/Users/deen/code/yugabyte-db/src/yb/util/string_trim.cc
Line | Count | Source |
1 | | // Copyright (c) YugaByte, Inc. |
2 | | // |
3 | | // Licensed under the Apache License, Version 2.0 (the "License"); you may not use this file except |
4 | | // in compliance with the License. You may obtain a copy of the License at |
5 | | // |
6 | | // http://www.apache.org/licenses/LICENSE-2.0 |
7 | | // |
8 | | // Unless required by applicable law or agreed to in writing, software distributed under the License |
9 | | // is distributed on an "AS IS" BASIS, WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express |
10 | | // or implied. See the License for the specific language governing permissions and limitations |
11 | | // under the License. |
12 | | // |
13 | | |
14 | | #include "yb/util/string_trim.h" |
15 | | |
16 | | #include <sstream> // for istringstream |
17 | | #include <string> |
18 | | #include <regex> |
19 | | |
20 | | using std::string; |
21 | | using std::vector; |
22 | | using std::istringstream; |
23 | | |
24 | | namespace yb { |
25 | | namespace util { |
26 | | |
27 | 485 | string ApplyEagerLineContinuation(const string& s) { |
28 | 485 | string result; |
29 | 485 | size_t length = s.size(); |
30 | 353k | for (size_t i = 0; i < length; ++i353k ) { |
31 | 353k | while (i > 0 && s[i - 1] == '\\'353k && s[i] == '\n'802 ) { |
32 | | // Remove the previous character (backslash). |
33 | 418 | result.resize(result.size() - 1); |
34 | | // Consume the leading whitespace on the new line. This may be different from how backslash |
35 | | // line continuation works in other places, but this is what we need for some of our expected |
36 | | // test output handling. |
37 | | |
38 | 418 | ++i; // skip the new line |
39 | | // Skip whitespace on the new line. |
40 | 2.78k | while (i < length && std::isspace(s[i], std::locale::classic())) { |
41 | 2.36k | ++i; |
42 | 2.36k | } |
43 | 418 | } |
44 | | |
45 | 353k | if (i < length) { |
46 | 353k | result.push_back(s[i]); |
47 | 353k | } |
48 | 353k | } |
49 | 485 | return result; |
50 | 485 | } |
51 | | |
52 | | namespace { |
53 | | |
54 | 11.6M | size_t CountLeadingSpaces(const string& line) { |
55 | 11.6M | size_t num_spaces = 0; |
56 | 63.0M | for (char c : line) { |
57 | 63.0M | if (c != ' ') |
58 | 11.6M | break; |
59 | 51.4M | num_spaces++; |
60 | 51.4M | } |
61 | 11.6M | return num_spaces; |
62 | 11.6M | } |
63 | | |
64 | | } // anonymous namespace |
65 | | |
66 | 415k | string LeftShiftTextBlock(const std::string& s) { |
67 | 415k | istringstream input(s); |
68 | 415k | vector<string> lines; |
69 | | |
70 | | // Split the string into lines. This could be implemented with boost::split with less data |
71 | | // copying and memory allocation. |
72 | 12.0M | while (!input.eof()) { |
73 | 11.6M | lines.emplace_back(); |
74 | 11.6M | getline(input, lines.back()); |
75 | 11.6M | } |
76 | | |
77 | 415k | size_t min_leading_spaces = std::numeric_limits<int>::max(); |
78 | 11.6M | for (const string& line : lines) { |
79 | 11.6M | const auto num_spaces = CountLeadingSpaces(line); |
80 | | // We're not counting empty lines when calculating the minimum number of leading spaces. |
81 | | // TODO: we're counting all-space lines as empty but not if they have e.g. tab chracters. |
82 | 11.6M | if (num_spaces != line.size() && num_spaces < min_leading_spaces11.6M ) { |
83 | 415k | min_leading_spaces = num_spaces; |
84 | 415k | } |
85 | 11.6M | } |
86 | | |
87 | 415k | string result; |
88 | 415k | bool need_newline = false; |
89 | 11.6M | for (const string& line : lines) { |
90 | 11.6M | if (need_newline) { |
91 | 11.1M | result.push_back('\n'); |
92 | 11.1M | } |
93 | 11.6M | need_newline = true; |
94 | 11.6M | if (min_leading_spaces <= line.size()) { |
95 | 11.6M | result += line.substr(min_leading_spaces, line.size() - min_leading_spaces); |
96 | 11.6M | } |
97 | 11.6M | } |
98 | 415k | return result; |
99 | 415k | } |
100 | | |
101 | 445 | std::string TrimCppComments(const std::string& s) { |
102 | 445 | static const std::regex kCppCommentRE = std::regex("\\s*//[^\n]+"); |
103 | 445 | return std::regex_replace(s, kCppCommentRE, ""); |
104 | 445 | } |
105 | | |
106 | 37 | std::string TrimTrailingWhitespaceFromEveryLine(std::string s) { |
107 | 37 | auto write_it = s.begin(); |
108 | 37 | auto first_it_to_delete = s.begin(); |
109 | 944 | for (auto i = s.begin(); i != s.end();) { |
110 | 907 | auto ch = *i; |
111 | 907 | ++i; |
112 | 907 | if (ch == '\r' || ch == '\n') { |
113 | 63 | *write_it++ = ch; |
114 | 63 | first_it_to_delete = i; |
115 | 844 | } else if (!std::isspace(ch)) { |
116 | 1.46k | while (first_it_to_delete != i) { |
117 | 796 | *write_it++ = *first_it_to_delete++; |
118 | 796 | } |
119 | 670 | first_it_to_delete = i; |
120 | 670 | } |
121 | 907 | } |
122 | 37 | s.erase(write_it, s.end()); |
123 | 37 | return s; |
124 | 37 | } |
125 | | |
126 | | } // namespace util |
127 | | } // namespace yb |