How do I dump gcc warnings into a structured format?

Question

Like many, I build my project with the an abundance of warning flags.
Since not all warning flags are detrimental, the compilation becomes noisy.

Warnings such as "unused variables", "shadowing members in initialization lists", "missing switch defaults", are all important to log, but they create too much clutter during builds, and it is hard to spot the important warnings.

Given a large project, there can be thousands of warnings mixed in with build statements, and parsing though it afterwards becomes burdensome. It's equally undesirable to maintain compiler pragmas and push/pop warnings inside code.

How can I dump compiler warnings in a structured format?
Whether it be XML, JSON, YAML, CSV, is there a way to tell the compiler to dump all emitted warnings? A format like this would allow me to view warnings more efficiently, and sort them by type, file, amount, etc.

Answer 1

GCC 9 added[1] support for outputting warnings and error messages in JSON format, just use -fdiagnostics-format=json option.

Compare the output of

$ gcc-9 -c cve-2014-1266.c -Wall
cve-2014-1266.c: In function ‘SSLVerifySignedServerKeyExchange’:
cve-2014-1266.c:629:2: warning: this ‘if’ clause does not guard... [-Wmisleading-indentation]
  629 |  if ((err = SSLHashSHA1.update(&hashCtx, &signedParams)) != 0)
      |  ^~
cve-2014-1266.c:631:3: note: ...this statement, but the latter is misleadingly indented as if it were guarded by the ‘if’
  631 |   goto fail;
      |   ^~~~

with JSON-formatted one:

[
    {
        "children": [
            {
                "kind": "note",
                "locations": [
                    {
                        "caret": {
                            "column": 3,
                            "file": "cve-2014-1266.c",
                            "line": 631
                        },
                        "finish": {
                            "column": 6,
                            "file": "cve-2014-1266.c",
                            "line": 631
                        }
                    }
                ],
                "message": "...this statement, but the latter is misleadingly indented as if it were guarded by the \u2018if\u2019"
            }
        ],
        "kind": "warning",
        "locations": [
            {
                "caret": {
                    "column": 2,
                    "file": "cve-2014-1266.c",
                    "line": 629
                },
                "finish": {
                    "column": 3,
                    "file": "cve-2014-1266.c",
                    "line": 629
                }
            }
        ],
        "message": "this \u2018if\u2019 clause does not guard...",
        "option": "-Wmisleading-indentation"
    }
]

[1] https://developers.redhat.com/blog/2019/03/08/usability-improvements-in-gcc-9/

Answer 2

• The coala project attempts to make the tools it wraps produce the same structured output formats (see http://wordpress.schuirmann.eu/2016/02/coala-at-fosdem-2016/#comment-123 ). Sadly there doesn't seem to be an already made "bear" to wrap a compiler.
• On a similar note to the above, Ericsson have built a tool called CodeChecker that lets you track a project's clang static analysis warnings over time.
• There is an moribund project called firehose that wrapped the output of gcc and several C static analysis tools.
• clang has -fdiagnostics-format=msvc which can make its output slightly more structured.

Something that might help you in the meantime is to turn those compiler warnings you deem critical into errors using -Werror= so you notice them breaking the build above the noise of the warnings.