-
Notifications
You must be signed in to change notification settings - Fork 4.8k
New issue
Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.
By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.
Already on GitHub? Sign in to your account
(chore) document and refactor release scripts #8078
Conversation
The other scripts use positional arguments, this script should use the same.
There have been some changes in the scripts as well: * Some steps were renamed (e.g `luarocks` became `publish_luarock`) * Steps that are now partially done by the CI only have the human part left (review the machine-generated prs)
@kikito Is there a specific parts needs to be reviewed other than trying this out locally (which I assume you have already did)? |
@fffonion that is in fact a very good point. There is no "dry-run" mode for these scripts so "manually trying" is often not even an option. I have tried some of the steps manually, but I have not tried all the combinations I am afraid T__T |
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
looks good overall, hard to actually test though without a proper dryrun mode.
CONFIRM "Press Enter to push the branch and open the release PR" \ | ||
"or Ctrl-C to cancel." | ||
|
||
set -e |
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
an overall comment on the set -e
and SUCCESS
interaction. A nicer and more userfriendly way would be to use traps. This maybe out of scope for PR but I'll leave it here for inspiration.
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
I admit that I have never used trap
in such a way (I try to avoid bash script when I can ...).
From what I have read, it seems that we would still need set -e
(to produce the signal that then would be captured by trap).
Let's call it out of scope for now, it seems a bit ... philosophical as well (one global rescue function vs lots of small traps?)
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
We had a live conversation about this.
@gszr , I think there must be a particular reason why set -e
is set/unset on every action instead of setting it once globally. Is there? What is your opinion of using trap
to recover? At the very least we should be able to print "the command blablabla failed with blebleble"
Co-authored-by: Joshua Schmid <joshua.schmid@konghq.com> Co-authored-by: Michael Martin <flrgh@protonmail.com>
d381c18
to
87125eb
Compare
Our jenkins server expects and uses some steps in scripts/make-patch-release that were removed in #8078. This commit adds them back as a temporary way of releasing 2.8.0
Our jenkins server expects and uses some steps in scripts/make-patch-release that were removed in #8078. This commit adds them back as a temporary way of releasing 2.8.0
This function was removed accidentally in #8078
usage
function in release scriptsSummary
SUMMARY_GOES_HERE
Full changelog
Issues resolved
Fix #XXX