imworkingon: Fuss with forge precedence

7 files changed, 177 insertions, 124 deletions

diff --git a/cmd/generate/forge_forgejo.go b/cmd/generate/forge_forgejo.go
index 62234b9..84988f2 100644
--- a/cmd/generate/forge_forgejo.go
+++ b/cmd/generate/forge_forgejo.go
@@ -15,10 +15,10 @@ type Forgejo struct {
 var _ Forge = Forgejo{}
 
 func (f Forgejo) FetchStatus(urls []string) (string, error) {
-	for _, u := range urls {
+	return fetchPerURLStatus(urls, func(u string) (string, error) {
 		m := reForgejoPR.FindStringSubmatch(u)
 		if m == nil || m[1] != f.Authority {
-			continue
+			return "", nil
 		}
 		authority := m[1]
 		user := m[2]
@@ -49,15 +49,14 @@ func (f Forgejo) FetchStatus(urls []string) (string, error) {
 		}
 
 		return ret, nil
-	}
-	return "", nil
+	})
 }
 
 func (f Forgejo) FetchSubmittedAt(urls []string) (time.Time, error) {
-	for _, u := range urls {
+	return fetchPerURLSubmittedAt(urls, func(u string) (time.Time, error) {
 		m := reForgejoPR.FindStringSubmatch(u)
 		if m == nil || m[1] != f.Authority {
-			continue
+			return time.Time{}, nil
 		}
 		authority := m[1]
 		user := m[2]
@@ -73,15 +72,14 @@ func (f Forgejo) FetchSubmittedAt(urls []string) (time.Time, error) {
 			return time.Time{}, err
 		}
 		return obj.CreatedAt, nil
-	}
-	return time.Time{}, nil
+	})
 }
 
 func (f Forgejo) FetchLastUpdated(urls []string) (time.Time, User, error) {
-	for _, u := range urls {
+	return fetchPerURLLastUpdated(urls, func(u string) (time.Time, User, error) {
 		m := reForgejoPR.FindStringSubmatch(u)
 		if m == nil || m[1] != f.Authority {
-			continue
+			return time.Time{}, User{}, nil
 		}
 		authority := m[1]
 		user := m[2]
@@ -181,6 +179,5 @@ func (f Forgejo) FetchLastUpdated(urls []string) (time.Time, User, error) {
 		}
 
 		return retUpdatedAt, retUser, nil
-	}
-	return time.Time{}, User{}, nil
+	})
 }
diff --git a/cmd/generate/forge_gerrit.go b/cmd/generate/forge_gerrit.go
index 1e6e073..31f2256 100644
--- a/cmd/generate/forge_gerrit.go
+++ b/cmd/generate/forge_gerrit.go
@@ -62,15 +62,10 @@ var _ Forge = Gerrit{}
 var reGoogleGerritCL = regexp.MustCompile(`https://([a-z]+-review\.googlesource\.com)/c/([^?#]+)/\+/([0-9]+)(?:\?[^#]*)?(?:#.*)?$`)
 
 func (Gerrit) FetchStatus(urls []string) (string, error) {
-	for _, u := range urls {
-		if reGitHubPR.MatchString(u) {
-			return "", nil
-		}
-	}
-	for _, u := range urls {
+	return fetchPerURLStatus(urls, func(u string) (string, error) {
 		m := reGoogleGerritCL.FindStringSubmatch(u)
 		if m == nil {
-			continue
+			return "", nil
 		}
 		authority := m[1]
 		projectID := m[2]
@@ -93,20 +88,15 @@ func (Gerrit) FetchStatus(urls []string) (string, error) {
 		case "ABANDONED":
 			return "closed", nil
 		}
-	}
-	return "", nil
+		return "", nil
+	})
 }
 
 func (Gerrit) FetchSubmittedAt(urls []string) (time.Time, error) {
-	for _, u := range urls {
-		if reGitHubPR.MatchString(u) {
-			return time.Time{}, nil
-		}
-	}
-	for _, u := range urls {
+	return fetchPerURLSubmittedAt(urls, func(u string) (time.Time, error) {
 		m := reGoogleGerritCL.FindStringSubmatch(u)
 		if m == nil {
-			continue
+			return time.Time{}, nil
 		}
 		authority := m[1]
 		projectID := m[2]
@@ -121,15 +111,14 @@ func (Gerrit) FetchSubmittedAt(urls []string) (time.Time, error) {
 			return time.Time{}, err
 		}
 		return obj.Created.Val, nil
-	}
-	return time.Time{}, nil
+	})
 }
 
 func (Gerrit) FetchLastUpdated(urls []string) (time.Time, User, error) {
-	for _, u := range urls {
+	return fetchPerURLLastUpdated(urls, func(u string) (time.Time, User, error) {
 		m := reGoogleGerritCL.FindStringSubmatch(u)
 		if m == nil {
-			continue
+			return time.Time{}, User{}, nil
 		}
 		authority := m[1]
 		projectID := m[2]
@@ -165,6 +154,5 @@ func (Gerrit) FetchLastUpdated(urls []string) (time.Time, User, error) {
 			}
 		}
 		return retUpdatedAt, retUser, nil
-	}
-	return time.Time{}, User{}, nil
+	})
 }
diff --git a/cmd/generate/forge_github.go b/cmd/generate/forge_github.go
index d29e3f7..9f475a3 100644
--- a/cmd/generate/forge_github.go
+++ b/cmd/generate/forge_github.go
@@ -7,10 +7,7 @@ import (
 	"time"
 )
 
-var (
-	reGitHubPR     = regexp.MustCompile(`^https://github\.com/([^/?#]+)/([^/?#]+)/pull/([0-9]+)(?:\?[^#]*)?(?:#.*)?$`)
-	reGitHubCommit = regexp.MustCompile(`^https://github\.com/([^/?#]+)/([^/?#]+)/commit/([0-9a-f]+)(?:\?[^#]*)?(?:#.*)?$`)
-)
+var reGitHubPR = regexp.MustCompile(`^https://github\.com/([^/?#]+)/([^/?#]+)/pull/([0-9]+)(?:\?[^#]*)?(?:#.*)?$`)
 
 func githubPagination(i int) url.Values {
 	params := make(url.Values)
@@ -23,11 +20,15 @@ type GitHub struct{}
 var _ Forge = GitHub{}
 
 func (GitHub) FetchStatus(urls []string) (string, error) {
-	// PR
 	for _, u := range urls {
+		if reGoogleGerritCL.MatchString(u) {
+			return "", nil
+		}
+	}
+	return fetchPerURLStatus(urls, func(u string) (string, error) {
 		m := reGitHubPR.FindStringSubmatch(u)
 		if m == nil {
-			continue
+			return "", nil
 		}
 		user := m[1]
 		repo := m[2]
@@ -57,40 +58,14 @@ func (GitHub) FetchStatus(urls []string) (string, error) {
 		}
 
 		return ret, nil
-	}
-	// Commits from a non-PR
-	var gitURL string
-	var gitCommits []string
-	for _, u := range urls {
-		if m := reGitHubCommit.FindStringSubmatch(u); m != nil {
-			user := m[1]
-			repo := m[2]
-			hash := m[3]
-
-			gitURL = "https://github.com/" + user + "/" + repo
-			gitCommits = append(gitCommits, hash)
-		}
-	}
-	if len(gitCommits) > 0 {
-		ret := statusMerged
-		tag, err := getGitTagThatContainsAll(gitURL, gitCommits...)
-		if err != nil {
-			return "", err
-		}
-		if tag != "" {
-			ret = fmt.Sprintf(statusReleasedFmt, tag)
-		}
-		return ret, nil
-	}
-	// Nope
-	return "", nil
+	})
 }
 
 func (GitHub) FetchSubmittedAt(urls []string) (time.Time, error) {
-	for _, u := range urls {
+	return fetchPerURLSubmittedAt(urls, func(u string) (time.Time, error) {
 		m := reGitHubPR.FindStringSubmatch(u)
 		if m == nil {
-			continue
+			return time.Time{}, nil
 		}
 		user := m[1]
 		repo := m[2]
@@ -105,16 +80,19 @@ func (GitHub) FetchSubmittedAt(urls []string) (time.Time, error) {
 			return time.Time{}, err
 		}
 		return obj.CreatedAt, nil
-	}
-	return time.Time{}, nil
+	})
 }
 
 func (GitHub) FetchLastUpdated(urls []string) (time.Time, User, error) {
-	// PR
 	for _, u := range urls {
+		if reGoogleGerritCL.MatchString(u) {
+			return time.Time{}, User{}, nil
+		}
+	}
+	return fetchPerURLLastUpdated(urls, func(u string) (time.Time, User, error) {
 		m := reGitHubPR.FindStringSubmatch(u)
 		if m == nil {
-			continue
+			return time.Time{}, User{}, nil
 		}
 		user := m[1]
 		repo := m[2]
@@ -213,42 +191,5 @@ func (GitHub) FetchLastUpdated(urls []string) (time.Time, User, error) {
 		}
 
 		return retUpdatedAt, retUser, nil
-	}
-	// Commits from a non-PR
-	{
-		var ret time.Time
-		for _, u := range urls {
-			if m := reGitHubCommit.FindStringSubmatch(u); m != nil {
-				user := m[1]
-				repo := m[2]
-				hash := m[3]
-
-				urlStr := "https://api.github.com/repos/" + user + "/" + repo + "/commits/" + hash
-				var obj struct {
-					Commit struct {
-						Author struct {
-							Date time.Time `json:"date"`
-						} `json:"author"`
-						Committer struct {
-							Date time.Time `json:"date"`
-						} `json:"committer"`
-					} `json:"commit"`
-				}
-				if err := httpGetJSON(urlStr, nil, &obj); err != nil {
-					return time.Time{}, User{}, err
-				}
-				if obj.Commit.Author.Date.After(ret) {
-					ret = obj.Commit.Author.Date
-				}
-				if obj.Commit.Committer.Date.After(ret) {
-					ret = obj.Commit.Committer.Date
-				}
-			}
-		}
-		if ret.IsZero() {
-			return time.Time{}, User{}, nil
-		}
-	}
-	// Nope
-	return time.Time{}, User{}, nil
+	})
 }
diff --git a/cmd/generate/forge_part_git.go b/cmd/generate/forge_part_git.go
new file mode 100644
index 0000000..5288286
--- /dev/null
+++ b/cmd/generate/forge_part_git.go
@@ -0,0 +1,80 @@
+package main
+
+import (
+	"fmt"
+	"regexp"
+	"time"
+)
+
+var reGitHubCommit = regexp.MustCompile(`^https://github\.com/([^/?#]+)/([^/?#]+)/commit/([0-9a-f]+)(?:\?[^#]*)?(?:#.*)?$`)
+
+type PartGit struct{}
+
+var _ Forge = PartGit{}
+
+func (PartGit) FetchStatus(urls []string) (string, error) {
+	var gitURL string
+	var gitCommits []string
+	for _, u := range urls {
+		if m := reGitHubCommit.FindStringSubmatch(u); m != nil {
+			user := m[1]
+			repo := m[2]
+			hash := m[3]
+
+			gitURL = "https://github.com/" + user + "/" + repo
+			gitCommits = append(gitCommits, hash)
+		}
+	}
+	if len(gitCommits) == 0 {
+		return "", nil
+	}
+	ret := statusMerged
+	tag, err := getGitTagThatContainsAll(gitURL, gitCommits...)
+	if err != nil {
+		return "", err
+	}
+	if tag != "" {
+		ret = fmt.Sprintf(statusReleasedFmt, tag)
+	}
+	return ret, nil
+}
+
+func (PartGit) FetchSubmittedAt(urls []string) (time.Time, error) {
+	return time.Time{}, nil
+}
+
+func (PartGit) FetchLastUpdated(urls []string) (time.Time, User, error) {
+	var ret time.Time
+	for _, u := range urls {
+		if m := reGitHubCommit.FindStringSubmatch(u); m != nil {
+			user := m[1]
+			repo := m[2]
+			hash := m[3]
+
+			urlStr := "https://api.github.com/repos/" + user + "/" + repo + "/commits/" + hash
+			var obj struct {
+				Commit struct {
+					Author struct {
+						Date time.Time `json:"date"`
+					} `json:"author"`
+					Committer struct {
+						Date time.Time `json:"date"`
+					} `json:"committer"`
+				} `json:"commit"`
+			}
+			if err := httpGetJSON(urlStr, nil, &obj); err != nil {
+				return time.Time{}, User{}, err
+			}
+			if obj.Commit.Author.Date.After(ret) {
+				ret = obj.Commit.Author.Date
+			}
+			if obj.Commit.Committer.Date.After(ret) {
+				ret = obj.Commit.Committer.Date
+			}
+		}
+	}
+	if ret.IsZero() {
+		return time.Time{}, User{}, nil
+	}
+	return time.Time{}, User{}, nil
+}
diff --git a/cmd/generate/forge_pipermail.go b/cmd/generate/forge_part_pipermail.go
index ccc58f0..af6a009 100644
--- a/cmd/generate/forge_pipermail.go
+++ b/cmd/generate/forge_part_pipermail.go
@@ -21,15 +21,15 @@ var (
 	rePiperMailReply   = regexp.MustCompile(`^\s*<LINK REL="made" HREF="(.*)">\s*$`)
 )
 
-type PiperMail struct{}
+type PartPiperMail struct{}
 
-var _ Forge = PiperMail{}
+var _ Forge = PartPiperMail{}
 
-func (PiperMail) FetchStatus(urls []string) (string, error) {
+func (PartPiperMail) FetchStatus(urls []string) (string, error) {
 	return "", nil
 }
 
-func (PiperMail) FetchSubmittedAt(urls []string) (time.Time, error) {
+func (PartPiperMail) FetchSubmittedAt(urls []string) (time.Time, error) {
 	for _, u := range urls {
 		if !rePiperMailMessage.MatchString(u) {
 			continue
@@ -47,7 +47,7 @@ func (PiperMail) FetchSubmittedAt(urls []string) (time.Time, error) {
 	return time.Time{}, nil
 }
 
-func (PiperMail) nextMonth(ym string) string {
+func (PartPiperMail) nextMonth(ym string) string {
 	yStr, mStr, ok := strings.Cut(ym, "-")
 	if !ok {
 		panic(fmt.Errorf("invalid year-month: %q", ym))
@@ -83,7 +83,7 @@ func (PiperMail) nextMonth(ym string) string {
 	}
 }
 
-func (p PiperMail) threadLen(thread *mailstuff.ThreadedMessage) int {
+func (p PartPiperMail) threadLen(thread *mailstuff.ThreadedMessage) int {
 	if thread == nil {
 		return 0
 	}
@@ -98,7 +98,7 @@ func (p PiperMail) threadLen(thread *mailstuff.ThreadedMessage) int {
 	return ret
 }
 
-func (p PiperMail) FetchLastUpdated(urls []string) (time.Time, User, error) {
+func (p PartPiperMail) FetchLastUpdated(urls []string) (time.Time, User, error) {
 	for _, u := range urls {
 		m := rePiperMailMessage.FindStringSubmatch(u)
 		if m == nil {
@@ -134,7 +134,7 @@ func (p PiperMail) FetchLastUpdated(urls []string) (time.Time, User, error) {
 
 			mboxGzStr, err := httpGet(uBase+ym+".txt.gz", nil)
 			if err != nil {
-				if (ym == uYM || !errors.Is(err, os.ErrNotExist)) {
+				if ym == uYM || !errors.Is(err, os.ErrNotExist) {
 					return time.Time{}, User{}, fmt.Errorf("could not fetch mbox for %s: %w", ym, err)
 				}
 				break
@@ -149,7 +149,7 @@ func (p PiperMail) FetchLastUpdated(urls []string) (time.Time, User, error) {
 				return time.Time{}, User{}, fmt.Errorf("could not parse mbox: %w", err)
 			}
 			if err := gzReader.Close(); err != nil {
-				return time.Time{}, User{},  fmt.Errorf("close gz: %w", err)
+				return time.Time{}, User{}, fmt.Errorf("close gz: %w", err)
 			}
 			mbox = append(mbox, _mbox...)
 			_, messages := mailstuff.ThreadMessages(mbox)
diff --git a/cmd/generate/mailstuff/jwz.md b/cmd/generate/mailstuff/jwz.md
index 54f0a45..91e03f5 100644
--- a/cmd/generate/mailstuff/jwz.md
+++ b/cmd/generate/mailstuff/jwz.md
@@ -11,11 +11,11 @@ feedback valuable.
 
 You write that the algorithm in RFC 5256 is merely a "restating" of
 your algorithm, but I noticed 3 (minor) differences:
-   
+
 1. In your step 1.C, the RFC says to check whether this would create a
    loop, and if it would to skip creating the link; your version only
    says to perform this check in step 1.B.
-   
+
 2. The RFC says to sort the messages by date between your steps 4 and
    5; that is: when grouping by subject, containers in the root set
    should be processed in date-order (you do not specify an order),
diff --git a/cmd/generate/src_contribs.go b/cmd/generate/src_contribs.go
index 39bc04b..5694156 100644
--- a/cmd/generate/src_contribs.go
+++ b/cmd/generate/src_contribs.go
@@ -108,15 +108,31 @@ type Forge interface {
 }
 
 var forges = []Forge{
+	// precedence only matters for .FetchStatus.
+
 	// highest precedence
-	Gerrit{}, // must be higher than GitHub because of golang
+	Gerrit{},
 	GitHub{},
 	GitLab{},
 	Forgejo{"codeberg.org"},
-	PiperMail{},
+	PartPiperMail{},
+	PartGit{},
 	// lowest precedence
 }
 
+func fetchPerURLStatus(urls []string, perURL func(string) (string, error)) (string, error) {
+	for _, u := range urls {
+		status, err := perURL(u)
+		if err != nil {
+			return "", err
+		}
+		if status != "" {
+			return status, nil
+		}
+	}
+	return "", nil
+}
+
 func (c Contribution) fetchStatus() (string, error) {
 	for _, forge := range forges {
 		status, err := forge.FetchStatus(c.URLs)
@@ -130,6 +146,20 @@ func (c Contribution) fetchStatus() (string, error) {
 	return "", fmt.Errorf("idk how to get status for %q", c.URLs[0])
 }
 
+func fetchPerURLSubmittedAt(urls []string, perURL func(string) (time.Time, error)) (time.Time, error) {
+	var ret time.Time
+	for _, u := range urls {
+		submittedAt, err := perURL(u)
+		if err != nil {
+			return time.Time{}, err
+		}
+		if !submittedAt.IsZero() && (ret.IsZero() || submittedAt.Before(ret)) {
+			ret = submittedAt
+		}
+	}
+	return ret, nil
+}
+
 func (c Contribution) fetchSubmittedAt() (time.Time, error) {
 	var ret time.Time
 	for _, forge := range forges {
@@ -155,6 +185,23 @@ func withinOneSecond(a, b time.Time) bool {
 	return d <= time.Second
 }
 
+func fetchPerURLLastUpdated(urls []string, perURL func(string) (time.Time, User, error)) (time.Time, User, error) {
+	var ret struct {
+		time.Time
+		User
+	}
+	for _, u := range urls {
+		updatedAt, updatedBy, err := perURL(u)
+		if err != nil {
+			return time.Time{}, User{}, err
+		}
+		if !updatedAt.IsZero() && (ret.Time.IsZero() || updatedAt.After(ret.Time)) {
+			ret.Time, ret.User = updatedAt, updatedBy
+		}
+	}
+	return ret.Time, ret.User, nil
+}
+
 func (c Contribution) fetchLastUpdated() (time.Time, User, error) {
 	var ret struct {
 		time.Time