On Sun Aug 9, 2026 at 2:28 PM EDT, yaydn wrote:
> På lørdag 8. august 2026 kl. 16:33, skrev Andrew Kloet
> <andrew@kloet.net>:
>
>> On Sat Aug 8, 2026 at 2:00 AM EDT, yaydn wrote:
>> > One of the patches doesn't apply?
>> >
>> > $ cd /usr/ports/net/arti/ && make clean build package test 2>&1 | /usr/ports/infrastructure/bin/portslogger .
>> >
>> > ===>  Patching for arti-2.5.1
>> > ===>   Applying OpenBSD patch patch-crates_arti_src_arti-example-config_toml
>> > Hmm...  Looks like a unified diff to me...
>> > The text leading up to this was:
>> > --------------------------
>> > |Index: crates/arti/src/arti-example-config.toml
>> > |--- crates/arti/src/arti-example-config.toml.orig
>> > |+++ crates/arti/src/arti-example-config.toml
>> > --------------------------
>> > Patching file crates/arti/src/arti-example-config.toml using Plan A...
>> > Hunk #1 succeeded at 94.
>> > Hunk #2 succeeded at 199.
>> > Hunk #3 succeeded at 209.
>> > done
>> > ===>   Applying OpenBSD patch patch-crates_arti_src_cfg_rs
>> > Hmm...  Looks like a unified diff to me...
>> > The text leading up to this was:
>> > --------------------------
>> > |Our config patch breaks this test
>> > |
>> > |Index: crates/arti/src/cfg.rs
>> > |--- crates/arti/src/cfg.rs.orig
>> > |+++ crates/arti/src/cfg.rs
>> > --------------------------
>> > Patching file crates/arti/src/cfg.rs using Plan A...
>> > Hunk #1 succeeded at 763.
>> > done
>> > ===>   Applying OpenBSD patch patch-crates_tor-dirserver_src_http_rs
>> > Hmm...  Looks like a unified diff to me...
>> > The text leading up to this was:
>> > --------------------------
>> > |Backported fix from upstream:
>> > |https://gitlab.torproject.org/tpo/core/arti/-/merge_requests/4286
>> > |
>> > |Index: crates/tor-dirserver/src/http.rs
>> > |--- crates/tor-dirserver/src/http.rs.orig
>> > |+++ crates/tor-dirserver/src/http.rs
>> > --------------------------
>> > Patching file crates/tor-dirserver/src/http.rs using Plan A...
>> > Hunk #1 succeeded at 887.
>> > done
>> > ===>   Applying OpenBSD patch patch-crates_tor-dirserver_src_mirror_operation_rs
>> > Hmm...  Looks like a unified diff to me...
>> > The text leading up to this was:
>> > --------------------------
>> > |Backported fix from upstream:
>> > |https://gitlab.torproject.org/tpo/core/arti/-/merge_requests/4286
>> > |
>> > |Index: crates/tor-dirserver/src/mirror/operation.rs
>> > |--- crates/tor-dirserver/src/mirror/operation.rs.orig
>> > |+++ crates/tor-dirserver/src/mirror/operation.rs
>> > --------------------------
>> > Patching file crates/tor-dirserver/src/mirror/operation.rs using Plan A...
>> > Hunk #1 succeeded at 924.
>> > Hunk #2 succeeded at 984.
>> > done
>> > ===>   Applying OpenBSD patch patch-crates_tor-memquota_Cargo_toml
>> > Hmm...  Looks like a unified diff to me...
>> > The text leading up to this was:
>> > --------------------------
>> > |Needed for patch-crates_tor-memquota_src_config_rs
>> > |
>> > |Index: crates/tor-memquota/Cargo.toml
>> > |--- crates/tor-memquota/Cargo.toml.orig
>> > |+++ crates/tor-memquota/Cargo.toml
>> > --------------------------
>> > Patching file crates/tor-memquota/Cargo.toml using Plan A...
>> > Hunk #1 succeeded at 20.
>> > done
>> > ===>   Applying OpenBSD patch patch-crates_tor-memquota_src_config_rs
>> > Hmm...  Looks like a unified diff to me...
>> > The text leading up to this was:
>> > --------------------------
>> > |The sysinfo crate does not have support for OpenBSD yet.
>> > |https://github.com/GuillaumeGomez/sysinfo/issues/1318
>> > |
>> > |Index: crates/tor-memquota/src/config.rs
>> > |--- crates/tor-memquota/src/config.rs.orig
>> > |+++ crates/tor-memquota/src/config.rs
>> > --------------------------
>> > Patching file crates/tor-memquota/src/config.rs using Plan A...
>> > Ignoring previously applied (or reversed) patch.
>> > Hunk #1 ignored at 1.
>> > Hunk #2 ignored at 292.
>> > 2 out of 2 hunks ignored--saving rejects to crates/tor-memquota/src/config.rs.rej
>> > done
>> > ***>   patch-crates_tor-memquota_src_config_rs did not apply cleanly
>> > ===>   Applying OpenBSD patch patch-modcargo-crates_aws-lc-sys-0_42_0_aws-lc_third_party_s2n-bignum_s2n-bignum-imported_x86_att_curve25519_curve25519_x25519base_alt_S
>> > Hmm...  Looks like a unified diff to me...
>> > The text leading up to this was:
>> > --------------------------
>> > |aws-lc-sys vendors AWS-LC, which vendors s2n-bignum.
>> > |The x86_att curve25519_x25519base_alt implementation placed
>> > |readonly Edwards25519 tables in .text. OpenBSD maps .text
>> > |execute-only, causing SIGSEGV when the assembly loads the tables.
>> > |
>> > |Issue will persist until aws-lc bumps the vendored s2n-bignum to commit:
>> > |4d9b2c1793412692d54b4e2a8cf08123078778c7
>> > |
>> > |Related issue for vendored s2n-bignum:
>> > |https://github.com/aws/aws-lc/issues/3408
>> > |
>> > |Index: modcargo-crates/aws-lc-sys-0.42.0/aws-lc/third_party/s2n-bignum/s2n-bignum-imported/x86_att/curve25519/curve25519_x25519base_alt.S
>> > |--- modcargo-crates/aws-lc-sys-0.42.0/aws-lc/third_party/s2n-bignum/s2n-bignum-imported/x86_att/curve25519/curve25519_x25519base_alt.S.orig
>> > |+++ modcargo-crates/aws-lc-sys-0.42.0/aws-lc/third_party/s2n-bignum/s2n-bignum-imported/x86_att/curve25519/curve25519_x25519base_alt.S
>> > --------------------------
>> > Patching file modcargo-crates/aws-lc-sys-0.42.0/aws-lc/third_party/s2n-bignum/s2n-bignum-imported/x86_att/curve25519/curve25519_x25519base_alt.S using Plan A...
>> > Hunk #1 succeeded at 2360.
>> > Hunk #2 succeeded at 2379.
>> > Hunk #3 succeeded at 2401.
>> > done
>> > ===>   Failed patches:
>> >
>> >     patch-crates_tor-memquota_src_config_rs
>> >
>> > *** Error 1 in . (/usr/ports/infrastructure/mk/bsd.port.mk:2949 '/usr/ports/pobj/arti-2.5.1/.patch_done': @if cd /usr/ports/net/arti/patches...)
>> > *** Error 2 in /usr/ports/net/arti (/usr/ports/infrastructure/mk/bsd.port.mk:2722 'build': @lock=arti-2.5.1;  export _LOCKS_HELD=" arti-2.5....)
>> > --
>> >
>> > Attached the log.
>>
>> Apologies, attatched is an updated tarball.
>> I had forgotten to test after making some final edits
>>
>> Andrew
>>
>
>
> So, I had to modify the patch:
>
> --- /tmp/arti/patches/patch-crates_tor-memquota_src_config_rs	Fri Aug  7 03:54:35 2026
> +++ /usr/ports/net/arti/patches/patch-crates_tor-memquota_src_config_rs	Sat Aug  8 22:59:47 2026
> @@ -4,12 +4,7 @@ https://github.com/GuillaumeGomez/sysinfo/issues/1318
>  Index: crates/tor-memquota/src/config.rs
>  --- crates/tor-memquota/src/config.rs.orig
>  +++ crates/tor-memquota/src/config.rs
> -@@ -1,4 +1,3 @@
> --
> - //! Configuration (private module)
> -
> - use std::sync::LazyLock;
> -@@ -293,66 +292,29 @@ fn compute_max_from_total_system_mem(mem: Result<usize
> +@@ -292,66 +292,29 @@ fn compute_max_from_total_system_mem(mem: Result<usize
>       Qty(mem)
>   }
>
Thanks, don't know how I had that newline in my original file.
>
> Basically, removing the first part to apply. I attached the fixed patch,
> but look over it just in case there is something missing or wrong?
>
> It took a longer than I'd like to figure that out, and the compile took
> a while to build too.
Yeah, clean build test repeat takes a long for Arti. I feel your pain.
>
> $ cd /usr/ports/net/arti/ && make build package test ; make port-lib-depends-check check-shlib-syms
> // Checks had no problems, test did(see below)
>
> $ cd /usr/ports/net/arti/ && make tests 2>&1 | /usr/ports/infrastructure/bin/portslogger .
> // Attached the test log.
Seems the CI doesn't like the filesystem permissions.
I've a bad habit building as root.
I suspect that's why I never caught this.
Will investigate and try to break my bad habits.
>
> With /usr/www/plus.html getting churned out already, this might not get in
> before the freeze, you never know. But at least with more testing this
> can at least get in after the the 8.0-stable release.
>
> Hope this helps. May you have success and happy hacking.
> --
> yaydn
Thank you for your testing, it's been very helpful.
I'll work on the port some more in the coming week.
I'm hopeful upstream imports my fixes so we need fewer patches.

-- 
Andrew