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
114
115
116
117
118
119
120
121
122
123
124
125
126
127
128
129
130
131
132
133
134
135
136
137
138
139
140
141
142
143
144
145
146
147
148
149
150
151
152
153
154
155
156
157
158
159
160
161
162
163
164
165
166
167
168
169
170
171
172
173
174
175
176
177
178
179
180
181
182
183
184
185
186
187
188
189
190
191
192
193
194
195
196
197
198
199
200
201
202
203
204
205
206
207
208
209
210
211
212
213
214
215
216
217
218
219
|
-------------------------------------------------------------------------------
Alignment thresholds:
Alignment threshold are used to prevent aligning of certain text.
Consider this example when aligning assignments:
a = 2;
gld.settings[somevariable] = 5;
Without thresholds, this would result in:
a = 2;
gld.settings[somevariable] = 5;
However, this is likely not desired.
So to handle this, we set a limit on the number of columns that a alignment
may move a token. If we set the threshold to, say, 3 characters, then
the two would not be aligned together because 'a = 2' would have to be moved
by 25 columns.
Here's a more complicated case:
a = 1;
bb = 2;
ccc = 3;
dddd = 4;
eeeee = 5;
ffffff = 6;
In this case, we may want this output:
a = 1;
bb = 2;
ccc = 3;
dddd = 4;
eeeee = 5;
ffffff = 6;
But, with a threshold of 3, 'a' and 'ffffff' cannot be aligned together.
So how should these thresholds be implemented?
One approach is to use the threshold as the maximum difference between the
current maximum and the new maximum. Using this approach, all 6 lines above
are aligned, because the change for each is 1 column.
Consider this with a line-span of 3 and a column-threshold of 3.
Note that we'll now need to keep track of the last line processed.
(line and column numbers added for clarity)
111111
123456789012345
---------------
1 | a = 1;
2 | eeeee = 5;
3 | ffffff = 6;
On line 1, the maxcol is set to 3.
On line 2, the maxcol would be set to 7, except that exceeds the threshold.
On line 3, the maxcol would be set to 8, except that exceeds the threshold.
So, it ends up with one item in the align stack - line 1.
Then we pick up from where we left off - on line 2.
On line 2, the maxcol is set to 7.
On line 3, the maxcol is set to 8.
End result:
111111
123456789012345
---------------
1 | a = 1;
2 | eeeee = 5;
3 | ffffff = 6;
Now lets get tricky.
111111
123456789012345
---------------
1 | a = 1;
2 | eeeee = 5;
3 | ccc = 3;
4 | ffffff = 6;
How should this be aligned?
I think we need to add another list - a 'skipped' list, that runs in parallel
with the 'aligned' list.
On line 1, the maxcol is set to 3, added to the aligned list.
On line 2, col exceeds the threshold, so it is added to the skipped list.
On line 3, the maxcol is set to 5, added to the aligned list.
Now, after an item is added to the aligned list, the skipped list must be re-scanned.
We now see that line 2 is only 2 columns different, so it is added and maxcol = 7.
The skipped list is now empty.
On line 4, the maxcol is set to 8, added to the aligned list.
So the output is:
111111
123456789012345
---------------
1 | a = 1;
2 | eeeee = 5;
3 | ccc = 3;
4 | ffffff = 6;
Now for a case where the skipped list is not absorbed:
111111
123456789012345
---------------
1 | a = 1;
2 | iiiiiiiiiiiiieeeee = 5;
3 | ccc = 3;
4 | ffffff = 6;
On line 1, the maxcol is set to 3, added to the aligned list.
On line 2, col exceeds the threshold, so it is added to the skipped list.
On line 3, the maxcol is set to 5, added to the aligned list.
Skipped list is scanned, nothing moved from the skipped list to the aligned list.
On line 4, the maxcol is set to 8, added to the aligned list.
So the output is:
111111
123456789012345
---------------
1 | a = 1;
2 | iiiiiiiiiiiiieeeee = 5;
3 | ccc = 3;
4 | ffffff = 6;
As a reminder, the old system would have produced:
1 | a = 1;
2 | iiiiiiiiiiiiieeeee = 5;
3 | ccc = 3;
4 | ffffff = 6;
Which is probably not wanted.
--===---===---===---===--
Absolute thresholds
To get a better grip on what thresholds do the absolute thresholds were introduced.
An absolute threshold means that the item, in this case the assign statement, is never
moved more than the threshold value.
See the below example:
Relative threshold = 10
000000000111111111122222222223
123456789012345678901234567890
1| a = 1
2| aaaaa = 1
3| bbbbbbbbbbb = 2
4| ccccccccccccccccccccc = 3
Absolute threshold:
000000000111111111122222222223
123456789012345678901234567890
1| a = 1
2| aaaaa = 1
3| bbbbbbbbbbb = 2
4| cccccccccccccccccccc = 3
--===---===---===---===--
How to do this generically in the code...
Easy case - one item per line.
An AlignStack is used to manage alignment.
Here are the functions and what each of them do:
AlignStack::Start(line_span, col_thresh)
AlignStack::Add(pc)
AlignStack::Flush()
AlignStack::Newline(count)
AlignStack::End()
For each entry, a sequence number is kept.
AlignStack::Start(line_span, col_thresh)
- initializes the align and skipped lists
- zero max_column
- zero cur_seqnum
- zero nl_count
AlignStack::Add(start, seqnum)
- update cur_seqnum, assign to item
- if item column is within threshold
- zero nl_count
- add to the aligned list
- if item column is > max_col
- update max_col
- re-adds all items on the skipped list
- if item column is not within threshold, add to skipped list
AlignStack::NewLines(count)
- adds count to nl_count
- if nl_count > line_span, call AlignStack::Flush()
AlignStack::Flush()
- step through all the items in aligned list and align the items to max_column
- keep the seq_num of the last aligned item
- zero max_column
- remove all items with a seq_num < last aligned seq_num
- call AlignStack::Add on all remaining items in the skipped list
AlignStack::End()
- call AlignStack::Flush
- clear the lists
- free resources, etc
Example usage: see align_assign function in align_assign.cpp
Chunk *align_assign(Chunk *first, size_t span, size_t thresh, size_t *p_nl_count);
|
