greatly improved Settings::loadCppcheckCfg() error handling
#5712
Conversation
|
Requires #5704 to be merged first. |
firewave
force-pushed
the
loadcfg
branch
4 times, most recently
from
1b827e0
to
6ef6767
Compare
…to latest dev version (#5710) The lastest release of picojson does not support creation of JSONs, so we need to switch to the current dev version.
The tests which are failing were in introduced in #5712. Those were not included in #5710 which updated `picojson` that resulted in the different lines being reported in the error messages. I also added some checks to `ScopeFile` which will indicate that a temporary file already exists highlighting multi-threading issues and leftover files from previously aborted testruns. I ran into his while looking into these failing tests
| mSettings.loadCppcheckCfg(); | ||
| return startsWith(mSettings.cppcheckCfgProductName, "Cppcheck Premium"); | ||
| bool CmdLineParser::isCppcheckPremium() { | ||
| Settings settings; |
There was a problem hiding this comment.
@firewave May I ask you why do we need to create the local settings here?
There was a problem hiding this comment.
So we do not accidentally load the config multiple times times. That should be done in a deterministic way once.
There was a problem hiding this comment.
Okay agree on that one. But you only creating local variable so it's not working.
adding static to it will help though.
There was a problem hiding this comment.
But you only creating local variable so it's not working.
What is not working? We only need to compare s single string from the object. Also this is only done during the parsing of the command-line options.
adding
staticto it will help though.
What? The method or the object?
There was a problem hiding this comment.
What? The method or the object?
Both but as you've already marked the method you need to add it only here.
There was a problem hiding this comment.
Both but as you've already marked the method you need to add it only here.
On the object it would not help as it still requires the function call.
I think I did look at the method but I am not sure. It is also just a limited usage.
| #endif | ||
| // cppcheck-suppress knownConditionTrueFalse | ||
| if (fileName.empty()) { | ||
| fileName = Path::getPathFromFilename(exename) + cfgFilename; |
There was a problem hiding this comment.
What is not working? We only need to compare s single string from the object. Also this is only done during the parsing of the command-line options.
loadCppcheckCfg() use member variable exename for storing relative path and it's empty if you create setting locally
There was a problem hiding this comment.
loadCppcheckCfg()use member variableexenamefor storing relative path and it's empty if you create setting locally
Oh right. That again. I ran into his in another context as well. Will take a look later today.
There was a problem hiding this comment.
If I may. I'm already working on it in a separate ticket and was just asking you to help me clarify your thought process behind this particular change.
There was a problem hiding this comment.
I see. I would have just hot fixed this this for now.
As it would have also tied into my upcoming changes to fix https://trac.cppcheck.net/ticket/12240 and #4424 I would have done any more refactoring within that.
This also fixes the issue that
cppcheck.cfgis no longer being loaded from executable path. That was introduced by #5704.