Geoff Reedy

It seems the most recent release that launches on macOS 10.13 is v0.8.943

Alternatively the directory layout could remain the same and the downloads could be de-duped by the sha in the library JSON. When the same file is present in other bundles/titles...

Instead of totally dropping the timestamp, antlr could be updated to support [SOURCE_DATE_EPOCH](https://reproducible-builds.org/docs/source-date-epoch/) for the build-time-stamp. This should leave any existing use-cases for the build-stamp undisturbed but make it easy...

I wrote `section’` because I thought this more general IH was going to be needed; turned out it wasn’t. I also left all the old stuff in there but commented...

I tracked this down to the criteria used for dependency resolution in opam-nix. The current default criteria is `-count[version-lag,request],-count[version-lag,changed],-count[avoid-version,request]`. Specifically, the position of the `-count[avoid-version,request]` leads to this for my...

I ran in to this problem and the core issue is that the not-valid-after timestamp in the JWT is computed by adding the validity period to the current time since...

Oh, I see now that this just happens with the sertop drivers. With coqc_time there is no error and it renders the HTML as I'd expect. So perhaps it's not...

It's not an urgent need. I'm using alectryon with sphinx to build documentation of my development and for now I'll use the coqc_time driver. I've got plenty of time before...

Does this strategy work with regular opam? I would think you’d have the same problem with that setup. In other words I don’t think this would be a opam-nix specific...

I thought to try changing the mode for `Decision` to `+` reasoning that the available decision instances shouldn't be used to determine what is being decided. I thought this would...