Skip to content

Commit

Permalink
update(pkgx)
Browse files Browse the repository at this point in the history
  • Loading branch information
@jhheider
jhheider committed Jan 10, 2025
1 parent 179dbab commit 5ef492d
Showing 1 changed file with 7 additions and 6 deletions.
13 changes: 7 additions & 6 deletions projects/pkgx.sh/package.yml
View file
Original file line number Diff line number Diff line change
Expand Up @@ -10,18 +10,19 @@ versions:
provides:
- bin/pkgx

dependencies:
linux/aarch64:
sourceware.org/libffi: 3 # denort

build:
dependencies:
deno.land: ~2 # since 1.3.0
working-directory: ${{prefix}}/bin
rust-lang.org: ^1.56 # since 2.0
# patchelf messes with Deno's ability to find the DATA block, making it
# operate as only a bloated Deno binary
skip: fix-patchelf
script: deno task --config "$SRCROOT"/deno.jsonc compile
script:
- run: deno task --config "$SRCROOT"/deno.jsonc compile
working-directory: ${{prefix}}/bin
if: <2
- run: cargo install --path . --root "{{prefix}}"
if: '>=2'
env:
linux/aarch64:
# deno's pre-built runtime doesn't support glibc old enough
Expand Down

3 comments on commit 5ef492d

@mxcl
Copy link
Member

@mxcl mxcl commented on 5ef492d Jan 10, 2025

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

I am using strip for the binaries that are generated at the main repo, reduces ~1MB. We generally don't as it can break things sometimes but we know it is safe for pkgx.

@jhheider
Copy link
Contributor Author

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

I wonder if we should either:

  • strip by default and add a skip: strip option, or
  • test, strip, test, and decide if strip was ok.

@mxcl
Copy link
Member

@mxcl mxcl commented on 5ef492d Jan 11, 2025

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

I have pushed us more towards “do less because we're the ones maintaining everything” so I think stripping by default will just end up with a less robust ecosystem unless our test: nodes are more thorough.

Could get AI to write better tests. Shame projects don't provide thorough integration tests that people like us can run.

Please sign in to comment.