IMO this is what "source in the preferred form for making modifications" is all about. GNU themselves give the example that C source generated from a perl script isn't good enough - if the way you change the program yourself is by changing the perl script, then you need to release that perl script.

I don't think it's reasonable to expect an open-source maintainer to create a better development experience for you than the one they use themselves. If you're not up to editing Whitney-style code in a text editor, you should get good.

But conversely, just publishing a "source code dump", when you yourself work in a much richer development environment, is not good enough. If there's a "how to set up your dev environment" wiki page that you point other developers to, that should be part of your open-source release. IDE project definitions, project structure guides - whatever you use for making modifications to the project is what you need to publish. And if as a maintainer you don't make those steps documented and reproducible, well, that's ok, but in the long run you're only hurting yourself.

In particular, e.g. RedHat's policy of publishing only the "final" source code for their patched kernels, and not the git repo with history (which is what they use when developing it internally) is a straight-up GPL violation as far as I'm concerned.