parse error for C code

[DrewDennison]

https://sgrep.live/Kl6

//low-end processors are slow on division
if (di->block_size == 1<<6 ){ //sha256
  nblocks = len >> 6;
  nbytes = len & 0xFFFFffC0;
} else if(di->block_size == 1<<7 ){ //sha512
  nblocks = len >> 7;
  nbytes = len & 0xFFFFff80;
} else {
  nblocks = len / di->block_size;
  nbytes = nblocks * di->block_size;
}

results in

{
  "matches": [],
  "errors": [
    {
      "check_id": "LexicalError",
      "path": "NO FILE INFO YET",
      "start": { "line": -1, "col": 0 },
      "end": { "line": -1, "col": 1 },
      "extra": {
        "message": "Lexical error: unrecognised symbol, in token rule:�",
        "line": "NO LINE"
      }
    }
  ]
}

[DrewDennison]

@aryx do you agree this is a bug? Can we figure out if this is a <1 hour fix or something more

[dlukeomalley]

@DrewDennison Following up, still affecting you?

[dgryski]

This is still an issue. I still get the following error with 0.15.0 running a simple query against the repo github.com/h2o/h2o.

~/go/src/github.com/h2o/h2o $ semgrep -l c -e 'if ($X , $Y) { ... }'
Traceback (most recent call last):
  File "/usr/local/bin/semgrep", line 33, in <module>
    sys.exit(load_entry_point('semgrep==0.15.0', 'console_scripts', 'semgrep')())
  File "/usr/local/Cellar/semgrep/0.15.0/libexec/lib/python3.8/site-packages/semgrep/__main__.py", line 11, in main
    cli()
  File "/usr/local/Cellar/semgrep/0.15.0/libexec/lib/python3.8/site-packages/semgrep/cli.py", line 322, in cli
    semgrep.semgrep_main.main(
  File "/usr/local/Cellar/semgrep/0.15.0/libexec/lib/python3.8/site-packages/semgrep/semgrep_main.py", line 343, in main
    semgrep_errors = [e.into_semgrep_error() for e in semgrep_core_errors]
  File "/usr/local/Cellar/semgrep/0.15.0/libexec/lib/python3.8/site-packages/semgrep/semgrep_main.py", line 343, in <listcomp>
    semgrep_errors = [e.into_semgrep_error() for e in semgrep_core_errors]
  File "/usr/local/Cellar/semgrep/0.15.0/libexec/lib/python3.8/site-packages/semgrep/core_exception.py", line 69, in into_semgrep_error
    with open(self._path, errors="replace") as f:
FileNotFoundError: [Errno 2] No such file or directory: 'NO FILE INFO YET'
~/go/src/github.com/h2o/h2o $

[dgryski]

Here's the smallest reproducer I could find.

int bt1_pop64(int x) { return  0xFllu; }

The C parser doesn't like the llu suffix.

[aryx]

@DrewDennison trying https://semgrep.live/Kl6 gives just "semgrep got an error" but I can't find a way to display the error.
Is it normal?

[ievans]

https://semgrep.dev/s/ZkEw works now! In the original snippet, there was an invalid whitespace character at the start of nblocks = len >> 6; that caused the parser to fail (it caused a real C compiler to fail too).