I was recently trying to debug a cycle reported by wolfictl in https://github.com/wolfi-dev/os/pull/28993. There is an internal slack discussion here. I was using melange at b4ef7f5a92234ad93d87ddcabac45d1b98753aa3.
I made a seemingly unrelated change here that resulted in wolfictl text failing with message like:
$ burl=https://packages.wolfi.dev/bootstrap/stage3
$ wolfictl text --pipeline-dir=./pipelines \
-r$burl -k$burl/wolfi-signing.rsa.pub 2>&1 |
tee out.log
2024/09/20 15:12:20 ERRO
unresolvable cycle:
py3-pygments:2.18.0-r1@local ->
py3-supported-hatchling:1.25.0-r2@local,
caused by:
py3-supported-hatchling:1.25.0-r2@local ->
py3-hatchling:1.25.0-r2@local ->
git:2.46.1-r0@local ->
curl-dev:8.10.1-r0@local ->
libpsl-dev:0.21.5-r3@local ->
libpsl:0.21.5-r3@local ->
gtk-doc:1.34.0-r0@local ->
py3-pygments:2.18.0-r1@local
Error: unable to build graph:
unable to re-add original edge with new dep still causes cycle
gtk-doc:1.34.0-r0@local -> py3-supported-hatchling
unable to build graph:
unable to re-add original edge with new dep still causes cycle
gtk-doc:1.34.0-r0@local -> py3-supported-hatchling
The output made me suspect that somehow git-bootstrap was not being considered. I'm fairly certain that assessment was wrong.
I'm not sure what was the actual problem, but it was fixed with this change (removing uses-strip from py3-pluggy). I got to that change just by dropping dependencies in packages and then re-running and repeating that cycle.
Some better output would probably help debug this. I'm not sure what said output would look like.
I was recently trying to debug a cycle reported by wolfictl in https://github.com/wolfi-dev/os/pull/28993. There is an internal slack discussion here. I was using melange at b4ef7f5a92234ad93d87ddcabac45d1b98753aa3.
I made a seemingly unrelated change here that resulted in
wolfictl text
failing with message like:The output made me suspect that somehow
git-bootstrap
was not being considered. I'm fairly certain that assessment was wrong.I'm not sure what was the actual problem, but it was fixed with this change (removing
uses-strip
from py3-pluggy). I got to that change just by dropping dependencies in packages and then re-running and repeating that cycle.Some better output would probably help debug this. I'm not sure what said output would look like.