fix #13362: Misra report on command line #7096
Conversation
|
|
||
| const std::vector<checkers::MisraInfo> &info = checkers::misraC2012Rules; | ||
| const auto it = std::find_if(info.cbegin(), info.cend(), [&](const checkers::MisraInfo &i) { | ||
| return i.a == std::stoi(components[0]) && i.b == std::stoi(components[1]); |
There was a problem hiding this comment.
std::stoi twice for each entry in info..
| @@ -1417,6 +1419,125 @@ void CppCheck::executeAddons(const std::string& dumpFile, const FileWithDetails& | |||
| } | |||
| } | |||
|
|
|||
| void CppCheck::setClassification(ErrorMessage &errMsg) const { | |||
There was a problem hiding this comment.
The ultimate goal would be that we only have functions to determine classification and guideline in one place. Right now we have it in both lib and gui.
There was a problem hiding this comment.
There is a use case in the GUI that the report type is changed after the analysis. Therefore the GUI needs to be able to take existing results and determine classification and guideline.
The GUI doesn't store the original ErrorMessage objects though as far as I remember. I therefore think it would be preferable if these functions did not use the whole ErrorMessage object.
| @@ -480,7 +480,11 @@ std::string ErrorMessage::toXML() const | |||
| tinyxml2::XMLPrinter printer(nullptr, false, 2); | |||
| printer.OpenElement("error", false); | |||
| printer.PushAttribute("id", id.c_str()); | |||
| if (!guideline.empty()) | |||
There was a problem hiding this comment.
You will need to tweak ErrorLogger::serialize and ErrorLogger::deserialize also otherwise it will not always work when running multiple threads with -j2
| @@ -1504,6 +1625,9 @@ void CppCheck::executeAddons(const std::vector<std::string>& files, const std::s | |||
| } | |||
| errmsg.file0 = file0; | |||
|
|
|||
| setGuideline(errmsg); | |||
There was a problem hiding this comment.
It's possible to adjust the guideline and classification here directly. It would also be possible to do it a bit later.
One advantage with doing it later would be that it avoids the serialization/deserialization problem.
However if we do it here then we should tweak the GUI.. it means the GUI doesn't need to determine the classification/guideline anymore.
WIP. Will add tests and HTML output.