Skip to content
体验新版
项目
组织
正在加载...
登录
切换导航
打开侧边栏
int
Rust
提交
a5dcf66a
R
Rust
项目概览
int
/
Rust
大约 1 年 前同步成功
通知
1
Star
0
Fork
0
代码
文件
提交
分支
Tags
贡献者
分支图
Diff
Issue
0
列表
看板
标记
里程碑
合并请求
0
DevOps
流水线
流水线任务
计划
Wiki
0
Wiki
分析
仓库
DevOps
项目成员
Pages
R
Rust
项目概览
项目概览
详情
发布
仓库
仓库
文件
提交
分支
标签
贡献者
分支图
比较
Issue
0
Issue
0
列表
看板
标记
里程碑
合并请求
0
合并请求
0
Pages
DevOps
DevOps
流水线
流水线任务
计划
分析
分析
仓库分析
DevOps
Wiki
0
Wiki
成员
成员
收起侧边栏
关闭侧边栏
动态
分支图
创建新Issue
流水线任务
提交
Issue看板
体验新版 GitCode,发现更多精彩内容 >>
提交
a5dcf66a
编写于
11月 04, 2011
作者:
D
David Rajchenbach-Teller
提交者:
Brian Anderson
11月 05, 2011
浏览文件
操作
浏览文件
下载
电子邮件补丁
差异文件
stdlib: Added a small rope library
上级
07574363
变更
3
隐藏空白更改
内联
并排
Showing
3 changed file
with
1134 addition
and
0 deletion
+1134
-0
src/lib/rope.rs
src/lib/rope.rs
+1131
-0
src/lib/std.rc
src/lib/std.rc
+2
-0
src/test/stdtest/stdtest.rc
src/test/stdtest/stdtest.rc
+1
-0
未找到文件。
src/lib/rope.rs
0 → 100644
浏览文件 @
a5dcf66a
/*
Module: rope
High-level text containers.
Ropes are a high-level representation of text that offers
much better performance than strings for common operations,
and generally reduce memory allocations and copies, while only
entailing a small degradation of less common operations.
More precisely, where a string is represented as a memory buffer,
a rope is a tree structure whose leaves are slices of immutable
strings. Therefore, concatenation, appending, prepending, substrings,
etc. are operations that require only trivial tree manipulation,
generally without having to copy memory. In addition, the tree
structure of ropes makes them suitable as a form of index to speed-up
access to Unicode characters by index in long chunks of text.
The following operations are algorithmically faster in ropes:
- extracting a subrope is logarithmic (linear in strings);
- appending/prepending is near-constant time (linear in strings);
- concatenation is near-constant time (linear in strings);
- char length is constant-time (linear in strings);
- access to a character by index is logarithmic (linear in strings);
*/
/*
Type: rope
The type of ropes.
*/
type
rope
=
node
::
root
;
/*
Section: Creating a rope
*/
/*
Function:empty
Create an empty rope
*/
fn
empty
()
->
rope
{
ret
node
::
empty
;
}
/*
Function: of_str
Adopt a string as a rope.
Parameters:
str - A valid string.
Returns:
A rope representing the same string as `str`. Depending of the length
of `str`, this rope may be empty, flat or complex.
Performance notes:
- this operation does not copy the string;
- the function runs in linear time.
*/
fn
of_str
(
str
:
@
str
)
->
rope
{
ret
of_substr
(
str
,
0u
,
str
::
byte_len
(
*
str
));
}
/*
Function: of_substr
As `of_str` but for a substring.
Performance note:
- this operation does not copy the substring.
Parameters:
byte_offset - The offset of `str` at which the rope starts.
byte_len - The number of bytes of `str` to use.
Returns:
A rope representing the same string as
`str::substr(str, byte_offset, byte_len)`.
Depending on `byte_len`, this rope may be empty, flat or complex.
Safety notes:
- this function does _not_ check the validity of the substring;
- this function fails if `byte_offset` or `byte_len` do not match `str`.
*/
fn
of_substr
(
str
:
@
str
,
byte_offset
:
uint
,
byte_len
:
uint
)
->
rope
{
if
byte_len
==
0u
{
ret
node
::
empty
;
}
ret
node
::
content
(
node
::
of_substr
(
str
,
byte_offset
,
byte_len
));
}
/*
Section: Adding things to a rope
*/
/*
Function: append_char
Add one char to the end of the rope
Performance note:
- this function executes in near-constant time
*/
fn
append_char
(
rope
:
rope
,
char
:
char
)
->
rope
{
ret
append_str
(
rope
,
@
str
::
from_chars
([
char
]));
}
/*
Function: append_str
Add one string to the end of the rope
Performance note:
- this function executes in near-linear time
*/
fn
append_str
(
rope
:
rope
,
str
:
@
str
)
->
rope
{
ret
append_rope
(
rope
,
of_str
(
str
))
}
/*
Function: prepend_char
Add one char to the beginning of the rope
Performance note:
- this function executes in near-constant time
*/
fn
prepend_char
(
rope
:
rope
,
char
:
char
)
->
rope
{
ret
prepend_str
(
rope
,
@
str
::
from_chars
([
char
]));
}
/*
Function: prepend_str
Add one string to the beginning of the rope
Performance note:
- this function executes in near-linear time
*/
fn
prepend_str
(
rope
:
rope
,
str
:
@
str
)
->
rope
{
ret
append_rope
(
of_str
(
str
),
rope
)
}
/*
Function: append_rope
Concatenate two ropes
*/
fn
append_rope
(
left
:
rope
,
right
:
rope
)
->
rope
{
alt
(
left
)
{
node
::
empty
.
{
ret
right
;
}
node
::
content
(
left_content
)
{
alt
(
right
)
{
node
::
empty
.
{
ret
left
;
}
node
::
content
(
right_content
)
{
ret
node
::
content
(
node
::
concat2
(
left_content
,
right_content
));
}
}
}
}
}
/*
Function: concat
Concatenate many ropes
*/
fn
concat
(
v
:
[
rope
])
->
rope
{
let
acc
=
node
::
empty
;
for
r
:
rope
in
v
{
acc
=
append_rope
(
acc
,
r
);
}
ret
bal
(
acc
);
}
/*
Section: Keeping ropes healthy
*/
/*
Function: bal
Balance a rope.
Returns:
A copy of the rope in which small nodes have been grouped in memory,
and with a reduced height.
If you perform numerous rope concatenations, it is generally a good idea
to rebalance your rope at some point, before using it for other purposes.
*/
fn
bal
(
rope
:
rope
)
->
rope
{
alt
(
rope
)
{
node
::
empty
.
{
ret
rope
}
node
::
content
(
x
)
{
alt
(
node
::
bal
(
x
))
{
option
::
none
.
{
rope
}
option
::
some
(
y
)
{
node
::
content
(
y
)
}
}
}
}
}
/*
Section: Transforming ropes
*/
/*
Function: sub_chars
Extract a subrope from a rope.
Performance note:
- on a balanced rope, this operation takes algorithmic time;
- this operation does not involve any copying
Safety note:
- this function fails if char_offset/char_len do not represent
valid positions in rope
*/
fn
sub_chars
(
rope
:
rope
,
char_offset
:
uint
,
char_len
:
uint
)
->
rope
{
if
char_len
==
0u
{
ret
node
::
empty
;
}
alt
(
rope
)
{
node
::
empty
.
{
fail
}
node
::
content
(
node
)
{
if
char_len
>
node
::
char_len
(
node
)
{
fail
}
else
{
ret
node
::
content
(
node
::
sub_chars
(
node
,
char_offset
,
char_len
))
}
}
}
}
/*
Function:sub_bytes
Extract a subrope from a rope.
Performance note:
- on a balanced rope, this operation takes algorithmic time;
- this operation does not involve any copying
Safety note:
- this function fails if byte_offset/byte_len do not represent
valid positions in rope
*/
fn
sub_bytes
(
rope
:
rope
,
byte_offset
:
uint
,
byte_len
:
uint
)
->
rope
{
if
byte_len
==
0u
{
ret
node
::
empty
;
}
alt
(
rope
)
{
node
::
empty
.
{
fail
}
node
::
content
(
node
)
{
if
byte_len
>
node
::
byte_len
(
node
)
{
fail
}
else
{
ret
node
::
content
(
node
::
sub_bytes
(
node
,
byte_offset
,
byte_len
))
}
}
}
}
/*
Section: Comparing ropes
*/
/*
Function: cmp
Compare two ropes by Unicode lexicographical order.
This function compares only the contents of the rope, not their structure.
Returns:
A negative value if `left < right`, 0 if eq(left, right) or a positive
value if `left > right`
*/
fn
cmp
(
left
:
rope
,
right
:
rope
)
->
int
{
alt
((
left
,
right
))
{
(
node
::
empty
.
,
node
::
empty
.
)
{
ret
0
;
}
(
node
::
empty
.
,
_
)
{
ret
-
1
;}
(
_
,
node
::
empty
.
)
{
ret
1
;}
(
node
::
content
(
a
),
node
::
content
(
b
))
{
ret
node
::
cmp
(
a
,
b
);
}
}
}
fn
eq
(
left
:
rope
,
right
:
rope
)
->
bool
{
ret
cmp
(
left
,
right
)
==
0
;
}
fn
leq
(
left
:
rope
,
right
:
rope
)
->
bool
{
ret
cmp
(
left
,
right
)
<=
0
;
}
fn
lt
(
left
:
rope
,
right
:
rope
)
->
bool
{
ret
cmp
(
left
,
right
)
<
0
;
}
fn
geq
(
left
:
rope
,
right
:
rope
)
->
bool
{
ret
cmp
(
left
,
right
)
>=
0
;
}
fn
gt
(
left
:
rope
,
right
:
rope
)
->
bool
{
ret
cmp
(
left
,
right
)
>
0
;
}
/*
Section: Iterating
*/
/*
Function: loop_chars
Loop through a rope, char by char
While other mechanisms are available, this is generally the best manner
of looping through the contents of a rope char by char. If you prefer a
loop that iterates through the contents string by string (e.g. to print
the contents of the rope or output it to the system), however,
you should rather use `traverse_components`.
Parameters:
rope - A rope to traverse. It may be empty.
it - A block to execute with each consecutive character of the rope.
Return `true` to continue, `false` to stop.
Returns:
`true` If execution proceeded correctly, `false` if it was interrupted,
that is if `it` returned `false` at any point.
*/
fn
loop_chars
(
rope
:
rope
,
it
:
block
(
char
)
->
bool
)
->
bool
{
alt
(
rope
)
{
node
::
empty
.
{
ret
true
}
node
::
content
(
x
)
{
ret
node
::
loop_chars
(
x
,
it
)
}
}
}
fn
iter_chars
(
rope
:
rope
,
it
:
block
(
char
))
{
loop_chars
(
rope
)
{|
x
|
it
(
x
);
ret
true
}
}
/*
Function: loop_leaves
Loop through a rope, string by string
While other mechanisms are available, this is generally the best manner of
looping through the contents of a rope string by string, which may be useful
e.g. to print strings as you see them (without having to copy their
contents into a new string), to send them to then network, to write them to
a file, etc.. If you prefer a loop that iterates through the contents
char by char (e.g. to search for a char), however, you should rather
use `traverse`.
Parameters:
rope - A rope to traverse. It may be empty.
it - A block to execute with each consecutive string component of the rope.
Return `true` to continue, `false` to stop.
Returns:
`true` If execution proceeded correctly, `false` if it was interrupted,
that is if `it` returned `false` at any point.
*/
fn
loop_leaves
(
rope
:
rope
,
it
:
block
(
node
::
leaf
)
->
bool
)
->
bool
{
alt
(
rope
)
{
node
::
empty
.
{
ret
true
}
node
::
content
(
x
)
{
ret
node
::
loop_leaves
(
x
,
it
)}
}
}
mod
iterator
{
mod
leaf
{
fn
start
(
rope
:
rope
)
->
node
::
leaf_iterator
::
t
{
alt
(
rope
)
{
node
::
empty
.
{
ret
node
::
leaf_iterator
::
empty
()
}
node
::
content
(
x
)
{
ret
node
::
leaf_iterator
::
start
(
x
)
}
}
}
fn
next
(
it
:
node
::
leaf_iterator
::
t
)
->
option
::
t
<
node
::
leaf
>
{
ret
node
::
leaf_iterator
::
next
(
it
);
}
}
mod
char
{
fn
start
(
rope
:
rope
)
->
node
::
char_iterator
::
t
{
alt
(
rope
)
{
node
::
empty
.
{
ret
node
::
char_iterator
::
empty
()
}
node
::
content
(
x
)
{
ret
node
::
char_iterator
::
start
(
x
)
}
}
}
fn
next
(
it
:
node
::
char_iterator
::
t
)
->
option
::
t
<
char
>
{
ret
node
::
char_iterator
::
next
(
it
)
}
}
}
/*
Section: Rope properties
*/
/*
Function: height
Returns: The height of the rope, i.e. a bound on the number of
operations which must be performed during a character access before
finding the leaf in which a character is contained.
Performance note: Constant time.
*/
fn
height
(
rope
:
rope
)
->
uint
{
alt
(
rope
)
{
node
::
empty
.
{
ret
0u
;
}
node
::
content
(
x
)
{
ret
node
::
height
(
x
);
}
}
}
/*
Function: char_len
Returns: The number of character in the rope
Performance note: Constant time.
*/
pure
fn
char_len
(
rope
:
rope
)
->
uint
{
alt
(
rope
)
{
node
::
empty
.
{
ret
0u
;
}
node
::
content
(
x
)
{
ret
node
::
char_len
(
x
)
}
}
}
/*
Function: char_len
Returns: The number of bytes in the rope
Performance note: Constant time.
*/
pure
fn
byte_len
(
rope
:
rope
)
->
uint
{
alt
(
rope
)
{
node
::
empty
.
{
ret
0u
;
}
node
::
content
(
x
)
{
ret
node
::
byte_len
(
x
)
}
}
}
/*
Function: char_at
Parameters:
pos - A position in the rope
Returns: The character at position `pos`
Safety notes: The function will fail if `pos`
is not a valid position in the rope.
Performance note: This function executes in a time
proportional to the height of the rope + the (bounded)
length of the largest leaf.
*/
fn
char_at
(
rope
:
rope
,
pos
:
uint
)
->
char
{
alt
(
rope
)
{
node
::
empty
.
{
fail
}
node
::
content
(
x
)
{
ret
node
::
char_at
(
x
,
pos
)
}
}
}
/*
Section: Implementation
*/
mod
node
{
/*
Enum: node::root
Implementation of type `rope`
Constants:
empty - An empty rope
content - A non-empty rope
*/
tag
root
{
empty
;
content
(
@
node
);
}
/*
Struct: node::leaf
A text component in a rope.
This is actually a slice in a rope, so as to ensure maximal sharing.
*/
type
leaf
=
{
/*
Field: byte_offset
The number of bytes skipped in `content`
*/
byte_offset
:
uint
,
/*
Field: byte_len
The number of bytes of `content` to use
*/
byte_len
:
uint
,
/*
Field: char_len
The number of chars in the leaf.
*/
char_len
:
uint
,
/*
Field: content
Contents of the leaf.
Note that we can have `char_len < str::char_len(content)`, if this
leaf is only a subset of the string. Also note that the string
can be shared between several ropes, e.g. for indexing purposes.
*/
content
:
@
str
};
/*
Struct node::concat
A node obtained from the concatenation of two other nodes
*/
type
concat
=
{
/*
Field: left
The node containing the beginning of the text.
*/
left
:
@
node
,
//TODO: Perhaps a `vec` instead of `left`/`right`
/*
Field: right
The node containing the end of the text.
*/
right
:
@
node
,
/*
Field: char_len
The number of chars contained in all leaves of this node.
*/
char_len
:
uint
,
/*
Field: byte_len
The number of bytes in the subrope.
Used to pre-allocate the correct amount of storage for serialization.
*/
byte_len
:
uint
,
/*
Field: height
Height of the subrope.
Used for rebalancing and to allocate stacks for
traversals.
*/
height
:
uint
};
/*
Enum: node::node
leaf - A leaf consisting in a `str`
concat - The concatenation of two ropes
*/
tag
node
{
leaf
(
leaf
);
concat
(
concat
);
}
/*
The maximal number of chars that _should_ be permitted in a single node.
This is not a strict value
*/
const
hint_max_leaf_char_len
:
uint
=
256u
;
/*
The maximal height that _should_ be permitted in a tree.
This is not a strict value
*/
const
hint_max_node_height
:
uint
=
16u
;
/*
Adopt a string as a node.
If the string is longer than `max_leaf_char_len`, it is
logically split between as many leaves as necessary. Regardless,
the string itself is not copied.
Performance note: The complexity of this function is linear in
the length of `str`.
*/
fn
of_str
(
str
:
@
str
)
->
@
node
{
ret
of_substr
(
str
,
0u
,
str
::
byte_len
(
*
str
));
}
/*
Adopt a slice of a string as a node.
If the slice is longer than `max_leaf_char_len`, it is logically split
between as many leaves as necessary. Regardless, the string itself
is not copied.
@param byte_start The byte offset where the slice of `str` starts.
@param byte_len The number of bytes from `str` to use.
*/
fn
of_substr
(
str
:
@
str
,
byte_start
:
uint
,
byte_len
:
uint
)
->
@
node
{
assert
(
byte_len
>
0u
);
let
char_len
=
str
::
char_len_range
(
*
str
,
byte_start
,
byte_len
);
let
candidate
=
@
leaf
({
byte_offset
:
byte_start
,
byte_len
:
byte_len
,
char_len
:
char_len
,
content
:
str
});
if
char_len
<=
hint_max_leaf_char_len
{
ret
candidate
;
}
else
{
//Firstly, split `str` in slices of hint_max_leaf_char_len
let
leaves
=
uint
::
div_ceil
(
char_len
,
hint_max_leaf_char_len
);
//Number of leaves
let
nodes
=
vec
::
init_elt_mut
(
candidate
,
leaves
);
let
i
=
0u
;
let
offset
=
byte_start
;
let
first_leaf_char_len
=
if
char_len
%
hint_max_leaf_char_len
==
0u
{
hint_max_leaf_char_len
}
else
{
char_len
%
hint_max_leaf_char_len
};
while
i
<
leaves
{
let
chunk_char_len
:
uint
=
if
i
==
0u
{
first_leaf_char_len
}
else
{
hint_max_leaf_char_len
};
let
chunk_byte_len
=
str
::
byte_len_range
(
*
str
,
offset
,
chunk_char_len
);
nodes
[
i
]
=
@
leaf
({
byte_offset
:
offset
,
byte_len
:
chunk_byte_len
,
char_len
:
chunk_char_len
,
content
:
str
});
offset
+=
chunk_byte_len
;
i
+=
1u
;
}
//Then, build a tree from these slices by collapsing them
while
leaves
>
1u
{
i
=
0u
;
while
i
<
leaves
-
1u
{
//Concat nodes 0 with 1, 2 with 3 etc.
nodes
[
i
/
2u
]
=
concat2
(
nodes
[
i
],
nodes
[
i
+
1u
]);
i
+=
2u
;
}
if
i
==
leaves
-
1u
{
//And don't forget the last node if it is in even position
nodes
[
i
/
2u
]
=
nodes
[
i
];
}
leaves
=
uint
::
div_ceil
(
leaves
,
2u
);
}
ret
nodes
[
0u
];
}
}
pure
fn
byte_len
(
node
:
@
node
)
->
uint
{
alt
(
*
node
)
{
//TODO: Could we do this without the pattern-matching?
leaf
(
y
)
{
ret
y
.byte_len
;
}
concat
(
y
){
ret
y
.byte_len
;
}
}
}
pure
fn
char_len
(
node
:
@
node
)
->
uint
{
alt
(
*
node
)
{
leaf
(
y
)
{
ret
y
.char_len
;
}
concat
(
y
)
{
ret
y
.char_len
;
}
}
}
fn
tree_from_forest_destructive
(
forest
:
[
mutable
@
node
])
->
@
node
{
let
i
=
0u
;
let
len
=
vec
::
len
(
forest
);
while
len
>
1u
{
i
=
0u
;
while
i
<
len
-
1u
{
//Concat nodes 0 with 1, 2 with 3 etc.
let
left
=
forest
[
i
];
let
right
=
forest
[
i
+
1u
];
let
left_len
=
char_len
(
left
);
let
right_len
=
char_len
(
right
);
let
left_height
=
height
(
left
);
let
right_height
=
height
(
right
);
if
left_len
+
right_len
>
hint_max_leaf_char_len
//TODO: Improve strategy
||
left_height
>=
hint_max_node_height
||
right_height
>=
hint_max_node_height
{
if
left_len
<=
hint_max_leaf_char_len
{
left
=
flatten
(
left
);
}
if
right_len
<=
hint_max_leaf_char_len
{
right
=
flatten
(
right
);
}
}
forest
[
i
/
2u
]
=
concat2
(
left
,
right
);
i
+=
2u
;
}
if
i
==
len
-
1u
{
//And don't forget the last node if it is in even position
forest
[
i
/
2u
]
=
forest
[
i
];
}
len
=
uint
::
div_ceil
(
len
,
2u
);
}
ret
forest
[
0
];
}
fn
serialize_node
(
node
:
@
node
)
->
str
unsafe
{
let
buf
=
vec
::
init_elt_mut
(
0u8
,
byte_len
(
node
));
let
offset
=
0u
;
//Current position in the buffer
let
it
=
leaf_iterator
::
start
(
node
);
while
true
{
alt
(
leaf_iterator
::
next
(
it
))
{
option
::
none
.
{
break
;
}
option
::
some
(
x
)
{
//TODO: Replace with memcpy or something similar
let
local_buf
:
[
u8
]
=
unsafe
::
reinterpret_cast
(
*
x
.content
);
let
i
=
x
.byte_offset
;
while
i
<
x
.byte_len
{
buf
[
offset
]
=
local_buf
[
i
];
offset
+=
1u
;
i
+=
1u
;
}
unsafe
::
leak
(
local_buf
);
}
}
}
let
str
:
str
=
unsafe
::
reinterpret_cast
(
buf
);
unsafe
::
leak
(
buf
);
//TODO: Check if this is correct
ret
str
;
}
/*
Replace a subtree by a single leaf with the same contents.
*/
fn
flatten
(
node
:
@
node
)
->
@
node
unsafe
{
alt
(
*
node
)
{
leaf
(
_
)
{
ret
node
}
concat
(
x
)
{
ret
@
leaf
({
byte_offset
:
0u
,
byte_len
:
x
.byte_len
,
char_len
:
x
.char_len
,
content
:
@
serialize_node
(
node
)
})
}
}
}
fn
bal
(
node
:
@
node
)
->
option
::
t
<@
node
>
{
if
height
(
node
)
<
hint_max_node_height
{
ret
option
::
none
}
else
{
//1. Gather all leaves as a forest
let
forest
=
[
mutable
];
let
it
=
leaf_iterator
::
start
(
node
);
while
true
{
alt
(
leaf_iterator
::
next
(
it
))
{
option
::
none
.
{
break
;
}
option
::
some
(
x
)
{
forest
+=
[
mutable
@
leaf
(
x
)];
}
}
}
//2. Rebuild tree from forest
let
root
=
@*
tree_from_forest_destructive
(
forest
);
ret
option
::
some
(
root
);
}
}
fn
sub_bytes
(
node
:
@
node
,
byte_offset
:
uint
,
byte_len
:
uint
)
->
@
node
{
let
node
=
node
;
let
result
=
node
;
//Arbitrary value
let
byte_offset
=
byte_offset
;
let
byte_len
=
byte_len
;
while
true
{
if
byte_offset
==
0u
&&
byte_len
==
node
::
byte_len
(
node
)
{
result
=
node
;
break
;
}
alt
(
*
node
)
{
node
::
leaf
(
x
)
{
let
char_len
=
str
::
char_len_range
(
*
x
.content
,
byte_offset
,
byte_len
);
result
=
@
leaf
({
byte_offset
:
byte_offset
,
byte_len
:
byte_len
,
char_len
:
char_len
,
content
:
x
.content
});
break
;
}
node
::
concat
(
x
)
{
let
left_len
:
uint
=
node
::
byte_len
(
x
.left
);
if
byte_offset
<=
left_len
{
if
byte_offset
+
byte_len
<=
left_len
{
//Case 1: Everything fits in x.left, tail-call
node
=
x
.left
;
}
else
{
//Case 2: A (non-empty, possibly full) suffix
//of x.left and a (non-empty, possibly full) prefix
//of x.right
let
left_result
=
sub_bytes
(
x
.left
,
byte_offset
,
left_len
);
let
right_result
=
sub_bytes
(
x
.right
,
0u
,
left_len
-
byte_offset
);
result
=
concat2
(
left_result
,
right_result
);
break
;
}
}
else
{
//Case 3: Everything fits in x.right
byte_offset
-=
left_len
;
node
=
x
.right
;
}
}
}
}
ret
result
;
}
fn
sub_chars
(
node
:
@
node
,
char_offset
:
uint
,
char_len
:
uint
)
->
@
node
{
alt
(
*
node
)
{
node
::
leaf
(
x
)
{
if
char_offset
==
0u
&&
char_len
==
x
.char_len
{
ret
node
;
}
let
byte_offset
=
str
::
byte_len_range
(
*
x
.content
,
0u
,
char_offset
);
let
byte_len
=
str
::
byte_len_range
(
*
x
.content
,
byte_offset
,
char_len
);
ret
@
leaf
({
byte_offset
:
byte_offset
,
byte_len
:
byte_len
,
char_len
:
char_len
,
content
:
x
.content
});
}
node
::
concat
(
x
)
{
if
char_offset
==
0u
&&
char_len
==
x
.char_len
{
ret
node
;}
let
left_len
:
uint
=
node
::
char_len
(
x
.left
);
if
char_offset
<=
left_len
{
if
char_offset
+
char_len
<=
left_len
{
//Case 1: Everything fits in x.left
ret
sub_chars
(
x
.left
,
char_offset
,
char_len
);
//TODO: Optimize manually this tail call?
}
else
{
//Case 2: A (non-empty, possibly full) suffix
//of x.left and a (non-empty, possibly full) prefix
//of x.right
let
left_result
=
sub_chars
(
x
.left
,
char_offset
,
left_len
);
let
right_result
=
sub_chars
(
x
.right
,
0u
,
left_len
-
char_offset
);
ret
concat2
(
left_result
,
right_result
)
}
}
else
{
//Case 3: Everything fits in x.right
ret
sub_chars
(
x
.right
,
char_offset
-
left_len
,
char_len
);
//TODO: Optimize manually this tail call?
}
}
}
}
fn
concat2
(
left
:
@
node
,
right
:
@
node
)
->
@
node
{
ret
@
concat
({
left
:
left
,
right
:
right
,
char_len
:
char_len
(
left
)
+
char_len
(
right
),
byte_len
:
byte_len
(
left
)
+
byte_len
(
right
),
height
:
math
::
max
(
height
(
left
),
height
(
right
))
+
1u
})
}
fn
height
(
node
:
@
node
)
->
uint
{
alt
(
*
node
)
{
leaf
(
_
)
{
ret
0u
;
}
concat
(
x
)
{
ret
x
.height
;
}
}
}
fn
cmp
(
a
:
@
node
,
b
:
@
node
)
->
int
{
let
ita
=
char_iterator
::
start
(
a
);
let
itb
=
char_iterator
::
start
(
b
);
let
result
=
0
;
let
pos
=
0u
;
while
result
==
0
{
alt
((
char_iterator
::
next
(
ita
),
char_iterator
::
next
(
itb
)))
{
(
option
::
none
.
,
option
::
none
.
)
{
break
;
}
(
option
::
some
(
chara
),
option
::
some
(
charb
))
{
result
=
char
::
cmp
(
chara
,
charb
);
}
(
option
::
some
(
_
),
_
)
{
result
=
1
;
}
(
_
,
option
::
some
(
_
))
{
result
=
-
1
;
}
}
pos
+=
1u
;
}
ret
result
;
}
fn
loop_chars
(
node
:
@
node
,
it
:
block
(
char
)
->
bool
)
->
bool
{
ret
loop_leaves
(
node
,
{|
leaf
|
ret
str
::
loop_chars_sub
(
*
leaf
.content
,
leaf
.byte_offset
,
leaf
.byte_len
,
it
)
})
}
fn
loop_leaves
(
node
:
@
node
,
it
:
block
(
leaf
)
->
bool
)
->
bool
{
let
result
=
true
;
let
current
=
node
;
while
true
{
alt
(
*
current
)
{
leaf
(
x
)
{
result
=
it
(
x
);
break
;
}
concat
(
x
)
{
if
loop_leaves
(
x
.left
,
it
)
{
//non tail call
current
=
x
.right
;
//tail call
}
else
{
result
=
false
;
break
;
}
}
}
}
ret
result
;
}
fn
char_at
(
node
:
@
node
,
pos
:
uint
)
->
char
{
let
node
=
node
;
let
pos
=
pos
;
let
result
=
'0'
;
while
true
{
alt
(
*
node
)
{
leaf
(
x
)
{
result
=
str
::
char_at
(
*
x
.content
,
pos
);
break
;
}
concat
({
left
:
left
,
right
:
right
,
char_len
:
_
,
byte_len
:
_
,
height
:
_
})
{
let
left_len
=
char_len
(
left
);
if
left_len
>
pos
{
node
=
left
;
}
else
{
node
=
right
;
pos
=
pos
-
left_len
;
}
}
}
};
ret
result
;
}
mod
leaf_iterator
{
type
t
=
{
stack
:
[
mutable
@
node
],
mutable
stackpos
:
int
};
fn
empty
()
->
t
{
let
stack
:
[
mutable
@
node
]
=
[
mutable
];
ret
{
stack
:
stack
,
mutable
stackpos
:
-
1
}
}
fn
start
(
node
:
@
node
)
->
t
{
let
stack
=
vec
::
init_elt_mut
(
node
,
height
(
node
)
+
1u
);
ret
{
stack
:
stack
,
mutable
stackpos
:
0
}
}
fn
next
(
it
:
t
)
->
option
::
t
<
leaf
>
{
if
it
.stackpos
<
0
{
ret
option
::
none
;
}
let
result
=
option
::
none
;
while
true
{
let
current
=
it
.stack
[
it
.stackpos
];
it
.stackpos
-=
1
;
alt
(
*
current
)
{
concat
(
x
)
{
it
.stackpos
+=
1
;
it
.stack
[
it
.stackpos
]
=
x
.right
;
it
.stackpos
+=
1
;
it
.stack
[
it
.stackpos
]
=
x
.left
;
}
leaf
(
x
)
{
result
=
option
::
some
(
x
);
break
;
}
}
}
ret
result
;
}
}
mod
char_iterator
{
type
t
=
{
leaf_iterator
:
leaf_iterator
::
t
,
mutable
leaf
:
option
::
t
<
leaf
>
,
mutable
leaf_byte_pos
:
uint
};
fn
start
(
node
:
@
node
)
->
t
{
ret
{
leaf_iterator
:
leaf_iterator
::
start
(
node
),
mutable
leaf
:
option
::
none
,
mutable
leaf_byte_pos
:
0u
}
}
fn
empty
()
->
t
{
ret
{
leaf_iterator
:
leaf_iterator
::
empty
(),
mutable
leaf
:
option
::
none
,
mutable
leaf_byte_pos
:
0u
}
}
fn
next
(
it
:
t
)
->
option
::
t
<
char
>
{
let
result
=
option
::
none
;
while
true
{
alt
(
get_current_or_next_leaf
(
it
))
{
option
::
none
.
{
break
;
}
option
::
some
(
leaf
)
{
let
next_char
=
get_next_char_in_leaf
(
it
);
alt
(
next_char
)
{
option
::
none
.
{
cont
;
}
option
::
some
(
_
)
{
result
=
next_char
;
break
;
}
}
}
}
}
ret
result
;
}
fn
get_current_or_next_leaf
(
it
:
t
)
->
option
::
t
<
leaf
>
{
alt
(
it
.leaf
)
{
option
::
some
(
_
)
{
ret
it
.leaf
}
option
::
none
.
{
let
next
=
leaf_iterator
::
next
(
it
.leaf_iterator
);
alt
(
next
)
{
option
::
none
.
{
ret
option
::
none
}
option
::
some
(
leaf
)
{
it
.leaf
=
next
;
it
.leaf_byte_pos
=
0u
;
ret
next
;
}
}
}
}
}
fn
get_next_char_in_leaf
(
it
:
t
)
->
option
::
t
<
char
>
{
alt
(
it
.leaf
)
{
option
::
none
.
{
ret
option
::
none
}
option
::
some
(
leaf
)
{
if
it
.leaf_byte_pos
>=
leaf
.byte_len
{
//We are actually past the end of the leaf
it
.leaf
=
option
::
none
;
ret
option
::
none
}
else
{
let
{
ch
,
next
}
=
str
::
char_range_at
(
*
leaf
.content
,
it
.leaf_byte_pos
+
leaf
.byte_offset
);
it
.leaf_byte_pos
=
next
-
leaf
.byte_offset
;
ret
option
::
some
(
ch
)
}
}
}
}
}
}
src/lib/std.rc
浏览文件 @
a5dcf66a
...
...
@@ -11,6 +11,7 @@ export box, char, float, int, str, ptr, uint, u8, u32, u64, vec;
export aio, comm, fs, io, net, run, sio, sys, task;
export ctypes, either, option, result, util;
export bitv, deque, fun_treemap, list, map, smallintmap, sort, treemap, ufind;
export rope;
export ebml, dbg, getopts, math, rand, sha1, term, time, unsafe;
export extfmt, test;
// FIXME: generic_os and os_fs shouldn't be exported
...
...
@@ -61,6 +62,7 @@ mod deque;
mod fun_treemap;
mod list;
mod map;
mod rope;
mod smallintmap;
mod sort;
mod treemap;
...
...
src/test/stdtest/stdtest.rc
浏览文件 @
a5dcf66a
...
...
@@ -33,6 +33,7 @@ mod uint;
mod float;
mod math;
mod result;
mod rope;
// Local Variables:
// mode: rust
...
...
编辑
预览
Markdown
is supported
0%
请重试
或
添加新附件
.
添加附件
取消
You are about to add
0
people
to the discussion. Proceed with caution.
先完成此消息的编辑!
取消
想要评论请
注册
或
登录