Bug 293188 - [NEW PORT] databases/vectorscale: higher performance embedding search and cost-efficient storage for AI applications
Summary: [NEW PORT] databases/vectorscale: higher performance embedding search and cos...
Status: New
Alias: None
Product: Ports & Packages
Classification: Unclassified
Component: Individual Port(s) (show other bugs)
Version: Latest
Hardware: Any Any
: --- Affects Some People
Assignee: freebsd-ports-bugs (Nobody)
URL:
Keywords:
Depends on:
Blocks:
 
Reported: 2026-02-15 11:10 UTC by Volodymyr Kostyrko
Modified: 2026-03-08 06:03 UTC (History)
1 user (show)

See Also:

Attachments
intial version - 0.9.0 (64.80 KB, patch)
2026-02-15 11:10 UTC, Volodymyr Kostyrko 		no flags Details | Diff
View All Add an attachment (proposed patch, testcase, etc.)

Note You need to log in before you can comment on or make changes to this bug.
Description Volodymyr Kostyrko 2026-02-15 11:10:07 UTC 
Created attachment 268070 [details]
intial version - 0.9.0
Comment 1 Kevin Bowling freebsd_committer freebsd_triage 2026-03-02 21:19:56 UTC 
General ack, do you want pgxs added or do you want to localize it like you did in vchord?
Comment 2 Volodymyr Kostyrko 2026-03-03 06:00:48 UTC 
(In reply to Kevin Bowling from comment #1)

I had issues with vchord install step and reverted to documentation and they actually are not directly using pgrx, instead they have their own way to build it. I'll recheck vectorscale docs for pointers today.
Comment 3 Volodymyr Kostyrko 2026-03-03 07:28:15 UTC 
No, vectorscale is directly using cargo-pgrx for building/installing:

https://github.com/timescale/pgvectorscale?tab=readme-ov-file#installing-from-source
Comment 4 Kevin Bowling freebsd_committer freebsd_triage 2026-03-07 06:10:41 UTC 
The build is failing for me, any ideas?

error: failed to run custom build command for `syn v1.0.109`

Caused by:
  process didn't exit successfully: `/wrkdirs/usr/ports/databases/pgvectorscale/work/pgvectorscale-0.9.0/target/release/build/syn-c8befa2064d9afef/build-script-build` (signal: 4, SIGILL: illegal instruction)
warning: build failed, waiting for other jobs to finish...
error: failed to run custom build command for `owo-colors v4.2.3`

Caused by:
  process didn't exit successfully: `/wrkdirs/usr/ports/databases/pgvectorscale/work/pgvectorscale-0.9.0/target/release/build/owo-colors-22a8c4fae6dbc1da/build-script-build` (signal: 4, SIGILL: illegal instruction)
  --- stdout
  cargo:rustc-check-cfg=cfg(doc_cfg)
  cargo:rustc-check-cfg=cfg(const_mut_refs)
*** Error code 1
Comment 5 Volodymyr Kostyrko 2026-03-07 10:36:11 UTC 
(In reply to Kevin Bowling from comment #4)

Quite a weird outcome. SIGILL means illegal instruction is used, but how it ended up there? I tried building on 3 different hosts:

1. AMD Ryzen 7 5825U, RUSTFLAGS=-Ctarget-cpu=native. 0k
2. ARM Neoverse-N1. 0k
3. AMD A8-5500 APU (quite old) - failed with SIGILL.

I need to verify by build env there thoroughly, as I'm using CPUTYPE and -O3 by default.

Host #3 had bdver2 arch if that rings a bell. Other difference is that host #3 is only one using PostgreSQL 18, will check that one too…
Comment 6 Kevin Bowling freebsd_committer freebsd_triage 2026-03-08 06:03:17 UTC 
(In reply to Volodymyr Kostyrko from comment #5)
My failing build is happening on a Intel Ivy Bridge class system without any unique make env flags/settings.  Nothing in the build jumped out right away to me that would indicate we are asking for special instructions.