Re: [sage-devel] attempting Sage build on M1

2021-06-04 Thread Dima Pasechnik
On Fri, 4 Jun 2021, 14:46 kcrisman, wrote: > > >>> Perhaps you need to run brew install openjpeg (or brew reinstall openjpeg) >>> Apparently doing this and sage -f pillow (+ its dependencies) fixed this >>> problem. I'm not sure why it was a problem in the first pla

Re: [sage-devel] attempting Sage build on M1

2021-06-04 Thread kcrisman
On what appears to be a completely different note, I cannot use Jupyter. I can do sage -n and it opens the GUI in the browser, but actually attempting to start a worksheet leads to a massive failure. $ ./sage -n ┌┐ │ SageMat

Re: [sage-devel] attempting Sage build on M1

2021-06-04 Thread kcrisman
> >> Perhaps you need to run >>> >>> brew install openjpeg >>> >>> (or brew reinstall openjpeg) >>> >>> >> Apparently doing this and sage -f pillow (+ its dependencies) fixed this >> problem. I'm not sure why it was a problem in the first place, though. >> > > basically, openjpeg install wa

Re: [sage-devel] attempting Sage build on M1

2021-06-04 Thread Dima Pasechnik
On Fri, 4 Jun 2021, 13:12 kcrisman, wrote: > > Perhaps you need to run >> >> brew install openjpeg >> >> (or brew reinstall openjpeg) >> >> > Apparently doing this and sage -f pillow (+ its dependencies) fixed this > problem. I'm not sure why it was a problem in the first place, though. > basic

Re: [sage-devel] attempting Sage build on M1

2021-06-04 Thread kcrisman
> Perhaps you need to run > > brew install openjpeg > > (or brew reinstall openjpeg) > > Apparently doing this and sage -f pillow (+ its dependencies) fixed this problem. I'm not sure why it was a problem in the first place, though. Think we need to add openjpeg to the brew list at https

Re: [sage-devel] attempting Sage build on M1

2021-06-04 Thread kcrisman
On Thursday, June 3, 2021 at 1:05:17 PM UTC-4 dim...@gmail.com wrote: > For cvxopt, you need to patch its setup.py (we already patch it, but > not this bit), > which at the moment says > if sys.platform.startswith("darwin"): > SUITESPARSE_LIB_DIR = '/usr/local/lib' > SUITESPARSE_INC_DIR = '

Re: [sage-devel] attempting Sage build on M1

2021-06-03 Thread Dima Pasechnik
On Wed, Jun 2, 2021 at 2:12 PM kcrisman wrote: > > Some interesting bits of these directly related to the errors in question - > similar final linking error in both cases. > > temp.macosx-11-arm64-3.9/src/map.o > build/temp.macosx-11-arm64-3.9/src/outline.o > build/temp.macosx-11-arm64-3.9/src/

Re: [sage-devel] attempting Sage build on M1

2021-06-03 Thread Dima Pasechnik
For cvxopt, you need to patch its setup.py (we already patch it, but not this bit), which at the moment says if sys.platform.startswith("darwin"): SUITESPARSE_LIB_DIR = '/usr/local/lib' SUITESPARSE_INC_DIR = '/usr/local/include' That is, replace /usr/local/ with the correct prefix for M

Re: [sage-devel] attempting Sage build on M1

2021-06-03 Thread Matthias Koeppe
On Thursday, June 3, 2021 at 9:05:20 AM UTC-7 kcrisman wrote: > Update: It still is looking in /usr/local even though >> /opt/homebrew/Cellar/openjpeg/2.4.0/ exists (and probably already did). It >> also looks for little-cms2 in /usr/local/Cellar even though that also >> exists in /opt/homebre

Re: [sage-devel] attempting Sage build on M1

2021-06-03 Thread kcrisman
> Update: It still is looking in /usr/local even though > /opt/homebrew/Cellar/openjpeg/2.4.0/ exists (and probably already did). It > also looks for little-cms2 in /usr/local/Cellar even though that also > exists in /opt/homebrew. Everything else it is looking for in > /opt/homebrew, it se

Re: [sage-devel] attempting Sage build on M1

2021-06-02 Thread kcrisman
Update: It still is looking in /usr/local even though /opt/homebrew/Cellar/openjpeg/2.4.0/ exists (and probably already did). It also looks for little-cms2 in /usr/local/Cellar even though that also exists in /opt/homebrew. Everything else it is looking for in /opt/homebrew, it seems. -- Yo

Re: [sage-devel] attempting Sage build on M1

2021-06-01 Thread Dima Pasechnik
On Tue, Jun 1, 2021 at 7:15 PM kcrisman wrote: > > >> apparently _imaging.cpython-39-darwin.so does not link to the right >> libraries. >> What's the output of >> >> otool -L >> /Users/karl.crisman/Downloads/BrewSage/sage-9.4.beta0/local/lib/python3.9/site-packages/PIL/_imaging.cpython-39-darwin

Re: [sage-devel] attempting Sage build on M1

2021-06-01 Thread kcrisman
> apparently _imaging.cpython-39-darwin.so does not link to the right > libraries. > What's the output of > > otool -L > /Users/karl.crisman/Downloads/BrewSage/sage-9.4.beta0/local/lib/python3.9/site-packages/PIL/_ > imaging.cpython-39-darwin.so > /Users/karl.crisman/Downloads/BrewSage/sage

Re: [sage-devel] attempting Sage build on M1

2021-06-01 Thread Dima Pasechnik
On Tue, Jun 1, 2021 at 4:23 PM Dima Pasechnik wrote: > > On Mon, May 31, 2021 at 9:24 PM kcrisman wrote: > > > > Not as bad as you might think on the test failures. make ptest gives the > > following - many or even most of which seem to be plotting errors exactly > > like this one (and even "p

Re: [sage-devel] attempting Sage build on M1

2021-06-01 Thread Dima Pasechnik
On Mon, May 31, 2021 at 9:24 PM kcrisman wrote: > > Not as bad as you might think on the test failures. make ptest gives the > following - many or even most of which seem to be plotting errors exactly > like this one (and even "plot(x)" does not work from the CLI, same error): > > ImportErr

Re: [sage-devel] attempting Sage build on M1

2021-05-31 Thread kcrisman
> it is better to install the source using git, as you will need to apply > patches on top of 9.4.beta0 > >> >> Sure, this was just the quickest thing now to make sure I was actually starting with the right thing after having messed up a couple attempts somehow. Currently a whole slew of stuf

Re: [sage-devel] attempting Sage build on M1

2021-05-31 Thread Dima Pasechnik
On Mon, 31 May 2021, 18:57 kcrisman, wrote: > Belay all of that. I am going to start from scratch with source for > 9.4.beta0 from http://mirrors.mit.edu/sage/devel/index.html because I am > mixing up two different sets of installs somehow. > it is better to install the source using git, as you

Re: [sage-devel] attempting Sage build on M1

2021-05-31 Thread kcrisman
Belay all of that. I am going to start from scratch with source for 9.4.beta0 from http://mirrors.mit.edu/sage/devel/index.html because I am mixing up two different sets of installs somehow. -- You received this message because you are subscribed to the Google Groups "sage-devel" group. To

Re: [sage-devel] attempting Sage build on M1

2021-05-31 Thread kcrisman
On Sunday, May 30, 2021 at 5:08:14 AM UTC-4 dim...@gmail.com wrote: > you need Homebrew Python3 installed, I think it comes with ssl. > >> >> Looks like I have that already. Warning: python@3.9 3.9.5 is already installed and up-to-date. To reinstall 3.9.5, run: brew reinstall python@3.9

Re: [sage-devel] attempting Sage build on M1

2021-05-30 Thread Dima Pasechnik
you need Homebrew Python3 installed, I think it comes with ssl. On Sun, 30 May 2021, 02:05 kcrisman, wrote: > > The ARM version of homebrew is installed in /opt/homebrew, not /usr/local. >> Best to first install it and put it in your PATH >> > > I see now that it isn't that already-installed Ho

Re: [sage-devel] attempting Sage build on M1

2021-05-29 Thread Matthias Koeppe
config.log please On Saturday, May 29, 2021 at 6:42:19 PM UTC-7 kcrisman wrote: > source SAGE_ROOT/.homebrew-build-env is currently not enough to get it to > use that version of openblas from the keg. > > I get instead there this below. It wouldn't let me attach full log, see > instead htt

Re: [sage-devel] attempting Sage build on M1

2021-05-29 Thread kcrisman
source SAGE_ROOT/.homebrew-build-env is currently not enough to get it to use that version of openblas from the keg. I get instead there this below. It wouldn't let me attach full log, see instead http://www.math.gordon.edu/~kcrisman/openblas-0.3.13.log - but in principle there shouldn't B

Re: [sage-devel] attempting Sage build on M1

2021-05-29 Thread kcrisman
> The ARM version of homebrew is installed in /opt/homebrew, not /usr/local. > Best to first install it and put it in your PATH > I see now that it isn't that already-installed Homebrew is smart enough to put ARM stuff in that location, but it actually needs two separate installs. This is go

Re: [sage-devel] attempting Sage build on M1

2021-05-29 Thread Matthias Koeppe
>From the top of config.log: PATH: /usr/local/bin PATH: /usr/bin PATH: /bin PATH: /usr/sbin PATH: /sbin PATH: /Library/TeX/texbin PATH: /usr/texbin PATH: /Applications/MathApps PATH: /Users/karl.crisman/Downloads/git-trac-command/bin PATH: /Users/karl.crisman/Library/Haskell/bin The ARM version o

Re: [sage-devel] attempting Sage build on M1

2021-05-28 Thread Dima Pasechnik
On Fri, 28 May 2021 at 20:46, kcrisman wrote: > I am on a MacBook Air with M1 chip, attempting to build Sage 9.3 from > scratch. I just installed as many Homebrew packages as I could from the > install directions, though some optional ones (ffmpeg, for instance) led to > aborts. I have sourced

[sage-devel] attempting Sage build on M1

2021-05-28 Thread kcrisman
I am on a MacBook Air with M1 chip, attempting to build Sage 9.3 from scratch. I just installed as many Homebrew packages as I could from the install directions, though some optional ones (ffmpeg, for instance) led to aborts. I have sourced the homebrew build env just to make sure. Hopefully