Cloc doesn't do a recursive search in Docker container

Question

When I run cloc inside a docker container, it does not seem to recursively search through the given directories, compared to running it stand-alone.

Dockerfile:

FROM python:3.6.2-alpine3.6

VOLUME "/data"
WORKDIR /data

RUN apk --no-cache add cloc=1.72-r2

Running cloc without the docker container I get the following:

cloc src\main\java\ --by-file --unix --report-file=temp.csv
      19 text files.
      19 unique files.
      12 files ignored.
Wrote temp.csv

When running it with the docker container the following happens:

docker run --rm -it -v C:\repos\code-repository\:/data cloc-image cloc src/main/java --by-file --unix --report-file=/data/temp2.csv
       0 text files.
       0 unique files.
       2 files ignored.

Any ideas? I have:

Checked the rights of the user in the docker container (root).
Using ash I checked inside the container if the volume was correctly mapped, and all files were present.
Checked if the version of cloc inside the container was indeed the same as the local installation (both 1.72).

EDIT 1:

Interesting finding, this behaviour only shows on Windows, the same Dockerfile/container works fine on a linux machine.

You might `docker run ... cloc-image ls -l` and `... pwd` to get a little more context. — David Maze, Aug 20 '18 at 11:37
I did david, the working directory is, as expected, /data. All files are owned by root and `whoami` shows the user as root as well. — Oromë, Aug 20 '18 at 12:23

score 1 · Answer 1 · answered Jan 24 '19 at 15:24

I encountered the same problem on a debian:stretch-based container running on Docker for Windows, where the volume cloc was running against was a local directory. The solution was to add --follow-links to cloc:

$ cloc .

# Only returned results of top-level directory

Fix:

$ cloc --follow-links .

# Included nested files

Cloc doesn't do a recursive search in Docker container

1 Answers1