Syntax error while trying to parse regex in raw string

[Korny0301]

Description

Sonar-scanner can't parse a specific regular expression that is defined in a raw string. When using capturing groups with round brackets () inside the regular expression, the closing bracket of a capturing group can not be placed just in front of the very last closing bracket of the raw string.

Steps to reproduce the problem

Create c++ file with regular expression:

#include <regex>
constexpr auto rgxStr = R"(([A-Z\d]))";

Run sonar-scanner with this file as inclusion (e.g. sonar-scanner -Dsonar.inclusion=test.cpp).

Expected behavior

Line with defined regular expression in raw string should be parsed without syntax error.

Actual behavior

Expression can't be parsed. Sonar-scanner resolves in error:
C++ skip parser error
C++ Parser can't read code. Declaration is skipped.

Known workarounds

Avoid closing bracket immediately before the very last closing bracket that is used to finish the raw string. You could e.g. use a dollar sign before the very last closing bracket:

constexpr auto rgxStr = R"(([A-Z\d])$)";

LOG file

10:27:42.924 DEBUG: finished preprocessing '/opt/wsp/project/test.cpp'
10:27:42.935 DEBUG: [/opt/wsp/project/test.cpp:2]: syntax error: const auto rgxStr = R "(([A-Z\d]\d{4}))"

Related information

cxx plugin version: 1.2.0
SonarQube version: 7.2.1 (14109)

[guwirth]

Hi @Korny0301,

thanks for your feedback

Raw string literal. Used to avoid escaping of any character. Anything between the delimiters becomes part of the string:

R "delimiter( raw_characters )delimiter"

• delimiter: A character sequence made of any source character but parentheses, backslash and spaces (can be empty, and at most 16 characters long)

Your sample: R"(([A-Z\d]))"

• empty / delimiter
• detection of )" seems to be wrong in this case (maybe we have to count opening / closing brackets)
• other possible workaround should be to set delimiter e.g. R"regex(([A-Z\d]))regex"

Code reading raw strings is here: https://github.com/SonarOpenCommunity/sonar-cxx/blob/master/cxx-squid/src/main/java/org/sonar/cxx/channels/StringLiteralsChannel.java#L90

Regards,

[Korny0301]

Hi @guwirth,

thanks for your response.

I already tried to use an additional delimiter but without success. The raw string R"regex(([A-Z\d]))regex" results in a syntax error as well.

Best regards

[guwirth]

Hi @Korny0301, fixed! You can try with https://ci.appveyor.com/project/SonarOpenCommunity/sonar-cxx/branch/master/artifacts. Regards

[Korny0301]

I tried the new version and it works now. Thanks for fixing!!