Open dholm opened 2 months ago
We suspect that this also impacts skipfile
because our reports now includes all the files in our skipfile
as well.
I encountered the same issue.
Changing report.file.id
to report.file.path
makes it work again.
Apparently, file.id
is the absolute path and file.path
the path with (optional) trimming applied.
I directly edited the contents installed by pip (/<venv-path>/lib/python3.12/site-packages/codechecker_report_converter/report/output/html/html.py
)
Edit:
This might be more robust (i.e., not have unintended side-effects):
diff --git a/tools/report-converter/codechecker_report_converter/report/output/html/html.py b/tools/report-converter/codechecker_report_converter/report/output/html/html.py
index c102199c..85bacfed 100644
--- a/tools/report-converter/codechecker_report_converter/report/output/html/html.py
+++ b/tools/report-converter/codechecker_report_converter/report/output/html/html.py
@@ -309,7 +309,7 @@ class HtmlBuilder:
table_reports = map(lambda data: {
'link': os.path.basename(data['link']),
- 'file-path': data['report']['fileId'],
+ 'file-path': self.files[data['report']['fileId']]['filePath'],
'report-hash': data['report']['reportHash'],
'checker-name': data['report']['checker']['name'],
'checker-url': data['report']['checker']['url'],
Describe the bug
We just upgraded from CodeChecker
6.23.1
to6.24.0
and the--trim-path-prefix
option toCodeChecker parse
appears to have stopped working as the generated HTML pages are using the absolute path to the build folders on our Jenkins machines now.CodeChecker version
To Reproduce
After analyzing the code, run
CodeChecker parse --export html --print-steps --trim-path-prefix "<absolute-path-to-source-code>" --ignore <our-skipfile> --output "<build-dir>/code-checker-report" "<build-dir>/code-checker"
. Steps to reproduce the behaviour:Expected behaviour
In the report generated by
6.23.1
it looks like this:In the report generated by
6.24.0
it looks like this:Desktop (please complete the following information)