Arch Linux User Repository

@MarsSeed I admit I'm a bit confused. Mind showing me what you'd like the PKGBUILD to look like, and if it still builds I can update it? I'm not sure either what you mean by "fixed version" vs "VCS type" packages. 4 years ago I had a problem, I fixed the problem, and I figured out just enough of the packaging system to distribute the fix. I'm rather confused why suddenly there's a problem with the solution, when everything still seems to work fine.

I'm not persuaded it's worth more than the time I've already spent to figure this out again when everything still works. What you're suggesting seems like a lot of work just to acquire context (which you already have, so please feel free to share the solution to the problem you've reported) to solve a problem I don't seem to have. Could you be having an X-Y problem, and hope if I solve Y then X will disappear?

And also you have to pin the main source to a specific commit or tag, not just declare the repo or a branch of it. As this package is a fixed-version one, not a VCS type.

Populating the source=() array is mandatory.

If you need to use Git submodules, please see the relevant section of the Arch VCS guidelines.

@MarsSeed I'm a bit confused about what you're asking - spent half an hour digging into the AUR and PKGBUILD documentation, and as far as I can tell I've not violated any spec and the package still builds fine. Mind making a pull request or giving some examples of what you'd like to see? I'll remove the flag in the meantime.

Please declare the source repos in the PKGBUILD source=() array.

Ah, gmail spam filter filed away your message. My bad, reading it now. Sorry I didn't notice it before.

Ah, node event-stream takeover by someone malicious. Gotcha. No, my intentions are simple. I've worked on speech-dispatcher in the past, kttsd which I renamed to Jovie, then killed in favor of QtSpeech library. Lately I've been trying to get pico-tts resurrected from the binary /kind of source package that google threw over the wall at some point years ago when they added it to ubuntu repositories then never did anything with it since. I like to use it myself and think it could be one of the main tts systems on linux if the licensing issues get sorted out.

Hmm, I didn't see any email, which address did you send it to? I agree piping to aplay or paplay makes it much easier to use. Looking through what you've got that enables that to do similarly. So far I have mostly been trying to fix the licensing issues by reverse engineering the picoloadfoo.exe into picoloadfoo.py scripts to generate the lingware from the lingware text files. Left todo are:

picoloadlex.exe picoloadfst.exe dt2pkb.exe pdf2pkb.exe

Also what's the event-stream version 3.3.6 you've referred to? I'm not familiar with it.

Gave a more detailed reply in email. Essentially open to the idea, but leery of what happened with "event-stream version 3.3.6", and the repo linked does not appear to support use in with pipes - a key feature for me. See for example https://unix.stackexchange.com/q/325019/9583.

Hey I've been trying to revive pico-tts (svox-pico is the name I chose for the update) Check it out here https://github.com/jpwhiting/svox-pico.git if you like. I've got a PKGBUILD here that will pull that from git, build it and put it into place, it's only missing the lingware files because they aren't yet completely built from source text files. Once that's ready I think it will be able to replace this (or this could be updated to pull from/build from there if that's preferred). Anyway, let me know what you think if you are able to check it out. If I get it all working the way I'd like to I'll create a new svox-pico package on AUR instead, but would prefer to update this package instead if possible.