`--ignore` does not appear to work for directories

[zfields]

Directories specified in the --ignore list still appear in the result list.

$ scancode --version
ScanCode version: 32.3.3
ScanCode Output Format version: 4.0.0
SPDX License list version: 3.26

I tried the following basic invocations...

scancode --ignore build/* --verbose --json-pp - .

The way it evaluated the glob without parenthesis in the options was interesting...

"options": {
        "input": [
          "build/CMakeFiles",
          "build/CTestTestfile.cmake",
          "build/DartConfiguration.tcl",
          "build/Makefile",
          "build/Testing",
          "build/_deps",
          "build/cmake_install.cmake",
          "build/compile_commands.json",
          "build/generated-includes",
          "build/libnote_c.a",
          "build/libnote_c.so",
          "build/test",
          "."
        ],
        "--ignore": [
          "build/CMakeCache.txt"
        ],
        "--json-pp": "-",
        "--verbose": true
      },

I tried putting parenthesis in around the ignore, then it captures the ignore literally, and gets the input correct (but that results in no filtering)...

scancode --ignore "build/*" --verbose --json-pp - .

but I still see these in my output...

$ scancode --ignore "note-c/build/*" --verbose --json-pp - .
 Setup plugin: pre_scan:ignore...
 Setup plugin: output:json-pp...
Collect file inventory...
Run pre-scans...
 Run pre-scan: ignore...
Save scan results...
 Save scan results as: json-pp...
{
  "headers": [
    {
      "tool_name": "scancode-toolkit",
      "tool_version": "32.3.3",
      "options": {
        "input": [
          "."
        ],
        "--ignore": [
          "build/*"
        ],
        "--json-pp": "-",
        "--verbose": true
      },
      "notice": "Generated with ScanCode and provided on an \"AS IS\" BASIS, WITHOUT WARRANTIES\nOR CONDITIONS OF ANY KIND, either express or implied. No content created from\nScanCode should be considered or used as legal advice. Consult an Attorney\nfor any legal advice.\nScanCode is a free software code scanning tool from nexB Inc. and others.\nVisit https://github.com/nexB/scancode-toolkit/ for support and download.",
      "start_timestamp": "2025-03-25T132932.971187",
      "end_timestamp": "2025-03-25T132933.313648",
      "output_format_version": "4.0.0",
      "duration": 0.3424708843231201,
      "message": null,
      "errors": [],
      "warnings": [],
      "extra_data": {
        "system_environment": {
          "operating_system": "linux",
          "cpu_architecture": "64",
          "platform": "Linux-6.8.0-55-generic-x86_64-with-glibc2.36",
          "platform_version": "#57-Ubuntu SMP PREEMPT_DYNAMIC Wed Feb 12 23:42:21 UTC 2025",
          "python_version": "3.11.2 (main, Nov 30 2024, 21:22:50) [GCC 12.2.0]"
        },
        "spdx_license_list_version": "3.26",
        "files_count": 202
      }
    }
  ],
  "files": [
    {
      "path": "note-c",
      "type": "directory",
      "scan_errors": []
    },
...
    {
      "path": "note-c/build",
      "type": "directory",
      "scan_errors": []
    },
    {
      "path": "note-c/build/_deps",
      "type": "directory",
      "scan_errors": []
    },
    {
      "path": "note-c/build/_deps/catch2-build",
      "type": "directory",
      "scan_errors": []
    },
    {
      "path": "note-c/build/_deps/catch2-build/CMakeFiles",
      "type": "directory",
      "scan_errors": []
    },
    {
      "path": "note-c/build/_deps/catch2-build/generated-includes",
      "type": "directory",
      "scan_errors": []
    },
...

truncated for readability

[pombredanne]

@zfields Thanks. looks like a nasty bug... the ignore feature needs love... but have you tried to quote your patterns too? Shell expansion may be at play otherwise?

[zfields]

To answer your question directly, "Yes."

I also tried all manner of tricks I could think of (e.g. single ticks, quotes, regex, shell wildcards, repeating flags with individual directories, etc...). Each one produced different outputs with different errors, none of them successful. All told, I probably spent a day trying everything under the sun, because I was assuming I just didn't understand how it worked.

To report the bug, I selected the format that produced the most reasonable (but still wrong) outcome, hoping it would be the easiest to understand and work from.