Skip to content

version 4.0.0 #492

New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

Sign up for GitHub

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We鈥檒l occasionally send you account related emails.

Already on GitHub? Sign in to your account

Merged
ofloveandhate merged 1 commit into from
Oct 6, 2026
Merged

version 4.0.0 #492

Show file tree
Hide file tree
Changes from all commits
Commits
Show all changes
1 commit
Select commit
File filter

Filter by extension

Filter by extension
Conversations
Failed to load comments.
Loading
Jump to
Jump to file
Failed to load files.
Loading
Diff view
Diff view
12 changes: 8 additions & 4 deletions CHANGELOG.md
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -67,13 +67,14 @@ _______________________________________________________________________________

_______________________________________________________________________________

## [4.0.0] - unreleased
## [4.0.0] - 2026-10-06

The release where a solve tells you what it is doing. Multiprecision numbers go straight onto a
matplotlib axis, every object describes itself when printed instead of giving you its address, and
a long solve reports how far it has got. Underneath that sits a large correctness pass --
precision, endpoint classification, the classic-format round trip, path crossings, start points --
and Intel macOS wheels are built and tested again.
and Intel macOS wheels are built and tested again. A seeded solve now gives the same bits however
it runs: serially, on any number of threads, or across MPI ranks.

**A major version because it breaks source compatibility.** A `System` no longer carries a
precision, because evaluation happens at the precision of the point it is handed.
Expand All @@ -85,8 +86,11 @@ and the gamma actually used. `Homogenize()` throws where it used to half-succee
**It also breaks quietly, which matters more**: code that still compiles can get different
answers. The default endgame is power series rather than Cauchy, `InEGOperatingZone` changed
meaning, a system can now declare coordinates that the finiteness and realness tests must ignore,
and `function(i)` addresses the patch rows on a patched system. Read Changed and Fixed before
upgrading a script whose answers you rely on.
and `function(i)` addresses the patch rows on a patched system. The thread count is
`BERTINI_NUM_THREADS`, which defaults to every available CPU: `OMP_NUM_THREADS` no longer reaches
b2, and MPI ranks no longer default to one thread. `set_random_seed(0)` is now an ordinary,
reproducible seed rather than a request for entropy. Read Changed and Fixed before upgrading a
script whose answers you rely on.

**Records written by 3.x are not recalled.** Both content encodings moved -- `b2sysenc/2` for
systems, `b2cfgenc/4` for configs -- so a record written by an older version describes a different
Expand Down
2 changes: 1 addition & 1 deletion VERSION
View file
Open in desktop
Original file line number Diff line number Diff line change
@@ -1 +1 @@
4.0.0.dev1
4.0.0