Hi Isuru, That sounds very promising. I will report back ...
- Marc On Tue, Apr 29, 2025 at 10:59 AM Isuru Fernando <isu...@gmail.com> wrote: > Hi Marc, > > > * Every mach binary should have a minimum deployment target of macOS > 10.13 for Intel and macOS 11 for Arm. > > Yes, we have this condition implicitly, but I made it explicit in the PR > now. > > > * Every load path for every mach binary in the sage tree must either be > a standard system library distributed with the OS or be a file inside of > the Sage tree. > > This is true for all conda packages. If not, it is considered a bug. > > > * Every symlink in the Sage tree must point to a file in the Sage tree. > > This is true for all conda packages. If not, it is considered a bug. > > > I don't want to install a bunch of software on my build system which has > the potential for breaking my current build system for the app by causing > sage to use different packages than what it now uses, without my knowing > which ones those are > > It's fully self-contained. Micromamba is downloaded into > `repo/sage/micromamba` and the conda packages are installed into > `repo/sage/local`. Cleaning up those directories and starting a new shell > should make everything go away. > > Isuru > > > > On Mon, Apr 28, 2025 at 5:00 PM Marc Culler <marc.cul...@gmail.com> wrote: > >> What do you mean by "trying it out"? I am confident that it builds >> sage. But there are other requirements. >> >> I don't have any simple way of running that build right now. I don't >> want to install a bunch of software on my build system which has the >> potential for breaking my current build system for the app by causing sage >> to use different packages than what it now uses, without my knowing which >> ones those are. I can't login to m1.sagemath.org any more, which would >> be the natural way for me to try building with the PR, since my ssh >> credentials seem to no longer work. Even if they did, since I don't do >> multi-arch builds for reasons that I have explained in this thread, that >> would only help with Apple silicon. >> >> Maybe you could help me? Here the major requirements: >> >> * Every mach binary should have a minimum deployment target of macOS >> 10.13 for Intel and macOS 11 for Arm. >> * Every load path for every mach binary in the sage tree must either be a >> standard system library distributed with the OS or be a file inside of the >> Sage tree. >> * Every symlink in the Sage tree must point to a file in the Sage tree. >> >> Since you have already used the PR to build a sage with many spkgs >> installed via conda, perhaps you could look at what you built and see how >> far it is from satisfying those requirements. (I am sure they aren't met, >> but the question is how many hours of work would be involved in changing >> things so that they are met.) >> >> Those requirements are the starting point. From there I have to replace >> every absolute mach load path by a relative path and construct a Sage >> framework and make various adjustments such as adding a tkinter module to >> the python. Then I need to construct an app bundle containing the Sage >> framework and the Tcl and Tk frameworks. That app has to be signed and >> notarized and packaged with the signed notarized pkg installers. >> >> - Marc >> >> >> On Mon, Apr 28, 2025 at 2:30 PM Isuru Fernando <isu...@gmail.com> wrote: >> >>> Marc, do you mind trying out >>> https://github.com/3-manifolds/Sage_macOS/pull/82 ? >>> >>> Isuru >>> >>> On Sat, Apr 26, 2025 at 12:44 PM Dima Pasechnik <dimp...@gmail.com> >>> wrote: >>> >>>> >>>> >>>> On Tuesday, April 22, 2025 at 11:42:21 AM UTC-5 Michael Orlitzky wrote: >>>> >>>> On 2025-04-22 10:19:54, Dima Pasechnik wrote: >>>> > >>>> > That's the first time I see a complaint that the Apple's libraries >>>> used for >>>> > Python or its modules are not accepted. >>>> > This is something we can fix, to an extent. And surely we can adjust >>>> the >>>> > check for xz. >>>> > >>>> > Regarding macOS native bzip2, it just happens to pass our tests, >>>> nothing >>>> > weird about it. >>>> > >>>> > Regarding libz, it, implictly, needs pkg-config. One can hand-craft >>>> > pkg-config .pc file for it, then macOS native libz probably will >>>> work, I >>>> > didn't try. >>>> >>>> We should investigate these independently. Is there any reason for >>>> someone to build zlib, bzip2, libffi, or liblzma in 2025? >>>> >>>> >>>> I don't think there is. I'm starting to prepare PRs to move them to >>>> pre-reqs, >>>> following the patch spkg removal (to appear in the coming beta). >>>> The most obvious one is bzip2, which is available everywhere, no extra >>>> fiddling needed. >>>> Please review (removing bzip2 spkg) >>>> https://github.com/sagemath/sage/pull/40011 >>>> >>>> Dima >>>> >>>> >>>> If there is, >>>> we can't really get rid of python yet. But if not, we should aim to >>>> eliminate them all to avoid conflicts between the system python (that >>>> we want to force) and the SPKG libraries. >>>> >>>> This won't immediately solve Marc's problem, but it would make him the >>>> main consumer of those SPKGs and grant him the freedom to nuke >>>> whatever gets in his way. That should eventually clear up the conflict >>>> between how easy it is to use sage to build these packages, and how >>>> annoying it is to get sage to accept them. And in the meantime, a >>>> ./configure override would let him keep doing what he is doing. >>>> >>>> -- >>>> 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 sage-devel+unsubscr...@googlegroups.com. >>>> To view this discussion visit >>>> https://groups.google.com/d/msgid/sage-devel/6bee67f4-e258-44ef-8674-2f07cc56c6e0n%40googlegroups.com >>>> <https://groups.google.com/d/msgid/sage-devel/6bee67f4-e258-44ef-8674-2f07cc56c6e0n%40googlegroups.com?utm_medium=email&utm_source=footer> >>>> . >>>> >>> -- >>> You received this message because you are subscribed to a topic in the >>> Google Groups "sage-devel" group. >>> To unsubscribe from this topic, visit >>> https://groups.google.com/d/topic/sage-devel/-ASHfAXqVYo/unsubscribe. >>> To unsubscribe from this group and all its topics, send an email to >>> sage-devel+unsubscr...@googlegroups.com. >>> To view this discussion visit >>> https://groups.google.com/d/msgid/sage-devel/CA%2B01voNzkDcxdKCF95MkOMVB7D37GjoR0tTOz21B4B64hn1%2BcA%40mail.gmail.com >>> <https://groups.google.com/d/msgid/sage-devel/CA%2B01voNzkDcxdKCF95MkOMVB7D37GjoR0tTOz21B4B64hn1%2BcA%40mail.gmail.com?utm_medium=email&utm_source=footer> >>> . >>> >> -- >> 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 sage-devel+unsubscr...@googlegroups.com. >> To view this discussion visit >> https://groups.google.com/d/msgid/sage-devel/CALcZXRGCANnyG3pPe2uMMnB_rcZMQ_UkJ7vZPH9QqcVb-Apnsg%40mail.gmail.com >> <https://groups.google.com/d/msgid/sage-devel/CALcZXRGCANnyG3pPe2uMMnB_rcZMQ_UkJ7vZPH9QqcVb-Apnsg%40mail.gmail.com?utm_medium=email&utm_source=footer> >> . >> > -- > You received this message because you are subscribed to a topic in the > Google Groups "sage-devel" group. > To unsubscribe from this topic, visit > https://groups.google.com/d/topic/sage-devel/-ASHfAXqVYo/unsubscribe. > To unsubscribe from this group and all its topics, send an email to > sage-devel+unsubscr...@googlegroups.com. > To view this discussion visit > https://groups.google.com/d/msgid/sage-devel/CA%2B01voN7CVLJjQMVcCT4Mj5Nxa%2BbsVB6XS-rnkJ5o%2BD9s7Ha3A%40mail.gmail.com > <https://groups.google.com/d/msgid/sage-devel/CA%2B01voN7CVLJjQMVcCT4Mj5Nxa%2BbsVB6XS-rnkJ5o%2BD9s7Ha3A%40mail.gmail.com?utm_medium=email&utm_source=footer> > . > -- 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 sage-devel+unsubscr...@googlegroups.com. To view this discussion visit https://groups.google.com/d/msgid/sage-devel/CALcZXRHhCN2XATBN12VyK8hny2g2ZqwO%2B0tFfu7Dxm%2BaaoVeBw%40mail.gmail.com.