diff options
| author | nick@chromium.org <nick@chromium.org@0039d316-1c4b-4281-b951-d872f2087c98> | 2010-05-28 20:18:20 +0000 |
|---|---|---|
| committer | nick@chromium.org <nick@chromium.org@0039d316-1c4b-4281-b951-d872f2087c98> | 2010-05-28 20:18:20 +0000 |
| commit | d06e3e078e734adbb98909f871e510eddcdd54a5 (patch) | |
| tree | 8fc0b814307a9dec0e157d902a4bb6d442a02fb3 /base/string_util.cc | |
| parent | d87f845659b2e54622850acfc780b7fc07cfed6c (diff) | |
| download | chromium_src-d06e3e078e734adbb98909f871e510eddcdd54a5.zip chromium_src-d06e3e078e734adbb98909f871e510eddcdd54a5.tar.gz chromium_src-d06e3e078e734adbb98909f871e510eddcdd54a5.tar.bz2 | |
Add a method that truncates strings to the end point of a valid UTF8 character, leaving the string's size to be less than or equal to a specified byte size.
BUG=43675
TEST=base/string_util_unittest.cc
Patch contributed by to Jerrica Jones (jerrica@chromium.org).
Review URL: http://codereview.chromium.org/2239007
git-svn-id: svn://svn.chromium.org/chrome/trunk/src@48518 0039d316-1c4b-4281-b951-d872f2087c98
Diffstat (limited to 'base/string_util.cc')
| -rw-r--r-- | base/string_util.cc | 35 |
1 files changed, 35 insertions, 0 deletions
diff --git a/base/string_util.cc b/base/string_util.cc index 0269ba1..e36ae51 100644 --- a/base/string_util.cc +++ b/base/string_util.cc @@ -483,6 +483,41 @@ bool TrimString(const std::string& input, return TrimStringT(input, trim_chars, TRIM_ALL, output) != TRIM_NONE; } +void TruncateUTF8ToByteSize(const std::string& input, + const size_t byte_size, + std::string* output) { + if (byte_size > input.length()) { + *output = input; + return; + } + + int32 truncation_length = static_cast<int32>(byte_size); + int32 char_index = truncation_length - 1; + const char* cstr = input.c_str(); + + // Using CBU8, we will move backwards from the truncation point + // to the beginning of the string looking for a valid UTF8 + // character. Once a full UTF8 character is found, we will + // truncate the string to the end of that character. + while (char_index >= 0) { + int32 prev = char_index; + uint32 code_point = 0; + CBU8_NEXT(cstr, char_index, truncation_length, code_point); + if (!base::IsValidCharacter(code_point) || + !base::IsValidCodepoint(code_point)) { + char_index = prev - 1; + } else { + break; + } + } + + DCHECK(output != NULL); + if (char_index >= 0 ) + *output = input.substr(0, char_index); + else + output->clear(); +} + TrimPositions TrimWhitespace(const std::wstring& input, TrimPositions positions, std::wstring* output) { |