About my Gantt pro PRs

[WMP]

Hi, I’d like to explain the PRs I’ve submitted to Kaneo.

First, one note about this message: I wrote the original version in Polish and used ChatGPT to translate it into English.

Unlike this text, my PRs are generated 100% by Claude Opus 4.8, which acts as the orchestrator, and Sonnet 5, which acts as the programmer.

I’m a systems and platform architect. I’ve written quite a lot in Python and Bash before, but languages from the *Script family have always been unfamiliar to me.

The company I work for gave me the responsibility of leading a very large migration project from our own data centers to cloud environments. It is a huge topic, and we need good, customizable software to manage the project schedule.

My manager mentioned Microsoft solutions such as Project or Planner, but I need something that supports MCP, OIDC, and does not have artificial limitations.

I liked Kaneo, but after a while I noticed that it was missing simple dependency visualization on the Gantt chart (#1785).

A bit later, AI pointed out that the backend and API already allow dependencies to be created between projects, but the frontend does not support this (#1784).

Additionally, in this PR, to avoid conflicts between later PRs, I added zoom, displaying dependent tasks from other projects, and the ability to move the timeline left and right by dragging with the mouse.

I searched the open PRs and did not find anything that addressed this, so I asked Claude to prepare CI/CD for building a Docker image and to add these features.

My Kaneo instance running in Kubernetes is already using this image, and I started to really enjoy working with it. Up to this point, the PRs mentioned above have been tested by me on a live Kaneo instance.

Later, however, I started to worry about how my project, which will involve many departments and many people, would cope once the complexity starts to grow.

I gave Claude a simple prompt: you are a professional IT project manager, test Kaneo using this image and tell me what functionality you would be missing in this software.

Unfortunately, I do not have much experience with traditional project management, and I was not previously aware of some of the things Claude pointed out.

Claude listed the functionality that, in its opinion, should be available in Kaneo for me to safely start running my project:

• percentage task progress. This was very useful in smaller projects I had managed before (feat(gantt): professional scheduling — progress, milestones, baselines, timescales, subtask rollup #1788)
• milestones. I’m not sure yet whether I will use them myself, but other people will also have access to this Kaneo workspace, and they may have more experience with Gantt-based planning (feat(gantt): professional scheduling — progress, milestones, baselines, timescales, subtask rollup #1788)
• choosing the timescale: weekly, monthly, quarterly, etc. (feat(gantt): professional scheduling — progress, milestones, baselines, timescales, subtask rollup #1788)
• expanding tasks and showing their subtasks (feat(gantt): professional scheduling — progress, milestones, baselines, timescales, subtask rollup #1788)
• task dependency types such as FS/SS/FF/SF (feat(gantt): dependency types (fs/ss/ff/sf) with lag, cycle detection and drag-to-link #1793)
• prevention of dependency loops between tasks (feat(gantt): dependency types (fs/ss/ff/sf) with lag, cycle detection and drag-to-link #1793)
• creating dependencies by dragging from task edges directly on the Gantt chart (feat(gantt): dependency types (fs/ss/ff/sf) with lag, cycle detection and drag-to-link #1793)
• automatic rescheduling of dependent tasks (feat(gantt): auto-reschedule, critical path, working calendar and date constraints #1794)
• marking tasks as critical, together with their critical golden dependency line (feat(gantt): auto-reschedule, critical path, working calendar and date constraints #1794)
• marking non-working days (feat(gantt): auto-reschedule, critical path, working calendar and date constraints #1794)
• defining when a task may start or must finish, for example because of customer approval, contract end date, etc. (feat(gantt): auto-reschedule, critical path, working calendar and date constraints #1794)
• calculating the number of tasks assigned to a given person within a specific period. This will be a future PR

All of my changes are continuously built into this image:

ghcr.io/wmp/kaneo:claude-gantt-plus-cross-project

At the moment, Claude is running /code-review three times on each branch.

Only after those reviews are complete and CI/CD no longer finds any additional issues to fix will I run my Kaneo instance on the final image and manually test everything during real work.

[WMP]

Currently i have all my features on:

ghcr.io/wmp/kaneo:sha-8a79651

[WMP]

Here is video with all features what I want merge:

https://github.com/WMP/kaneo/blob/claude/new-session-t7ee31/demo-kaneo-tutorial.mp4

[tinsever]

To be blunt, this is a bit frustrating. You've asked Claude what features Kaneo should have, had it generate them, and opened a stack of large, overlapping PRs before we've agreed that any of this belongs in the project. Now the work of understanding, validating, and maintaining all of it lands on us.

Running /code-review three times doesn't resolve that. You need to understand and take responsibility for the code you submit. Manual testing of the full feature set should happen before you ask maintainers to review it.

Kaneo is deliberately simple. "Claude suggested it" is not enough justification to add a scheduling subsystem, especially when you're unsure whether you’ll use some of these features yourself.

Please stop expanding this stack. Bring us a concrete problem, discuss the scope, and submit a focused change you can explain and support. You're welcome to develop whatever you need in your fork, but opening PRs doesn't create an obligation for us to absorb that work.

[WMP]

I understand your approach, and everything is in line with the direction of your discussion. There are currently two PRs open that I have tested and fully understand – #1784 and #1785. The others haven’t been tested yet; they probably don’t contain any bugs, but I can’t guarantee that I fully understand what I’m doing – that’s why they’re marked as drafts. As I need these features straight away, I’ve gone ahead and implemented them. We can discuss them now – if we start talking about them, I’ll feel the need to deliver high-quality code for this project. However, if we don’t discuss them, I won’t even consider opening any more or trying to finalise the current ones, and before long I’ll close the PRs marked as drafts.

So now there are two options – either we start developing Kaneo in the direction I’ve outlined, and you decide to merge #1784 and #1785, or you’re not interested in such an extensive Gantt chart.

[sondr3]

The others haven’t been tested yet; they probably don’t contain any bugs, but I can’t guarantee that I fully understand what I’m doing

Dude, come on. Just have Claude make whatever software you want instead of dropping tens of thousands of lines of code of slop on unsuspecting maintainers.