You signed in with another tab or window. Reload to refresh your session.You signed out in another tab or window. Reload to refresh your session.You switched accounts on another tab or window. Reload to refresh your session.Dismiss alert
Copy file name to clipboardExpand all lines: README.md
+7-3Lines changed: 7 additions & 3 deletions
Display the source diff
Display the rich diff
Original file line number
Diff line number
Diff line change
@@ -18,7 +18,10 @@ Results from analysis below, also summarized in c\_testing\_slides.pdf.
18
18
More GCC flags which we could have added (would have detected some additional bugs): _"-Wshadow -Wundef -Og"_,
19
19
additionally, GCC can provide runtime checks with e.g. _"-fsanitize=undefined"_[1] (not sure how this compares to the Valgrind suite).
20
20
21
-
I can also add that Cppcheck looks somewhat useful and found _one_ bug not discovered by any of the other tools. (Cppcheck is FOSS)
21
+
On later releases GCC (>=10) also provides some static analysis functionality with **_"-fanalyzer"_**[2]. On this code base, GCC (12.1.0) -fanalyzer finds only two issues, namely the double frees in _positives.c_. No FPs.
22
+
23
+
24
+
I can also add that **Cppcheck** looks somewhat useful and found _one_ bug not discovered by any of the other tools. (Cppcheck is FOSS)
22
25
Cppcheck was not evaluated below, so I provide a very brief summary here,
23
26
24
27
```
@@ -42,7 +45,7 @@ Cppcheck 1.82
42
45
Example (if included cmake file doesn't work): cppcheck . --enable=all --inconclusive
43
46
```
44
47
45
-
SonarCloud (SonarQube) is another tool not evaluated here, which detected some interesting stuff when tested on another code base.
48
+
SonarCloud (**SonarQube**) is another tool not evaluated here, which detected some interesting stuff when tested on another code base (similar to CodeQL few hits, but also very few FP -> that's usable).
46
49
(It is proprietary but free to use for open-source software, https://sonarcloud.io/)
47
50
48
51
There seems to be a never ending list of static analysis tools (especially if we include proprietary and/or expensive products),
@@ -52,7 +55,8 @@ duplicates and FPs is concerning. Only choosing one, the tool of choice would be
0 commit comments