Files are not parsed on dockers codecompass/*

[sebastopol06]

Hi there,

Despite no warning nor error disclaimed during the parsing step, the files stated on docker containers are not parsed. Thus most of the features are not available.

I have tried manual installation, scripts-based installation, and released dockers. Same results.

Is it a known issue or known limitation? Am I missing something? Any help on that would be highly appreciated.

[mcserep]

Hi @sebastopol06,

Thanks for your interest in CodeCompass. For a start, could you share the parsing command (CodeCompass_parser) you have executed?

[sebastopol06]

Thanks for coming back to me.

Say I start from the official docker release web-sqlite and connect to the container with the below command:
$ docker run --rm -ti --env DATABASE=sqlite --env BUILD_TYPE=Release --volume ~/dev/CodeCompass:/CodeCompass --volume ~/dev/<source_code_location>/:/projects/<code_label> -p 8001:8080 codecompass:web /bin/bash

In the container:
# CodeCompass_parser -d "sqlite:database=~/dev/CodeCompass/workspace/<code_label>/data.sqlite" -w ~/dev/CodeCompass/workspace -n <code_label> -i /workspace/<code_label>
And returns the below
CodeCompass_parser: error while loading shared libraries: libmagic.so.1: cannot open shared object file: No such file or directory
which is a duplicate of issue 742. However, reinstalling the parser engine on top of the released image with the tricks mentioned in the issue does not make it.

Is there a new released image drop out planned?

[mcserep]

The web-* docker images indeed contain an ODB installation bug, as documented in #742. I made a PR recently to fix that in #756.

However, the web-* docker images are used for serving the CodeCompass webserver, not for parsing. You should use the runtime-* images for parsing a project, as documented.

Moreover, in your example you aim to execute the following command:

CodeCompass_parser -d "sqlite:database=~/dev/CodeCompass/workspace/<code_label>/data.sqlite" \
  -w ~/dev/CodeCompass/workspace -n <code_label> -i /workspace/<code_label>

However, the ~/dev/ folder is on your host machine based on your example, so it won't be accessible inside the docker container. Instead, you should map it as a volume to the container, see: --volume ~/dev/<source_code_location>/:/projects/<code_label>