The fix which works for me is here:
https://github.com/sagemath/sage/pull/39850
It's in fact a missing static_cast in Boost interface. As conda is
using clang 18, it's extra sensitive to this sort of
type mis-casts; so one needs
--- a/src/sage/graphs/base/boost_interface.cpp
+++ b/src/sage/graphs
On Wed, Apr 2, 2025 at 3:33 AM John Cremona wrote:
>
> I successfully built 10.6 from scratch on 5 different machines (all
> ubuntu), then (as is my usual practice) did "make gap_packages". This
> worked on all but one machine where I see this, of which the relevant
> line seems to be "UnicodeDec
Hi,
yes, I can reproduce this problem with a fresh Conda. That is, our instructions
https://doc.sagemath.org/html/en/installation/conda.html#using-conda-to-provide-all-dependencies-for-the-sage-library
to build Sage 10.6.on Apple arm64 (M1/2/3/4) using Conda are not working.
That's, well, a bit emb
On Mon, 31 Mar 2025 at 14:36, Dima Pasechnik wrote:
>
> On Mon, Mar 31, 2025 at 4:45 AM John Cremona wrote:
> >
> > I still have the problem. Although what I did in my previous post "fixed"
> > the issue once last week, this week I am back to not being able to use the
> > notebook as it asks f
I looked through past messages and didn't see a similar issue (though I
don't doubt that it did and I just missed it). I updated and upgraded
Homebrew and reinstalled perl but got exactly the same error. Here is the
output again and, below that, is info on my perl version.
make configure
./
Don't try to build Sage's Python (it's broken on macOS - but some devs
are in denial about it, and refuse to remove this usless, even
harmful, package from Sage).
Use a good, already installed, Python 3.11 or 3.12.
E.g. you can install python3.12 with homebrew:
brew install python@3.12
make su
