================
@@ -409,9 +412,9 @@ void SarifDocumentWriter::appendResult(const SarifResult
&Result) {
if (!Result.RelatedLocations.empty()) {
----------------
zeyi2 wrote:
I found a case where repeated notes produce schema-invalid SARIF. With
`-checks='-*,bugprone-*'` on cppcheck's `simplecpp.cpp`, the regular text
output contains the following notes (messages shortened):
```
simplecpp.cpp:2723:5: note:
simplecpp.cpp:2724:5: note: 'int' and 'std::ptrdiff_t' may be implicitly
converted ...
simplecpp.cpp:2723:5: note:
simplecpp.cpp:2725:5: note: 'int' and 'std::size_t' may be implicitly converted
...
```
The exporter turns them into identical `relatedLocations` entries (Other
entries are omitted):
```
"relatedLocations": [
{
"physicalLocation": {
"artifactLocation": { "index": 0, "uri": "file:///.../simplecpp.cpp" },
"region": {
"startLine": 2723, "startColumn": 5,
"endLine": 2723, "endColumn": 6
}
}
},
{
"physicalLocation": {
"artifactLocation": { "index": 0, "uri": "file:///.../simplecpp.cpp" },
"region": {
"startLine": 2723, "startColumn": 5,
"endLine": 2723, "endColumn": 6
}
}
}
]
```
Although the empty messages appear to come from an existing check issue, there
may be other cases that produce repeated notes. IMO it would be better for the
exporter to help handle these diagnostics in a way that still produces valid
SARIF.
https://github.com/llvm/llvm-project/pull/219182
_______________________________________________
cfe-commits mailing list
[email protected]
https://lists.llvm.org/cgi-bin/mailman/listinfo/cfe-commits
