search

TeamPyOgg / PyOgg

Simple OGG Vorbis, Opus and FLAC bindings for Python
The Unlicense
64 stars 27 forks source link

Automated tests, library loader changes, documentation #39

Closed mattgwwalker closed 4 years ago

mattgwwalker commented 4 years ago

Hi Zuzu-Typ,

I gave up :o/ Here is everything. In one mighty pull request. I've been trying to backport things from my version to yours, but it all just got too hairy.

This pull request contains:

Sorry this is so massive. However with the newly-added automated tests at least you can demonstrate that the code works ;o)

Cheers,

Matthew

mattgwwalker commented 4 years ago

Hi again Zuzu-Typ,

I've added to this pull request macOS wheels:

If you have any questions, please fire away. We can update the documentation if there are any issues that aren't clear.

Cheers,

Matthew

Zuzu-Typ commented 4 years ago

Hi there @mattgwwalker ,

I'm sorry that I didn't have a look at this earlier. It's amazing how much work you put into the library. Frankly, I don't currently have enough time to keep up and I really don't want your work to be wasted, so I would like to ask you if you'd like to become a collaborator of this project (here and on PyPI). If you'd prefer I could also transfer ownership of this repository to you, because I feel like you've done more work for it than I did already.

Please do tell me what you think about it.

Cheers --Zuzu_Typ--

mattgwwalker commented 4 years ago

Hi Zuzu-Typ,

I'd be very happy to be added in as a collaborator. Thanks for asking me.

The project I'm currently working on calls your library at least once every 20ms for hours on end, so I've been very engaged with PyOgg because of it. As a contributor, here are the things I'd like to set up for the PyOgg GitHub repository:

I'd also like it if we did a PyPi release as I think there are a number of newly-added features that developers would find useful.

I was in your position with another project I started and I was very pleased to see the community of developers form around it. It was only a good thing for that project as, just like you, I didn't have time to work on it. However, I also feel it's your baby, and ownership of the project is best left with you.

Cheers,

Matthew

Zuzu-Typ commented 4 years ago

Hi there Matthew,

I'm glad that you decided to join (:

I thought about adding CI to PyOgg, but in my opinion it would only make sense if the binaries for ogg, vorbis, opus, flac, etc. would also be built. Unfortunately most of those libraries don't really work well together and require quite a bit of tweaking to get them to build. That's why I decided not to add CI. Of course, using pre-generated libraries would also be fine, but since I didn't have tests at the time, there was no real reason to incorporate CI.

I haven't used ReadTheDocs before, but it does sound like a good idea. The only thing I dislike about it is that it uses reStructuredText.. I know you can also use markdown, but I'm not sure if I would consider that to be much better. I find most markup languages cumbersome to write, especially when things are automatically formatted. I've considered creating a parser that's able to convert from different markup languages to HTML and back, because there are much better editors out there for HTML. Anyway, unrelated.

Do you have a PyPI account by any chance? Because I could add you as a collaborator there too.

Cheers, --Zuzu_Typ--

mattgwwalker commented 4 years ago

Hi again,

I've accepted the collaboration invitation. Would you like me to merge this pull request myself, or would you prefer to do it?

On the subject of the binaries, this pull request includes pre-built binaries for Windows 32-bit, Windows 64-bit and macOS 64-bit. I agree that automating the build process for all the required libraries would be ideal, but the amount of work required would be very large. Using the pre-built binaries will be satisfactory, if not ideal. Combining the pre-built libraries with the automated tests in this pull request would allow us to implement continuous integration. Would you like me to look into this? I can't make it my top priority but I'd consider it interesting to tick off and it would make future development more reliable.

The documentation I have included in this pull request uses Sphinx with restructured text. It's not the most fun of markup languages, I agree, but from now on, we can just copy markup from what has already been written. Would you like me to look into hooking us into ReadTheDocs?

As for PyPi, my username there is mattgwwalker.

Cheers,

Matthew

Zuzu-Typ commented 4 years ago

Hey Matthew,

We're kinda abusing this PR as a means of communication, aren't we 😅 Anyway, if you want to, you can merge this PR yourself.

I can try setting up CI, as I've done it before, but it has been a while and it might take a few tries to get right. We can set up CI to automatically upload the releases so we don't have to do that manually.

I'm not sure wether or not it would also be possible to generate the docs via CI and upload them to ReadTheDocs.io or wheresoever the documentation would be hosted, as I haven't looked into it yet, but do feel free to try setting it up. We can look into "continuous documentation" later (:

Again, thank you for your time and effort!

Cheers, --Zuzu_Typ--