forked from flang-compiler/flang-driver
-
Notifications
You must be signed in to change notification settings - Fork 0
/
FormatToken.cpp
306 lines (269 loc) · 10.6 KB
/
FormatToken.cpp
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
220
221
222
223
224
225
226
227
228
229
230
231
232
233
234
235
236
237
238
239
240
241
242
243
244
245
246
247
248
249
250
251
252
253
254
255
256
257
258
259
260
261
262
263
264
265
266
267
268
269
270
271
272
273
274
275
276
277
278
279
280
281
282
283
284
285
286
287
288
289
290
291
292
293
294
295
296
297
298
299
300
301
302
303
304
305
306
//===--- FormatToken.cpp - Format C++ code --------------------------------===//
//
// Part of the LLVM Project, under the Apache License v2.0 with LLVM Exceptions.
// See https://llvm.org/LICENSE.txt for license information.
// SPDX-License-Identifier: Apache-2.0 WITH LLVM-exception
//
//===----------------------------------------------------------------------===//
///
/// \file
/// This file implements specific functions of \c FormatTokens and their
/// roles.
///
//===----------------------------------------------------------------------===//
#include "FormatToken.h"
#include "ContinuationIndenter.h"
#include "llvm/ADT/SmallVector.h"
#include "llvm/Support/Debug.h"
#include <climits>
namespace clang {
namespace format {
const char *getTokenTypeName(TokenType Type) {
static const char *const TokNames[] = {
#define TYPE(X) #X,
LIST_TOKEN_TYPES
#undef TYPE
nullptr};
if (Type < NUM_TOKEN_TYPES)
return TokNames[Type];
llvm_unreachable("unknown TokenType");
return nullptr;
}
// FIXME: This is copy&pasted from Sema. Put it in a common place and remove
// duplication.
bool FormatToken::isSimpleTypeSpecifier() const {
switch (Tok.getKind()) {
case tok::kw_short:
case tok::kw_long:
case tok::kw___int64:
case tok::kw___int128:
case tok::kw_signed:
case tok::kw_unsigned:
case tok::kw_void:
case tok::kw_char:
case tok::kw_int:
case tok::kw_half:
case tok::kw_float:
case tok::kw_double:
case tok::kw__Float16:
case tok::kw___float128:
case tok::kw_wchar_t:
case tok::kw_bool:
case tok::kw___underlying_type:
case tok::annot_typename:
case tok::kw_char8_t:
case tok::kw_char16_t:
case tok::kw_char32_t:
case tok::kw_typeof:
case tok::kw_decltype:
return true;
default:
return false;
}
}
TokenRole::~TokenRole() {}
void TokenRole::precomputeFormattingInfos(const FormatToken *Token) {}
unsigned CommaSeparatedList::formatAfterToken(LineState &State,
ContinuationIndenter *Indenter,
bool DryRun) {
if (State.NextToken == nullptr || !State.NextToken->Previous)
return 0;
if (Formats.size() == 1)
return 0; // Handled by formatFromToken
// Ensure that we start on the opening brace.
const FormatToken *LBrace =
State.NextToken->Previous->getPreviousNonComment();
if (!LBrace || !LBrace->isOneOf(tok::l_brace, TT_ArrayInitializerLSquare) ||
LBrace->BlockKind == BK_Block || LBrace->Type == TT_DictLiteral ||
LBrace->Next->Type == TT_DesignatedInitializerPeriod)
return 0;
// Calculate the number of code points we have to format this list. As the
// first token is already placed, we have to subtract it.
unsigned RemainingCodePoints =
Style.ColumnLimit - State.Column + State.NextToken->Previous->ColumnWidth;
// Find the best ColumnFormat, i.e. the best number of columns to use.
const ColumnFormat *Format = getColumnFormat(RemainingCodePoints);
// If no ColumnFormat can be used, the braced list would generally be
// bin-packed. Add a severe penalty to this so that column layouts are
// preferred if possible.
if (!Format)
return 10000;
// Format the entire list.
unsigned Penalty = 0;
unsigned Column = 0;
unsigned Item = 0;
while (State.NextToken != LBrace->MatchingParen) {
bool NewLine = false;
unsigned ExtraSpaces = 0;
// If the previous token was one of our commas, we are now on the next item.
if (Item < Commas.size() && State.NextToken->Previous == Commas[Item]) {
if (!State.NextToken->isTrailingComment()) {
ExtraSpaces += Format->ColumnSizes[Column] - ItemLengths[Item];
++Column;
}
++Item;
}
if (Column == Format->Columns || State.NextToken->MustBreakBefore) {
Column = 0;
NewLine = true;
}
// Place token using the continuation indenter and store the penalty.
Penalty += Indenter->addTokenToState(State, NewLine, DryRun, ExtraSpaces);
}
return Penalty;
}
unsigned CommaSeparatedList::formatFromToken(LineState &State,
ContinuationIndenter *Indenter,
bool DryRun) {
// Formatting with 1 Column isn't really a column layout, so we don't need the
// special logic here. We can just avoid bin packing any of the parameters.
if (Formats.size() == 1 || HasNestedBracedList)
State.Stack.back().AvoidBinPacking = true;
return 0;
}
// Returns the lengths in code points between Begin and End (both included),
// assuming that the entire sequence is put on a single line.
static unsigned CodePointsBetween(const FormatToken *Begin,
const FormatToken *End) {
assert(End->TotalLength >= Begin->TotalLength);
return End->TotalLength - Begin->TotalLength + Begin->ColumnWidth;
}
void CommaSeparatedList::precomputeFormattingInfos(const FormatToken *Token) {
// FIXME: At some point we might want to do this for other lists, too.
if (!Token->MatchingParen ||
!Token->isOneOf(tok::l_brace, TT_ArrayInitializerLSquare))
return;
// In C++11 braced list style, we should not format in columns unless they
// have many items (20 or more) or we allow bin-packing of function call
// arguments.
if (Style.Cpp11BracedListStyle && !Style.BinPackArguments &&
Commas.size() < 19)
return;
// Limit column layout for JavaScript array initializers to 20 or more items
// for now to introduce it carefully. We can become more aggressive if this
// necessary.
if (Token->is(TT_ArrayInitializerLSquare) && Commas.size() < 19)
return;
// Column format doesn't really make sense if we don't align after brackets.
if (Style.AlignAfterOpenBracket == FormatStyle::BAS_DontAlign)
return;
FormatToken *ItemBegin = Token->Next;
while (ItemBegin->isTrailingComment())
ItemBegin = ItemBegin->Next;
SmallVector<bool, 8> MustBreakBeforeItem;
// The lengths of an item if it is put at the end of the line. This includes
// trailing comments which are otherwise ignored for column alignment.
SmallVector<unsigned, 8> EndOfLineItemLength;
bool HasSeparatingComment = false;
for (unsigned i = 0, e = Commas.size() + 1; i != e; ++i) {
// Skip comments on their own line.
while (ItemBegin->HasUnescapedNewline && ItemBegin->isTrailingComment()) {
ItemBegin = ItemBegin->Next;
HasSeparatingComment = i > 0;
}
MustBreakBeforeItem.push_back(ItemBegin->MustBreakBefore);
if (ItemBegin->is(tok::l_brace))
HasNestedBracedList = true;
const FormatToken *ItemEnd = nullptr;
if (i == Commas.size()) {
ItemEnd = Token->MatchingParen;
const FormatToken *NonCommentEnd = ItemEnd->getPreviousNonComment();
ItemLengths.push_back(CodePointsBetween(ItemBegin, NonCommentEnd));
if (Style.Cpp11BracedListStyle &&
!ItemEnd->Previous->isTrailingComment()) {
// In Cpp11 braced list style, the } and possibly other subsequent
// tokens will need to stay on a line with the last element.
while (ItemEnd->Next && !ItemEnd->Next->CanBreakBefore)
ItemEnd = ItemEnd->Next;
} else {
// In other braced lists styles, the "}" can be wrapped to the new line.
ItemEnd = Token->MatchingParen->Previous;
}
} else {
ItemEnd = Commas[i];
// The comma is counted as part of the item when calculating the length.
ItemLengths.push_back(CodePointsBetween(ItemBegin, ItemEnd));
// Consume trailing comments so the are included in EndOfLineItemLength.
if (ItemEnd->Next && !ItemEnd->Next->HasUnescapedNewline &&
ItemEnd->Next->isTrailingComment())
ItemEnd = ItemEnd->Next;
}
EndOfLineItemLength.push_back(CodePointsBetween(ItemBegin, ItemEnd));
// If there is a trailing comma in the list, the next item will start at the
// closing brace. Don't create an extra item for this.
if (ItemEnd->getNextNonComment() == Token->MatchingParen)
break;
ItemBegin = ItemEnd->Next;
}
// Don't use column layout for lists with few elements and in presence of
// separating comments.
if (Commas.size() < 5 || HasSeparatingComment)
return;
if (Token->NestingLevel != 0 && Token->is(tok::l_brace) && Commas.size() < 19)
return;
// We can never place more than ColumnLimit / 3 items in a row (because of the
// spaces and the comma).
unsigned MaxItems = Style.ColumnLimit / 3;
std::vector<unsigned> MinSizeInColumn;
MinSizeInColumn.reserve(MaxItems);
for (unsigned Columns = 1; Columns <= MaxItems; ++Columns) {
ColumnFormat Format;
Format.Columns = Columns;
Format.ColumnSizes.resize(Columns);
MinSizeInColumn.assign(Columns, UINT_MAX);
Format.LineCount = 1;
bool HasRowWithSufficientColumns = false;
unsigned Column = 0;
for (unsigned i = 0, e = ItemLengths.size(); i != e; ++i) {
assert(i < MustBreakBeforeItem.size());
if (MustBreakBeforeItem[i] || Column == Columns) {
++Format.LineCount;
Column = 0;
}
if (Column == Columns - 1)
HasRowWithSufficientColumns = true;
unsigned Length =
(Column == Columns - 1) ? EndOfLineItemLength[i] : ItemLengths[i];
Format.ColumnSizes[Column] = std::max(Format.ColumnSizes[Column], Length);
MinSizeInColumn[Column] = std::min(MinSizeInColumn[Column], Length);
++Column;
}
// If all rows are terminated early (e.g. by trailing comments), we don't
// need to look further.
if (!HasRowWithSufficientColumns)
break;
Format.TotalWidth = Columns - 1; // Width of the N-1 spaces.
for (unsigned i = 0; i < Columns; ++i)
Format.TotalWidth += Format.ColumnSizes[i];
// Don't use this Format, if the difference between the longest and shortest
// element in a column exceeds a threshold to avoid excessive spaces.
if ([&] {
for (unsigned i = 0; i < Columns - 1; ++i)
if (Format.ColumnSizes[i] - MinSizeInColumn[i] > 10)
return true;
return false;
}())
continue;
// Ignore layouts that are bound to violate the column limit.
if (Format.TotalWidth > Style.ColumnLimit && Columns > 1)
continue;
Formats.push_back(Format);
}
}
const CommaSeparatedList::ColumnFormat *
CommaSeparatedList::getColumnFormat(unsigned RemainingCharacters) const {
const ColumnFormat *BestFormat = nullptr;
for (SmallVector<ColumnFormat, 4>::const_reverse_iterator
I = Formats.rbegin(),
E = Formats.rend();
I != E; ++I) {
if (I->TotalWidth <= RemainingCharacters || I->Columns == 1) {
if (BestFormat && I->LineCount > BestFormat->LineCount)
break;
BestFormat = &*I;
}
}
return BestFormat;
}
} // namespace format
} // namespace clang