Skip to content

No test for malformed graph #208

Closed
Closed

Description

@GREENRAT-K405

The current tests for cli only capture two scenarios:

  • file is missing or
  • file is valid
    it lacks to test the cases where the file does exist but the graphml file has corrupted XML.

I propose adding a corrupted XML test

Activity

  1. GREENRAT-K405 commented on Feb 4, 2026

    @GREENRAT-K405
    ContributorAuthor

    @pradeeban , what are your thoughts on this? is this required?

    I'd like to work on this issue!

  2. avinxshKD commented on Feb 7, 2026

    @avinxshKD
    Contributor

    @pradeeban , what are your thoughts on this? is this required?

    I'd like to work on this issue!

    Makes sense ,, I'd suggest we test at least:

    1. Broken XML syntax
    2. Missing required GraphML attributes
    3. Invalid data types in node properties

    @pradeeban want me to draft a PR with a few test cases?

  3. GREENRAT-K405 commented on Feb 7, 2026

    @GREENRAT-K405
    ContributorAuthor

    @avinxshKD I am already working on this issue. Thanks for offering your help but please work on different issue.

  4. avinxshKD commented on Feb 7, 2026

    @avinxshKD
    Contributor

    @avinxshKD I am already working on this issue. Thanks for offering your help but please work on different issue.

    Just to clarify, I was adding suggestions to the discussion, not claiming the issue. All good, I’ll put up a PR for the test cases I mentioned. If thats okay @pradeeban, thanks

  5. pradeeban commented on Feb 8, 2026

    @pradeeban
    Member

    It is a dangerous task. There was a PR I closed: #213.
    concore supports directed hypergraphs.

    If you are simply validating whether a graphml is valid, it should be ok.
    But assuming all the graphml files are produced by concore-editor, I believe we won't have invalid graphml files? Just be cautious with this fix. We do not want to unintentionally introduce bugs while trying to fix a problem that does not exist in practice (since this is handled by concore-editor, unless someone mistakenly tries to run a graphml not created by concore-editor, which is unlikely).

  6. pradeeban commented on Feb 8, 2026

    @pradeeban
    Member

    Closing this since #218 is merged.

Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment

Metadata

Metadata

Assignees

No one assigned

    Labels

    No labels
    No labels

    Type

    No type

    Projects

    No projects

      Milestone

      No milestone

      Relationships

      None yet

      Development

      No branches or pull requests

      Issue actions