[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

Re: [tlaplus] Historical tla2tools.jar rolling builds



This is motivated by my early musings about TLA+ release engineering, and learning nix for personal use. It would be nice to have reproducible builds for the TLA+ tools and especially TLAPS, with transitive dependencies locked, and an easy way to specify all of that. Especially if we are using a rolling release model. Both Nix and Guix offer something of this sort. There's an interesting article in Nature about it through the lens of reproducible research (https://www.nature.com/articles/s41597-022-01720-9). Since TLA+ and TLAPS are used in research projects & papers, we should consider providing researchers an easy format to encode the exact versions their dependencies at the end of the paper, which other people can copy & paste to their local environment to reproduce their results indefinitely.

I'm not 100% sure that keeping all rolling builds around would be important for this purpose but it seems likely; more research is necessary on nix/guix packaging and what free build artifact caches exist.

Andrew
On Monday, January 8, 2024 at 12:45:20 AM UTC+8 Markus Kuppe wrote:
We don't archive those builds currently. However, let's continue the discussion in a GitHub issue, if you believe archiving is necessary.

By the way, I've been using git bisect for finding regressions, in case that's what you’re after.

Markus

> On Jan 7, 2024, at 1:23 AM, Andrew Helwer <andrew...@xxxxxxxxx> wrote:
>
> Primarily I'm interested in the build of every commit of main that is uploaded to the 1.8.0. rolling prerelease; are those archived anywhere or just overwritten?


--
You received this message because you are subscribed to the Google Groups "tlaplus" group.
To unsubscribe from this group and stop receiving emails from it, send an email to tlaplus+unsubscribe@xxxxxxxxxxxxxxxx.
To view this discussion on the web visit https://groups.google.com/d/msgid/tlaplus/c5125934-9ecb-4ae6-82ca-80e346da9441n%40googlegroups.com.