Elixir v1.20.0-rc.2 and v1.20.0-rc.3 released

As with previous RCs, give it a try and let us know about compilation times and any false or unclear warnings!

Files: 2994, Lines: 412276, Code: 349918, Comments: 6592
CPU: Apple M1 Max (10-cores), 32GB RAM
MIX_OS_DEPS_COMPILE_PARTITION_COUNT not set

Elixir 1.20.0-rc.2-otp-28

mix deps.compile 312.59s user 60.35s system 226% cpu 2:44.85 total
mix compile --force 120.38s user 22.69s system 429% cpu 33.347 total

Elixir 1.19.5-otp-28

mix deps.compile 286.97s user 57.77s system 241% cpu 2:22.93 total
mix compile --force 114.93s user 21.28s system 414% cpu 32.892 total

All run with Erlang 28.4, using time mix compile --force --profile time

1.19.5-otp-28

[profile] Finished cycle resolution in 0ms
[profile] Finished compilation cycle of 492 modules in 8234ms
[profile] Finished writing modules to disk in 53ms
[profile] Finished after compile callback in 1095ms
[profile] Finished group pass check of 492 modules in 201ms

Executed in   36.69 secs    fish           external
   usr time   41.91 secs    0.42 millis   41.91 secs
   sys time    7.75 secs    4.52 millis    7.74 secs

1.20.0-rc.0

Application does not compile

1.20.0-rc.1-otp-28

[profile] Finished cycle resolution in 0ms
[profile] Finished compilation cycle of 492 modules in 8444ms
[profile] Finished writing modules to disk in 50ms
[profile] Finished after compile callback in 1193ms
[profile] Finished group pass check of 492 modules in 212ms

Executed in   38.67 secs    fish           external
   usr time   42.52 secs    0.44 millis   42.51 secs
   sys time    8.45 secs    4.03 millis    8.45 secs

1.20.0-rc.2-otp-28

[profile] Finished cycle resolution in 0ms
[profile] Finished compilation cycle of 492 modules in 6833ms
[profile] Finished writing modules to disk in 40ms
[profile] Finished after compile callback in 344ms
[profile] Finished group pass check of 492 modules in 322ms

Executed in   34.96 secs    fish           external
   usr time   26.77 secs    0.29 millis   26.77 secs
   sys time    5.98 secs    4.01 millis    5.98 secs

@josevalim Type checker crash: FunctionClauseError in Module.Types.Pattern.badpattern/2 (rc.2) · Issue #15131 · elixir-lang/elixir · GitHub

Already running it on prod, my claude.md file in my monorepo has guardrails for tests and now for compile type errors. Thank you!

Hey! I ran into a compilation issue with open_api_spex on rc.2 that doesn’t happen on rc.1.

On rc.1 the whole thing compiles fine in ~29s. On rc.2 (and current main at 6fd161d) it just hangs forever on two modules: cast.ex and deprecated_cast.ex. I waited 5+ minutes before giving up and killing it. The compiler prints the “it’s taking more than 10s” message for both and never moves past them.

Both modules have a ton of function clauses (~20+) doing pattern matching on nested structs and maps with overlapping shapes, stuff like:

def cast(%__MODULE__{value: nil, schema: %{nullable: true}}), do: ...

def cast(%__MODULE__{value: nil, schema: %{nullable: false}} = ctx), do: ...

def cast(%__MODULE__{value: nil, schema: %{oneOf: list}} = ctx) when is_list(list), do: ...

def cast(%__MODULE__{schema: %{type: :object}} = ctx), do: ...

def cast(%__MODULE__{schema: %{type: :string}} = ctx), do: ...

*# ... many more*

def cast(%{} = ctx), do: cast(struct(\__MODULE_\_, ctx))

My guess is the type checker is getting tripped up by all these overlapping map/struct patterns and blowing up in time.

Phew, there was an extreme regression in compiler performance, for the project I used to evaluate the last RC

timestamp             elixir_version      stage     partition     seconds
--------------------  ------------------  --------  ------------  -------
2026-03-05T13:39:45Z  1.18.4-otp-28       deps      unset         77.32
2026-03-05T13:39:58Z  1.18.4-otp-28       project   default       12.77
2026-03-05T13:41:12Z  1.19.3-otp-28       deps      unset         74.10
2026-03-05T13:41:56Z  1.19.3-otp-28       deps      2             43.29
2026-03-05T13:42:29Z  1.19.3-otp-28       deps      4             33.32
2026-03-05T13:42:43Z  1.19.3-otp-28       project   default       13.82
2026-03-05T13:43:53Z  1.19.5-otp-28       deps      unset         70.23
2026-03-05T13:44:35Z  1.19.5-otp-28       deps      2             42.40
2026-03-05T13:45:08Z  1.19.5-otp-28       deps      4             32.47
2026-03-05T13:45:21Z  1.19.5-otp-28       project   default       13.20
2026-03-05T13:46:33Z  1.20.0-rc.1-otp-28  deps      unset         72.00
2026-03-05T13:47:16Z  1.20.0-rc.1-otp-28  deps      2             42.63
2026-03-05T13:47:49Z  1.20.0-rc.1-otp-28  deps      4             33.29
2026-03-05T13:48:03Z  1.20.0-rc.1-otp-28  project   default       13.74
2026-03-05T13:49:11Z  1.20.0-rc.2-otp-28  deps      unset         68.06
2026-03-05T13:49:53Z  1.20.0-rc.2-otp-28  deps      2             42.51
2026-03-05T13:50:26Z  1.20.0-rc.2-otp-28  deps      4             32.89
2026-03-05T13:53:14Z  1.20.0-rc.2-otp-28  project   default       167.24

I used time mix compile --force --profile time to get an idea for what is the cause. One file containing ~35 objects and input opjects for a graphql api jumped from ~220ms to 74s. The file doesn’t contain any macros that would explain the massive increase in compile time. Is there anything I can do get get a better idea on what is causing this?

I didn’t check the out the warnings yet, because I think this is a more glaring issue

Edit: Sorry, this was not supposed to be a reply

Yes, some of the new checks are expensive and we got bug reports, we are looking into optimized them right now.

We have released v1.20.0-rc.3 with many performance improvements around the type system. See the CHANGELOG below.

For those using Absinthe, I have created this pull request which I recommend using until v1.9.1 is out: perf: Break function dispatch into groups by josevalim · Pull Request #1414 · absinthe-graphql/absinthe · GitHub

1. Enhancements

IEx

• [IEx] Optimize autocompleting modules

2. Bug fixes

Elixir

• [Enum] Fix Enum.slice/2 for ranges with step > 1 sliced by step > 1
• [File] Preserve directory permissions in File.cp_r/3
• [File] Fix File.cp_r/3 infinite loop with symlink cycles
• [File] Fix File.cp_r/3 infinite loop when copying into subdirectory of source
• [File] Warn when defining @type record(), fixes CI on Erlang/OTP 29
• [File] Fix File.Stream Enumerable.count for files without trailing newline
• [Float] Fix Float.parse/1 inconsistent error handling for non-scientific notation overflow
• [Kernel] Process fields even when structs are unknown (regression)
• [Kernel] Improve performance on several corner cases in the type system (regression)
• [Kernel] Fix regression when using Kernel.in/2 in defguard (regression)

Apple M1 Max (10-cores), MIX_OS_DEPS_COMPILE_PARTITION_COUNT not set

Elixir 1.19.5-otp-28 + Erlang 28.4
mix deps.compile --force 284.31s user 54.21s system 231% cpu 2:25.97 total
mix compile --force 114.06s user 20.10s system 416% cpu 32.182 total

Elixir 1.20.0-rc.3-otp-28 + Erlang 28.4
mix deps.compile --force 235.06s user 39.24s system 202% cpu 2:15.61 total
mix compile --force 87.31s user 15.89s system 377% cpu 27.335 total