Manual QA for cabal-install

Hi everyone,
The cabal team is trying out a new programme to onboard people onto cabal development.
Some contributions that directly impact the command-line interface highly benefit
from manual quality assurance (QA), to supplement the test suite.
This allows to test the contributions we receive in the larger context of cabal-install,
and make sure that high-level behaviours hold on various setups and environments (operating system, CPU architecture, use of XDG, reliability of internet connection, etc).

As such, we are devising a workflow for manual testers, that you can read here: Manual QA · haskell/cabal Wiki · GitHub

The idea is that we collect the PRs that will need manual QA in a ticket board, and people who are new to cabal development can follow the instructions written in the PR or ticket, report their findings (everything behaves as expected or something wrong was found), and we can ship new features with more confidence.

Being a manual QA tester for cabal does not require any kind of long-term commitment,
you can pick a ticket whose PR was merged, compile cabal on your machine, see how it
behaves, and tell us what you see. :slight_smile:

Let yourself known in a ticket or PR that needs manual QA and you will be added to the Manual QA team, so that you can sort the ticket in the QA board.

Let us know if you have any question or remark!

Cheers.

15 Likes

It sounds re-adding ghcup compile cabal might be worthwhile in this context? It would be a simple wrapper that handles git, cabal build and renames binaries automatically based on git describe or detected version in the cabal file.

3 Likes

Sounds like it would be worth it indeed.