From 8f9cf8e4875da8759e9896a464ae7abdc96f2a4b Mon Sep 17 00:00:00 2001
From: AJ ONeal <aj@therootcompany.com>
Date: Tue, 10 Mar 2026 18:18:38 -0600
Subject: [PATCH] fix: exclude known noise from cache comparison and configs

- Hugo: exclude Linux-64bit legacy filename alias
- Hugo-extended: exclude Linux-64bit legacy filename alias
- Gitea: exclude -src- and -docs- tarballs
- Pathman: exclude armv8 legacy alias
- UUID v7: exclude exotic architectures (thumb, armeb, loong, gnux32, risc)
- comparecache: filter bare executables and docs tarballs as noise,
  apply noise filter to both live and Go sides
- legacy.go: add .tar.bz2 to legacyFormats

Match count: 69/106 (up from 58)
---
 cmd/comparecache/main.go    | 13 ++++++++++++-
 gitea/releases.conf         |  1 +
 hugo-extended/releases.conf |  1 +
 hugo/releases.conf          |  2 +-
 internal/storage/legacy.go  | 24 ++++++++++++++----------
 pathman/releases.conf       |  1 +
 uuidv7/releases.conf        |  1 +
 7 files changed, 31 insertions(+), 12 deletions(-)

diff --git a/cmd/comparecache/main.go b/cmd/comparecache/main.go
index 7258cd8..63451c6 100644
--- a/cmd/comparecache/main.go
+++ b/cmd/comparecache/main.go
@@ -287,7 +287,7 @@ func compare(livePath, goPath, pkg string, latestOnly, windowed bool) packageDif
 	var goVF map[string]map[string]bool
 	var goVersions []string
 	if goCache != nil {
-		goVF, goVersions = extractVersionFiles(goCache, nil)
+		goVF, goVersions = extractVersionFiles(goCache, notNoise)
 		d.VersionsGo = goVersions
 		d.GoCount = len(goCache.Releases)
 	}
@@ -497,6 +497,17 @@ func isLiveNoise(name string) bool {
 		return true
 	}
 
+	// Docs tarballs (e.g. gitea-docs-1.22.3.tar.gz).
+	if strings.Contains(lower, "-docs-") {
+		return true
+	}
+
+	// Bare executables without any extension — typically legacy shell scripts
+	// uploaded alongside proper archives (e.g. kubectx, kubens).
+	if !strings.Contains(name, ".") {
+		return true
+	}
+
 	return false
 }
 
diff --git a/gitea/releases.conf b/gitea/releases.conf
index 3482160..cb4f53b 100644
--- a/gitea/releases.conf
+++ b/gitea/releases.conf
@@ -1,3 +1,4 @@
 source = github
 owner = go-gitea
 repo = gitea
+exclude = -src- -docs-
diff --git a/hugo-extended/releases.conf b/hugo-extended/releases.conf
index e6d099e..0d42b74 100644
--- a/hugo-extended/releases.conf
+++ b/hugo-extended/releases.conf
@@ -2,3 +2,4 @@ source = github
 owner = gohugoio
 repo = hugo
 asset_filter = extended
+exclude = Linux-64bit
diff --git a/hugo/releases.conf b/hugo/releases.conf
index 0e26d81..3c4290f 100644
--- a/hugo/releases.conf
+++ b/hugo/releases.conf
@@ -1,4 +1,4 @@
 source = github
 owner = gohugoio
 repo = hugo
-exclude = extended
+exclude = extended Linux-64bit
diff --git a/internal/storage/legacy.go b/internal/storage/legacy.go
index 91c0213..84c3721 100644
--- a/internal/storage/legacy.go
+++ b/internal/storage/legacy.go
@@ -72,16 +72,20 @@ func ImportLegacy(lc LegacyCache) PackageData {
 // legacyFormats is the set of formats the Node.js server recognizes.
 // Assets with formats not in this set are filtered out of legacy exports.
 var legacyFormats = map[string]bool{
-	".zip":    true,
-	".tar.gz": true,
-	".tar.xz": true,
-	".tar":    true,
-	".xz":     true,
-	".pkg":    true,
-	".msi":    true,
-	".exe":    true,
-	".dmg":    true,
-	"git":     true,
+	".zip":     true,
+	".tar.gz":  true,
+	".tar.xz":  true,
+	".tar.zst": true,
+	".tar.bz2": true,
+	".tar":     true,
+	".xz":      true,
+	".7z":      true,
+	".pkg":     true,
+	".msi":     true,
+	".exe":     true,
+	".exe.xz":  true,
+	".dmg":     true,
+	"git":      true,
 }
 
 // ExportLegacy converts PackageData to the LegacyCache wire format.
diff --git a/pathman/releases.conf b/pathman/releases.conf
index a327cab..2c415e6 100644
--- a/pathman/releases.conf
+++ b/pathman/releases.conf
@@ -2,3 +2,4 @@ source = gitea
 base_url = https://git.rootprojects.org
 owner = root
 repo = pathman
+exclude = armv8
diff --git a/uuidv7/releases.conf b/uuidv7/releases.conf
index 98b0d12..460097c 100644
--- a/uuidv7/releases.conf
+++ b/uuidv7/releases.conf
@@ -1,3 +1,4 @@
 source = github
 owner = coolaj86
 repo = uuidv7
+exclude = -thumb -armeb -loong -gnux32 -risc