rhs
)])
class
PxdPostParse
(
CythonTransform
,
SkipDeclarations
):
class
PxdPostParse
(
CythonTransform
,
SkipDeclarations
):
"""
"""
Basic interpretation/validity checking that should only be
Basic interpretation/validity checking that should only be
...
...
Cython/Compiler/Parsing.pxd
View file @
21205bb0
...
@@ -59,8 +59,6 @@ cpdef p_testlist(PyrexScanner s)
...
@@ -59,8 +59,6 @@ cpdef p_testlist(PyrexScanner s)
#
#
#-------------------------------------------------------
#-------------------------------------------------------
cpdef
flatten_parallel_assignments
(
input
,
output
)
cpdef
p_global_statement
(
PyrexScanner
s
)
cpdef
p_global_statement
(
PyrexScanner
s
)
cpdef
p_expression_or_assignment
(
PyrexScanner
s
)
cpdef
p_expression_or_assignment
(
PyrexScanner
s
)
cpdef
p_print_statement
(
PyrexScanner
s
)
cpdef
p_print_statement
(
PyrexScanner
s
)
...
...
Cython/Compiler/Parsing.py
View file @
21205bb0
...
@@ -917,129 +917,20 @@ def p_expression_or_assignment(s):
...
@@ -917,129 +917,20 @@ def p_expression_or_assignment(s):
return
Nodes
.
PassStatNode
(
expr
.
pos
)
return
Nodes
.
PassStatNode
(
expr
.
pos
)
else
:
else
:
return
Nodes
.
ExprStatNode
(
expr
.
pos
,
expr
=
expr
)
return
Nodes
.
ExprStatNode
(
expr
.
pos
,
expr
=
expr
)
else
:
expr_list_list
=
[]
flatten_parallel_assignments
(
expr_list
,
expr_list_list
)
nodes
=
[]
for
expr_list
in
expr_list_list
:
lhs_list
=
expr_list
[:
-
1
]
rhs
=
expr_list
[
-
1
]
rhs
=
expr_list
[
-
1
]
if
len
(
lhs_list
)
==
1
:
if
len
(
expr_list
)
==
2
:
node
=
Nodes
.
SingleAssignmentNode
(
rhs
.
pos
,
node
=
Nodes
.
SingleAssignmentNode
(
rhs
.
pos
,
lhs
=
lhs
_list
[
0
],
rhs
=
rhs
)
lhs
=
expr
_list
[
0
],
rhs
=
rhs
)
else
:
else
:
node
=
Nodes
.
CascadedAssignmentNode
(
rhs
.
pos
,
node
=
Nodes
.
CascadedAssignmentNode
(
rhs
.
pos
,
lhs_list
=
lhs_list
,
rhs
=
rhs
)
lhs_list
=
expr_list
[:
-
1
],
rhs
=
rhs
)
nodes
.
append
(
node
)
if
len
(
nodes
)
==
1
:
return
nodes
[
0
]
else
:
return
Nodes
.
ParallelAssignmentNode
(
nodes
[
0
].
pos
,
stats
=
nodes
)
def
flatten_parallel_assignments
(
input
,
output
):
# The input is a list of expression nodes, representing the LHSs
# and RHS of one (possibly cascaded) assignment statement. For
# sequence constructors, rearranges the matching parts of both
# sides into a list of equivalent assignments between the
# individual elements. This transformation is applied
# recursively, so that nested structures get matched as well.
rhs
=
input
[
-
1
]
if
not
rhs
.
is_sequence_constructor
or
not
sum
([
lhs
.
is_sequence_constructor
for
lhs
in
input
[:
-
1
]]):
output
.
append
(
input
)
return
complete_assignments
=
[]
if
sum
([
1
for
expr
in
expr_list
if
expr
.
is_sequence_constructor
])
>
1
:
# at least one parallel assignment
rhs_size
=
len
(
rhs
.
args
)
return
Nodes
.
ParallelAssignmentNode
(
node
.
pos
,
stats
=
[
node
])
lhs_targets
=
[
[]
for
_
in
range
(
rhs_size
)
]
starred_assignments
=
[]
for
lhs
in
input
[:
-
1
]:
if
not
lhs
.
is_sequence_constructor
:
if
lhs
.
is_starred
:
error
(
lhs
.
pos
,
"starred assignment target must be in a list or tuple"
)
complete_assignments
.
append
(
lhs
)
continue
lhs_size
=
len
(
lhs
.
args
)
starred_targets
=
sum
([
1
for
expr
in
lhs
.
args
if
expr
.
is_starred
])
if
starred_targets
:
if
starred_targets
>
1
:
error
(
lhs
.
pos
,
"more than 1 starred expression in assignment"
)
output
.
append
([
lhs
,
rhs
])
continue
elif
lhs_size
-
starred_targets
>
rhs_size
:
error
(
lhs
.
pos
,
"need more than %d value%s to unpack"
%
(
rhs_size
,
(
rhs_size
!=
1
)
and
's'
or
''
))
output
.
append
([
lhs
,
rhs
])
continue
map_starred_assignment
(
lhs_targets
,
starred_assignments
,
lhs
.
args
,
rhs
.
args
)
else
:
if
lhs_size
>
rhs_size
:
error
(
lhs
.
pos
,
"need more than %d value%s to unpack"
%
(
rhs_size
,
(
rhs_size
!=
1
)
and
's'
or
''
))
output
.
append
([
lhs
,
rhs
])
continue
elif
lhs_size
<
rhs_size
:
error
(
lhs
.
pos
,
"too many values to unpack (expected %d, got %d)"
%
(
lhs_size
,
rhs_size
))
output
.
append
([
lhs
,
rhs
])
continue
else
:
for
targets
,
expr
in
zip
(
lhs_targets
,
lhs
.
args
):
targets
.
append
(
expr
)
if
complete_assignments
:
complete_assignments
.
append
(
rhs
)
output
.
append
(
complete_assignments
)
# recursively flatten partial assignments
for
cascade
,
rhs
in
zip
(
lhs_targets
,
rhs
.
args
):
if
cascade
:
cascade
.
append
(
rhs
)
flatten_parallel_assignments
(
cascade
,
output
)
# recursively flatten starred assignments
for
cascade
in
starred_assignments
:
if
cascade
[
0
].
is_sequence_constructor
:
flatten_parallel_assignments
(
cascade
,
output
)
else
:
output
.
append
(
cascade
)
def
map_starred_assignment
(
lhs_targets
,
starred_assignments
,
lhs_args
,
rhs_args
):
# Appends the fixed-position LHS targets to the target list that
# appear left and right of the starred argument.
#
# The starred_assignments list receives a new tuple
# (lhs_target, rhs_values_list) that maps the remaining arguments
# (those that match the starred target) to a list.
# left side of the starred target
for
i
,
(
targets
,
expr
)
in
enumerate
(
zip
(
lhs_targets
,
lhs_args
)):
if
expr
.
is_starred
:
starred
=
i
lhs_remaining
=
len
(
lhs_args
)
-
i
-
1
break
targets
.
append
(
expr
)
else
:
else
:
raise
InternalError
(
"no starred arg found when splitting starred assignment"
)
return
node
# right side of the starred target
for
i
,
(
targets
,
expr
)
in
enumerate
(
zip
(
lhs_targets
[
-
lhs_remaining
:],
lhs_args
[
-
lhs_remaining
:])):
targets
.
append
(
expr
)
# the starred target itself, must be assigned a (potentially empty) list
target
=
lhs_args
[
starred
].
target
# unpack starred node
starred_rhs
=
rhs_args
[
starred
:]
if
lhs_remaining
:
starred_rhs
=
starred_rhs
[:
-
lhs_remaining
]
if
starred_rhs
:
pos
=
starred_rhs
[
0
].
pos
else
:
pos
=
target
.
pos
starred_assignments
.
append
([
target
,
ExprNodes
.
ListNode
(
pos
=
pos
,
args
=
starred_rhs
)])
def
p_print_statement
(
s
):
def
p_print_statement
(
s
):
# s.sy == 'print'
# s.sy == 'print'
...
...
Write
Preview
Markdown
is supported
0%
Try again
or
attach a new file
Attach a file
Cancel
You are about to add
0
people
to the discussion. Proceed with caution.
Finish editing this message first!
Cancel
Please
register
or
sign in
to comment