Vincent Parrett

To add to the VS information - you can use Visual Studio Community Edition for commercial use: it's free for individual developers and non-enterprise organizations with fewer than 250 PCs or under $1 million USD in annual revenue. It's trivial to install and manage the license for the community edition - and it does a lot (no db limitations etc). That makes Delphi a very unattractive proposition for people starting out. The only place where Delphi wins out (ignoring the price) is for Desktop Windows applications. For everything else, there are better options. I wish it were not the case, but it's too far behind now to ever be able to compete in many spaces (like web, mobile etc).

FWIW I am not using WSL - running in a windows terminal cmd. I have a simialr batch file - but for some reason it only works half the time. I'll try the sh route as claude code is using bash even under windows.

I gave up trying to get it to compile and run tests, it worked at first with the batch file I created - but then it would randomly fail but claude would just blindly carry on and run the old exe - and then it started using the wrong command line parameters - that is despite me spending ages spelling it all out in claude.md Thanks for the tip about using the separate session - found several issues which point to an incomplete implementation of the algorithm 😢 More to do still

That's exactly what I did - I used various ai's to generate a long markdown document with as much info as I could on the algorithm and current implementations - including code flow analysis - I ran it through several iterations and ai's to refine it before using GPT5 to create the plan which I fed to claude code. I

I still don't know, the unit test project is 38K lines of code - I need to look through algorithm code to look for hard coded values. I have been here 3 times before - with all tests passing, but then I added more with different package names and things failed - that's when I discovered the hard coding of names in some functions.

claude finally managed it after 4 days - If I had done a CS degree maybe I would have learned SAT solvers and figured this out myself. Not to check claude hasn't fecking hard coded values in again to make the tests pass. Oh, and to refactor the 1000 line functions it created!

I did use opus for the plan (till I hit the limit, even on the $100 plan), plus I added to it extensively. it's a complex algorithm - https://nex3.medium.com/pubgrub-2fb6470504f - and the reference implementation is the dart pub package manager - dart -> delphi is difficult due to extreme language differences, memory management etc.

I am using claude code

I am starting to agree - on my 4th iteration with ai implementing an algorithm and for 4th time it started hard coding values in the algorithm to make tests pass 🤦‍♂️ Even with using it to generate unit tests - I am finding unit tests which create competing expectations - so the ai goes around in circles flip floping which bunch of tests fail. I ended up deleting a bunch of tests and focusing on ones I know for sure are valid scenarios. Even then, the code quality and accuracy goes up and down like a yo yo and I am stuck in a loop grrr

Is that a problem? If so, why? Yes - applications should not be installing root certificates into windows - the only scenario where I can see this being needed is when using self signed code signing certificates - which is happening more an more in enterprises for internal applications/scripts - especially ones where internet access is limited.

How do you do that? I tried, I even gave it a batch file to run, but on windows it's still using bash and it fails more often than it succeeds and then it doesn't see the compiler errors and just runs the code anyway. Is this something an mcp could do?

I did exactly that, I have had other ai's refine it - and I had chat write an annotated program flow for the dart (reference) implementation of the algorithm - but eventually it starts to wander off and do it's own thing rather than stick to the algorithm - and twice so far it got to 100% tests passing and when I investigated the 'breakthrough' I discovered it had hard coded values in the algorithm to satisfy the tests. The algorithm is complex - and implementing it in delphi is is not helped by the major differences in the programming languages - there are constructs in Dart that are not easy to replicate in delphi without major design changes.

I am not against ai specific forums - but like others I look at the activity view to find things of interest on a daily basis

I just wasted 2 days with claude and gemini trying to get it to implement the PubGrub dependency solving algorithm that the Dart (pub) and Rust (cargo) package manager uses - in delphi. Round and round in circles, but never quite getting there, until claude gives up (I used up my gemini allowence in an hour) and starts hard coding things to make the unit tests pass 🤦‍♂️ "You're absolutely right. I apologize for taking the wrong approach by hardcoding specific package names and version constraints to satisfy the test scenario. That's not a proper solution - it's just making the code pass a specific test rather than implementing the correct PubGrub algorithm."

Nice - I do need to add some sample apps at some point - this would be a good one!