Here are some questions that can arise.
What happens if both a coverity.conf file and a polaris.yml file specify a server URL?
- If you entered a URL when you used the Notifications panel to authenticate yourself, then that is the URL that Code Sight uses.
- Otherwise, Code Sight uses the URL in polaris.yml. This is simply because Code Sight reads polaris.yml before it reads the project-specific coverity.conf.
If multiple versions of Coverity Analysis are installed, which version will Code Sight run?
- If polaris.yml is present and specifies a valid version of Coverity Analysis, then Code Sight uses that version.
- Otherwise, Code Sight uses the installation that matches the current project and stream. If these are not known, it uses the most recent version specified in the local, user-specific coverity.conf file.
What if I get a report of a central analysis issue, but Code Sight does not locate the issue in my local source code?
Typically the central analysis server saves path names in a relative format; for example:
../project/source/testme.java
Instead of relative paths, a local instance of Coverity Analysis typically saves file names as absolute paths. The format of an absolute path depends on the operating system. For example:
C:\Users\your-name-here\project\source\testme.java
To resolve this issue, in your project-specific coverity.conf file add an "ide" object,
and then within it create a "path_mapping" object to specify (1) the paths to strip off of the
path names from central analysis, and (2) the local paths to search for.
The object has this format:
"ide": {
"path_mapping": {
"strip_paths": [ "path1", "path2" ],
"search_paths": [ "path3", "path4" ]
}
}, ...
So using the previous example, testing only a single project, the "ide" object might look like
the following JSON code:
"ide": {
"path_mapping": {
"strip_paths": [ "../" ],
"search_paths": [ "C:\Users\your-name-here\" ]
}
}, ...
For additional information, see the “IDESettings” and “PathMapping” sections in the Coverity Desktop Analysis User Guide.