This repository has been archived by the owner on Jul 30, 2021. It is now read-only.
build/git-version: Only call 'git' once #951
Merged
Conversation
This file contains bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
|
Can one of the admins verify this patch? |
3 similar comments
|
Can one of the admins verify this patch? |
|
Can one of the admins verify this patch? |
|
Can one of the admins verify this patch? |
k8s-ci-robot
added
the
cncf-cla: yes
Simplify the logic from 04848d1 (Add release build process, 2016-05-27, kubernetes-retired#47) to: * Move from Bash to a generic POSIX shell. This makes the script more portable (it will work on POSIX systems without Bash). It does mean we need to drop the 'set' (which is not in POSIX [1]), but that's not a major problem because: * We can use &&-chaining instead of 'set -e'. * We do only have the one locally-defined variable, so we don't need 'set -u'. * The simpler logic has no pipes, so 'set -o pipefail' would no longer matter. * Replace the previous multiple Git calls with a single call to 'git describe'. * Use --dirty so we can drop the previous diff call. * Drop --tags, because we don't want to use lightweight (non-annotated) tags. We want to use annotated tags, because we want to use signed release tags. There's no code check for tag signatures here, but RELEASING.md requests 'tag -s vX.Y.Z', which will create signed, annotated tags. * Git's SHA-1 hashes are 40 hex digits, but I've used --abbrev=100 to get unabbreviated hashes even in the face of future hash transitions [2]. This is a bit of a hack compared to the previous approach, because at some point Git may have >100-char hashes. But if that happens, it will be far enough in the future that I think the simplification we get here is worth it (and this constant is easy to bump if we need to). * The new call always returns the most recent tag, and (for commits which are not exact tag matches) also includes the abbreviated commit. To isolate the hash (and possible dirty-mark) for commits that are not exact tags, I use POSIX's ${parameter##[word]} prefix-pattern removal [3]. Exact tags are unlikely to include '-g' (unless it is in the tag itself), so for them the prefix-removal is a no-op. Again, this is a bit of a hack, since a tag might include '-g', but I think that's unlikely enough (as mentioned above, RELEASING.md requests vX.Y.Z for tags) that the simplification we get here is worth it. [1]: http://pubs.opengroup.org/onlinepubs/9699919799/idx/utilities.html [2]: https://github.com/git/git/blob/v2.16.3/Documentation/technical/hash-function-transition.txt [3]: http://pubs.opengroup.org/onlinepubs/9699919799/utilities/V3_chap02.html#tag_18_06_02
k8s-ci-robot
added
the
size/S
wking
force-pushed
the
git-describe-simplification
branch
from
193ce7c
to
3b0431f
Compare
|
ok to test |
rphillips
reviewed
Apr 3, 2018
| fi | ||
|
|
||
| echo $VERSION | ||
| DESCRIPTION=$(git describe --abbrev=100 --dirty) && |
There was a problem hiding this comment.
has an extra && at the end
There was a problem hiding this comment.
has an extra && at the end
Yup, that's the &&-chaining I mention in the initial post. It ensured that the script exits with a non-zero exit code if for some reason the subshell here does. You can test that behavior by calling the script from outside of aa Git repo (e.g. from /) and confirming its non-zero exit code.
rphillips
approved these changes
Apr 3, 2018
Sign up for free
to subscribe to this conversation on GitHub.
Already have an account?
Sign in.
Add this suggestion to a batch that can be applied as a single commit.
This suggestion is invalid because no changes were made to the code.
Suggestions cannot be applied while the pull request is closed.
Suggestions cannot be applied while viewing a subset of changes.
Only one suggestion per line can be applied in a batch.
Add this suggestion to a batch that can be applied as a single commit.
Applying suggestions on deleted lines is not supported.
You must change the existing code in this line in order to create a valid suggestion.
Outdated suggestions cannot be applied.
This suggestion has been applied or marked resolved.
Suggestions cannot be applied from pending reviews.
Suggestions cannot be applied on multi-line comments.
Suggestions cannot be applied while the pull request is queued to merge.
Suggestion cannot be applied right now. Please check back later.
Simplify the logic from #47 to:
Move from Bash to a generic POSIX shell. This makes the script more portable (it will work on POSIX systems without Bash). It does mean we need to drop the
set(which is not in POSIX), but that's not a major problem because:&&-chaining instead ofset -e.set -u.set -o pipefailwould no longer matter.Replace the previous multiple Git calls with a single call to
git describe.Use
--dirtyso we can drop the previousdiffcall.Drop
--tags, because we don't want to use lightweight (non-annotated) tags. We want to use annotated tags, because we want to use signed release tags. There's no code check for tag signatures here, butRELEASING.mdrequeststag -s vX.Y.Z, which will create signed, annotated tags.Git's SHA-1 hashes are 40 hex digits, but I've used
--abbrev=100to get unabbreviated hashes even in the face of future hash transitions. This is a bit of a hack compared to the previous approach, because at some point Git may have >100-char hashes. But if that happens, it will be far enough in the future that I think the simplification we get here is worth it (and this constant is easy to bump if we need to).The new call always returns the most recent tag, and (for commits which are not exact tag matches) also includes the abbreviated commit. To isolate the hash (and possible dirty-mark) for commits that are not exact tags, I use POSIX's
${parameter##[word]}prefix-pattern removal. Exact tags are unlikely to include-g(unless it is in the tag itself), so for them the prefix-removal is a no-op. Again, this is a bit of a hack, since a tag might include-g, but I think that's unlikely enough (as mentioned above,RELEASING.mdrequestsvX.Y.Zfor tags) that the simplification we get here is worth it.Fixes #132.