Urgency of having Mercurial better presented (in part. new website)

[PIERRE AUGIER]

Hi,

I'm in a situation in which I'd like to propose to colleagues of other institutions and universities that a new project (related to in-person Python trainings in French universities) would be hosted on foss.heptapod.net with Mercurial.

Indeed, we would need a neutral host, which would not be affiliated to a particular institution (not the UGA - my university - Gitlab instance, and not others). Of course, I don't want to use Github and I know other interested people agree. So foss.heptapod.net would make sense.

A new open-source project using Mercurial is a crazy proposition in 2025 and I'm sure that there will be oppositions since they are all Git users. But I want to be able to at least propose this possibility and that these colleagues have a look to modern Mercurial.

However, before doing this proposition I really need that the official Mercurial website would be updated or replaced by my new version.

I will have to propose this soon, something like in few days.

So would it be possible to go forward on the subject of the Mercurial website?

Then there is the issue of installation and setup of Mercurial + evolve for people that do not know Mercurial. Currently the situation is fine on Unix (pipx is ok for my case which involves Python devs) but really bad on Windows, where standard modern solutions do not work (and the Windows installers currently available are not much better). Work is being done to improve the situation but in my point of view it's going very slowly. The state of my personal todo list on Mercurial (http://legi.grenoble-inp.fr/people/Pierre.Augier/mercurial-plan-in-2025.html) is a bit depressing.

PS: We do not care about the language when we use a software, but using a Python app (Mercurial) for a projet on Python would still make sense. IMHO, the Mercurial project could gain at communicating more towards the Python community, which is still huge and healthy.

PS2: For mainstream open-source usage, the main issue of Mercurial is not Python or performance (also because a lot of efforts has been spent on that). The main problem is that Mercurial is not natively Git/Github compatible, in contrast with jj. And that even with hg-git, the UX is not that good, both in terms of memory (2 local repos) and overall in terms of incompatibility with modern Mercurial workflow involving topics, evolve and absorb. Getting modern Mercurial natively compatible with Git servers would be a game changer, even if it is only for the "simple" things without collaborative history edition (which is of course incompatible with a Git server). IMHO, it should become a priority if Mercurial wants to regain popularity and increase its user base.

--
Pierre Augier - CR CNRS                 http://www.legi.grenoble-inp.fr
LEGI (UMR 5519) Laboratoire des Ecoulements Geophysiques et Industriels
BP53, 38041 Grenoble Cedex, France                tel:+33.4.56.52.86.16