On Sat, Aug 1, 2020 at 10:03 AM Samuel Lelievre <[email protected]> wrote: > > 2020-08-01 08:00:32 UTC, Dima Pasechnik: > > > > Sorry for shouting, but "batteries-included" monolithic monsters > > don't work (certainly not long-term), in case of Sage it has become > > increasingly obvious in the past ~10 years. > > An illusion that it is sustainable has been, ironically, kept alive by > > OpenDreamKit, where a number of people were full-time doing the > > "dirty" system programming work. > > > > Every package with spkg-configure.m4 should become part of Sage > > "toolchain" and be removed from Sage proper. It's really become a > > waste of everyone's time, duplication of effort to maintain things > > like curl, libpng, gcc/gfortran, python, maxima, etc etc etc. > > > > If one likes to do system programming, there are much better ways to > > help Sage project along than maintaining the monolithic monster. E.g., > > removing dead bits, such as MPIR, Atlas, enabling system Python > > packages. > > > > Speaking of macOS, on modern macOS properly installing 3rd party > > software into /Applications/ needs various Apple's blessings, unless > > you turn off their security mechanisms. We badly need someone who can > > use Xcode to revive Sage macOS app. > > Besides, how about working with Homebrew formulae to offload more of > > Sage packages there. > > How about helping Conda's Sage distribution. (The latter potentially > > very helpful for Conda on Linux, too). > > Thanks Dima for suggesting better ways to spend my time. > I profoundly agree that I would be more useful if I could > > - help prepare spkg-configure.m4 files for more packages
learn on the job - pick up e.g. https://trac.sagemath.org/ticket/28991 which already has a draft spkg-configure.m4 posted as attachment, create a branch with it (needs a bit of tweaking, see comments there) test it (run ./bootstrap && ./configure, study its output, config.log, configure source (generated - this is really last resort)), bang head on the keyboard, ask questions... After it is done, move on to #29630 (which depends on #28991) > - help enable system Python packages > - help revive the macOS app You'd be on your own here - but surely there is plenty on the net on how to use Xcode, package apps, sign them, etc. See e.g. https://medium.com/swlh/the-easiest-way-to-build-macos-installer-for-your-application-34a11dd08744 > - help with any or all of the Arch, Conda, Cygwin, Debian, > Fedora, Guix, Homebrew, Nix, ZeroInstall distributions I can offer a quick tour of https://github.com/sagemath/homebrew-science with tweaked formulae. I believe it's actually very easy to tweak Homebrew formulae - I did this a bit, the main difficulty is that they are in Ruby, and propose fixes upstream with PRs. > > but sadly I don't know how to do any of those. > > If anybody can give a tutorial on any generally useful > but underpowered task that could take extra help, > such as one of the above, or any other one really, > I can make myself available at any time of day or night, > host the tutorial on my department's BigBlueButton server, > record the session if that would seem useful, > share the recording if the recorded people agree, > and try to start doing useful work. see above. Needless to say, we can also do an informal Sage Day via Google Meet or whatever. > > I want to take this opportunity to thank you for your > relentless work on the spkg-configure.m4 front and on > many other fronts, in Sage and the wider math software > ecosystem, and for your invaluable and patient help with > install problems, crash reports, never tired of going > through build logs and crash logs to find the relevant > piece of information, make the appropriate suggestion, > ask the right follow-up question to figure out the > missing bits of information, and getting things going > when they would otherwise easily stall. Thanks --- don't send me more papers, theses, and books to referee, and joint papers to write, for the time being :-) > > -- > 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 on the web visit > https://groups.google.com/d/msgid/sage-devel/57d0de6e-b549-4835-82b6-774f39ad0f2do%40googlegroups.com. -- 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 on the web visit https://groups.google.com/d/msgid/sage-devel/CAAWYfq0kFtJVxYXkfBJX6Z7DGQk39ghBYuk3879u1zDAopiUwA%40mail.gmail.com.
