On Monday, September 29, 2025 at 8:01:58 AM UTC-7 [email protected] wrote:
On Sun, Sep 28, 2025 at 10:52 AM Kwankyu Lee <[email protected]> wrote: > > On Friday, September 26, 2025 at 7:47:25 AM UTC+9 John H Palmieri wrote: > > Here are some recent occurrences in Sage development: > > 1. The documentation is not built by default. > > > Previously "make" built doc. Now it doesn't. This change affects all developers using sage-the-distro. And it is a positive change, as one does building of the code more often than building of the code and the docs. I understand that opinion, but it should have been discussed by the Sage community before being implemented. > As such, It should have been announced in sage-devel. I don't know if it was. And this is fine. You were not visibly doing anything Sage-related for at least half a year, after a minor scandal, with you demanding more "respect" legacy of a developer removed from the project for major misconduct. I don't know if I should read this as an attack on Kwankyu or a more general message, conveying something like, if you don't post here much, if you are not very involved in Sage development, then you should not express your opinion here. Either way: good job! This is a great way to make the Sage community more open and welcoming! Thanks! > > 2. There has been the assertion that Conda is the recommended approach for compiling from source. > > > This is a major change. This should be approved in sage-devel before it is officially adopted (appear in docs). No, why? It's just how things are at the moment. Conda is the easiest cross-platform way to install Sage, full stop. It's hard to argue otherwise. If someone writes that 2+2==4 in the manual, will you demand an approval for this too? There is a difference between saying "conda seems to work best these days" and "conda is the recommended approach". The second version connotes that a person or people in authority have deemed this to be the case, and I don't think that has happened. (The only authority would come from a discussion and vote on sage-devel, given our current governance structure.) In contrast, the first version is clearly just an opinion of an individual poster. > > > 3. Kwankyu has brought up some issues about github release creation. > > > We are keeping changelogs here: https://www.sagemath.org/changelogs/index.html. They are automatically generated by a few github workflows. > > To keep the format (contributions divided by releases) of the changelogs, I made the PR https://github.com/sagemath/sage/pull/39194. > > Recently https://github.com/sagemath/sage/pull/40709 removed the code from #39194 without any discussion. > > The author and the reviewer of #40709 made no efforts to fix the regression, as seen in https://github.com/sagemath/sage/pull/40840 and https://github.com/sagemath/sage/pull/40843. #40709 was fixing a much much bigger regression than this. https://github.com/sagemath/sage/pull/39194 solves a very minor (I think non-existent, and I am going to put it to a vote here) issue in a hard to maintain way. CI tasks like one tackled in #40709 like this are hard to fix, as you never know whether it will actually work when deployed in the real, in this case while making a new release. Yes, there was a lot of CI code added, removed or redone lately. I don't know why you have made such a fuss about your 20 lines. > > > 4. Historically (at least in my experience) Sage developers were careful to maintain backwards compatibility, whereas there are at least some now who are willing to break things and then maybe fix them later. Item 3 arose, and some other issues arose, because code was removed without carefully thinking about the consequences. Please stop making unfunded assumptions about someone being careless or not. It's impossible to think about consequences of removing of something that's broken and serves an unclear purpose, while its author is not around. Note that it is about the CI code, which is as a rule hardly ever documented. If someone removes a line saying something like "from X import Y", then the author and reviewer should just do a quick search to see if Y is still used anywhere, rather than just leaving it as an undefined symbol. I think it's fair to call that careless. If someone is working on the build system and neither the author nor the reviewer think to try it out on a fresh git clone, I think it's also fair to call that careless. Others may disagree with me. > > > I agree. Item 3 is one but trivial example. There could be other non-trivial items related with the work of conda/meson-based sage. > > The sage community did not approve transition from the sage-the-distro-based sage to conda/meson-based sage, as we did for python3 transition and github migration. Oh, do you mean to say that abandoning python2 needed some sort of approval? > Hence conda/meson-based sage should be developed while not breaking sage-the-distro-based sage. Nothing got broken in sage the distro in any serious way. Few quirks might be due to the move to the meson-based build of sagemath (the library). We have bid a well-deserved farewell to the old setuptools-and-what-no-based spaghetti, which in part resulted in creation of passagemath. By the way, the latter is being aggressively "marketed" as the "right" Sage, including sticking incompatible with SageMath bits into Sage downstream projects, see e.g. https://github.com/abelfunctions/abelfunctions/commit/0d8a4d5a9a69943ed3ecde4c929e23473cc187d8 Clearly, sage the distro is overdue for a major overhaul, as we are drowning in necessary for going forward updates such as https://github.com/sagemath/sage/pull/40892 which means doing manually what tools like pip and uv do automatically. > > However, the main developer of conda/meson-based sage proposes PRs without testing with sage-the-distro-based sage. An example: https://github.com/sagemath/sage/pull/40841. > > Even though he is not "willing to break things and then maybe fix them later", such a PR, if merged without a careful review, is likely to break sage-the-distro-based sage. Anyone can propose a PR which is not tested in some ways. By the way, I have given that PR a negative review a while ago, what is your problem with it? Dima -- You received this message because you are subscribed to the Google Groups "sage-devel" group. To unsubscribe from this group and stop receiving emails from it, send an email to [email protected]. To view this discussion visit https://groups.google.com/d/msgid/sage-devel/41ed1e62-a07f-45b3-817b-bb72cb75b249n%40googlegroups.com.
