Skip to content

Update 'Get-ChangeLog' to make it more accurate #6764

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.

Sign up for GitHub

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

Merged
TravisEz13 merged 2 commits into from
Apr 30, 2018
Merged

Update 'Get-ChangeLog' to make it more accurate #6764

Changes from all commits
Commits
Show all changes
2 commits
Select commit Hold shift + click to select a range
2c90aa5
Update 'Get-ChangeLog' to make it more accurate
daxian-dbw Apr 27, 2018
1676f60
Add comment about '--first-parent' option
daxian-dbw Apr 30, 2018
File filter

Filter by extension

Filter by extension
Conversations
Failed to load comments.
Loading
Jump to
Jump to file
Failed to load files.
Loading
Diff view
Diff view
60 changes: 42 additions & 18 deletions tools/releaseTools.psm1
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -27,6 +27,22 @@ class CommitNode {
}
}

# These powershell team members don't use 'microsoft.com' for Github email or choose to not show their emails.
# We have their names in this array so that we don't need to query Github to find out if they are powershell team members.
$Script:powershell_team = @(
"Klaudia Algiz"
"Robert Holt"
"Dan Travison"
)

# They are very active contributors, so we keep their email-login mappings here to save a few queries to Github.
$Script:community_login_map = @{
"darpa@yandex.ru" = "iSazonov"
"c.bergmeister@gmail.com" = "bergmeister"
"github@markekraus.com" = "markekraus"
"info@powercode-consulting.se" = "powercode"
}

##############################
#.SYNOPSIS
#In the release workflow, the release branch will be merged back to master after the release is done,
Expand Down Expand Up @@ -55,10 +71,10 @@ function Get-ChildMergeCommit
)

$tag_hash = $CommitHash
if ($PSCmdlet.ParameterSetName -eq "TagName") { git rev-parse "$LastReleaseTag^0" }
if ($PSCmdlet.ParameterSetName -eq "TagName") { $tag_hash = git rev-parse "$LastReleaseTag^0" }

## Get the merge commits that are reachable from 'HEAD' but not from the release tag
$merge_commits_not_in_release_branch = git --no-pager log --merges "$tag_hash..HEAD" --format='%H|%P'
$merge_commits_not_in_release_branch = git --no-pager log --merges "$tag_hash..HEAD" --format='%H||%P'
## Find the child merge commit, whose parent-commit-hashes contains the release tag hash
$child_merge_commit = $merge_commits_not_in_release_branch | Select-String -SimpleMatch $tag_hash
return $child_merge_commit.Line
Expand Down Expand Up @@ -88,7 +104,7 @@ function New-CommitNode
)

Process {
$hash, $parents, $name, $email, $subject = $CommitMetadata.Split("|")
$hash, $parents, $name, $email, $subject = $CommitMetadata.Split("||")
$body = (git --no-pager show $hash -s --format=%b) -join "`n"
return [CommitNode]::new($hash, $parents, $name, $email, $subject, $body)
}
Expand Down Expand Up @@ -124,9 +140,15 @@ function Get-ChangeLog
)

$tag_hash = git rev-parse "$LastReleaseTag^0"
$format = '%H|%P|%aN|%aE|%s'
$format = '%H||%P||%aN||%aE||%s'
$header = @{"Authorization"="token $Token"}

# Find the merge commit that merged the release branch to master.
$child_merge_commit = Get-ChildMergeCommit -CommitHash $tag_hash
$commit_hash, $parent_hashes = $child_merge_commit.Split("||")
# Find the other parent of the merge commit, which represents the original head of master right before merging.
$other_parent_hash = ($parent_hashes -replace $tag_hash).Trim()

if ($HasCherryPick) {
## Sometimes we need to cherry-pick some commits from the master branch to the release branch during the release,
## and eventually merge the release branch back to the master branch. This will result in different commit nodes
Expand All @@ -137,16 +159,10 @@ function Get-ChangeLog
## but not reachable from the last release tag. Instead, we need to exclude the commits that were cherry-picked,
## and only include the commits that are not in the last release into the change log.

# Find the merge commit that merged the release branch to master.
$child_merge_commit = Get-ChildMergeCommit -CommitHash $tag_hash
$commit_hash, $parent_hashes = $child_merge_commit.Split("|")
# Find the other parent of the merge commit, which represents the original head of master right before merging.
$other_parent_hash = ($parent_hashes -replace $tag_hash).Trim()

# Find the commits that were only in the orginal master, excluding those that were cherry-picked to release branch.
$new_commits_from_other_parent = git --no-pager log --no-merges --cherry-pick --right-only "$tag_hash...$other_parent_hash" --format=$format | New-CommitNode
$new_commits_from_other_parent = git --no-pager log --first-parent --cherry-pick --right-only "$tag_hash...$other_parent_hash" --format=$format | New-CommitNode
# Find the commits that were only in the release branch, excluding those that were cherry-picked from master branch.
$new_commits_from_last_release = git --no-pager log --no-merges --cherry-pick --left-only "$tag_hash...$other_parent_hash" --format=$format | New-CommitNode
$new_commits_from_last_release = git --no-pager log --first-parent --cherry-pick --left-only "$tag_hash...$other_parent_hash" --format=$format | New-CommitNode
# Find the commits that are actually duplicate but having different patch-ids due to resolving conflicts during the cherry-pick.
$duplicate_commits = Compare-Object $new_commits_from_last_release $new_commits_from_other_parent -Property PullRequest -ExcludeDifferent -IncludeEqual -PassThru
if ($duplicate_commits) {
Expand All @@ -155,18 +171,26 @@ function Get-ChangeLog
}

# Find the commits that were made after the merge commit.
$new_commits_after_merge_commit = git --no-pager log --no-merges "$commit_hash..HEAD" --format=$format | New-CommitNode
$new_commits_after_merge_commit = @(git --no-pager log --first-parent "$commit_hash..HEAD" --format=$format | New-CommitNode)
Copy link
Copy Markdown
Member

@TravisEz13 TravisEz13 Apr 27, 2018

Choose a reason for hiding this comment

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

--first-parent find only the parent commit in a set of merge commits?

Copy link
Copy Markdown
Member

@TravisEz13 TravisEz13 Apr 27, 2018

Choose a reason for hiding this comment

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

confirmed... this explained it well for me: https://opendatascience.com/git-first-parent-have-your-messy-history-and-eat-it-too/

Copy link
Copy Markdown
Member

@TravisEz13 TravisEz13 Apr 27, 2018

Choose a reason for hiding this comment

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

Maybe the comment should be updated

Copy link
Copy Markdown
Member Author

@daxian-dbw daxian-dbw Apr 28, 2018

Choose a reason for hiding this comment

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

--first-parent basically means for any merge node, only follow the parent node on the master branch side. In case we merge a branch to master for a PR, only the merge node will show up in this way, the individual commits from that branch will be ignored. This is what we want because the merge commit itself already represents the PR.

I will update the comment to explain --first-parent

$new_commits = $new_commits_after_merge_commit + $new_commits_from_other_parent
} else {
## No cherry-pick was involved in the last release branch.
## We can get all new commits using the revision range "$tag_hash..HEAD", meaning the commits that are
## reachable from 'HEAD' but not reachable from the last release tag.
$new_commits = git --no-pager log --no-merges "$tag_hash..HEAD" --format=$format | New-CommitNode
## Using a ref rang like "$tag_hash..HEAD" with 'git log' means getting the commits that are reachable from 'HEAD' but not reachable from the last release tag.

## We use '--first-parent' for 'git log'. It means for any merge node, only follow the parent node on the master branch side.
## In case we merge a branch to master for a PR, only the merge node will show up in this way, the individual commits from that branch will be ignored.
## This is what we want because the merge commit itself already represents the PR.

## First, we want to get all new commits merged during the last release
$new_commits_during_last_release = @(git --no-pager log --first-parent "$tag_hash..$other_parent_hash" --format=$format | New-CommitNode)
## Then, we want to get all new commits merged after the last release
$new_commits_after_last_release = @(git --no-pager log --first-parent "$commit_hash..HEAD" --format=$format | New-CommitNode)
## Last, we get the full list of new commits
$new_commits = $new_commits_during_last_release + $new_commits_after_last_release
}

$community_login_map = @{}
foreach ($commit in $new_commits) {
if ($commit.AuthorEmail.EndsWith("@microsoft.com")) {
if ($commit.AuthorEmail.EndsWith("@microsoft.com") -or $powershell_team -contains $commit.AuthorName) {
$commit.ChangeLogMessage = "- {0}" -f $commit.Subject
} else {
if ($community_login_map.ContainsKey($commit.AuthorEmail)) {
Expand Down