A quoted name that runs on into a title; .xNNN/.zNNN volumes

"lois 1258 Billy Joel - Extremes.part01.rar lois 1258 Billy Joel - Extremes ",
"GrooveCoverage.part26.rarGrooveCoveragebestof", "Goga Sekulic.part134.rar -
Goga Sekulic - …": the stem kept the title, and each part was a release of its
own (lois, CPP, DieViper, Maveri, tigra: ~2,400 files). runOn cuts the name
after its archive extension when a space, a dash or the name's own start
follows. And three-digit .x/.z volumes ("Reply 987986177631730.x004"),
not .x264/.x265.

Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>

edge/src/names.rs

@@ -71,16 +71,24 @@ pub fn file_stem_re(name: &str) -> String {
 
 /// Old-style rar volumes after .r99 and split zip volumes, as the file's own extension, lower case; .xNN but not
 /// .x32/.x64/.x86 (no lookahead here: the builds are put back by old_volume_at).
-static OLD_RAR_VOLUME: LazyLock<Regex> = LazyLock::new(|| re(r"\.(?:[s-vxz])[0-9]{2}$"));
+static OLD_RAR_VOLUME: LazyLock<Regex> = LazyLock::new(|| re(r"\.(?:[s-v][0-9]{2}|[xz][0-9]{2,3})$"));
 
 /// Whether a name ends with an old rar or split zip volume (OLD_RAR_VOLUME without the .x builds).
 fn old_volume_at(b: &[u8]) -> bool {
     let len = b.len();
-    len >= 4
-        && b[len - 4] == b'.'
-        && matches!(b[len - 3], b's'..=b'v' | b'x' | b'z')
-        && b[len - 2..].iter().all(u8::is_ascii_digit)
-        && !(b[len - 3] == b'x' && matches!(&b[len - 2..], b"32" | b"64" | b"86"))
+    let digits = b.iter().rev().take_while(|c| c.is_ascii_digit()).count();
+    if !(2..=3).contains(&digits) || len < digits + 2 || b[len - digits - 2] != b'.' {
+        return false;
+    }
+    let letter = b[len - digits - 1];
+    let number = &b[len - digits..];
+    match letter {
+        b's'..=b'v' => digits == 2,
+        b'z' => true,
+        // Not the builds .x32, .x64, .x86, .x264, .x265.
+        b'x' => !matches!(number, b"32" | b"64" | b"86" | b"264" | b"265"),
+        _ => false,
+    }
 }
 static QUOTED: LazyLock<Regex> = LazyLock::new(|| re(r#""[^"]*""#));
 static COUNTERS: LazyLock<Regex> = LazyLock::new(|| re(r"\[[0-9]+/[0-9]+\]|\([0-9]+/[0-9]+\)"));
@@ -229,7 +237,7 @@ pub fn file_stem(name: &str) -> String {
     let mut len = b.len();
     // Old-style rar volumes after .r99 (".s41"), as the file's own extension, lower case.
     if old_volume_at(b) {
-        len -= 4;
+        len -= b.iter().rev().take_while(|c| c.is_ascii_digit()).count() + 2;
     }
     let mut changed = true;
     while changed {
@@ -679,6 +687,27 @@ pub fn random_poster(from: &str) -> bool {
     looks_random(&local)
 }
 
+/// RUN_ON (lib/release/identity.ts): an archive's or par2's extension.
+static RUN_ON: LazyLock<Regex> = LazyLock::new(|| re(r"(?i)\.(?:part[0-9]+\.rar|vol[0-9]+[+-][0-9]+\.par2|par2|rar|r[0-9]{2,3}|7z(?:\.[0-9]{3})?|zip|nzb|nfo|sfv)"));
+
+/// runOn (lib/release/identity.ts): a quoted name that runs on past its extension into a title, cut there.
+pub fn run_on(name: &str) -> &str {
+    let b = name.as_bytes();
+    let ascii = b.len() >= 4 && b[..4].is_ascii();
+    for m in RUN_ON.find_iter(b) {
+        let rest = &name[m.end()..];
+        if rest.is_empty() {
+            return name;
+        }
+        let spaced = rest.starts_with(js::is_space) || rest.trim_start_matches(js::is_space).starts_with('-');
+        let repeats = ascii && rest.len() >= 4 && rest.as_bytes()[..4].eq_ignore_ascii_case(&b[..4]);
+        if spaced || repeats {
+            return &name[..m.end()];
+        }
+    }
+    name
+}
+
 /// sizeAfterName (lib/release/identity.ts): the subject without a number after the file's unquoted name at its end.
 pub fn size_after_name(subject: &str, name: &str) -> String {
     let end = subject.trim_end_matches(js::is_space);
@@ -703,7 +732,7 @@ pub fn release_identity(poster: &str, subject: &str, name: &str) -> Release {
     let poster = if random_poster(poster) { "" } else { poster };
     // The post rule's key takes the [i/M] total; the release's file total any counter's (counterTotal).
     let post_total = file_counter_total(subject);
-    let stem = file_stem(name);
+    let stem = file_stem(run_on(name));
     // A name that isn't quoted (unquoted_name) is taken out as the quotes would have been.
     // The subject here has its spaces collapsed; the name may not.
     // Other quoted text stays: it names the release (">>> CREATIVE COMMONS NZB <<< "0190 alarm 5.0 " - File 1 of 9: …").
@@ -1445,6 +1474,8 @@ mod image_set_tests {
             ("App.x86", "App.x86"),
             ("Licence 700790900887990.z18", "Licence 700790900887990"),
             ("Licence 077887994400668.x12", "Licence 077887994400668"),
+            ("Reply 987986177631730.x004", "Reply 987986177631730"),
+            ("Some.Movie.2010.x264", "Some.Movie.2010.x264"),
             ("chappelle1.part02.rev", "chappelle1"),
             ("chappelle1.rev", "chappelle1"),
             ("VwE3i90MVX9dqMFuDfzdS7Q6.part1.par2", "VwE3i90MVX9dqMFuDfzdS7Q6"),
@@ -1459,6 +1490,21 @@ mod image_set_tests {
         assert_eq!(super::size_after_name("(Hot MILF getting banged hard)[27/70] - 2407.jpg - 28", "2407.jpg"), "(Hot MILF getting banged hard)[27/70] - 2407.jpg");
         assert_eq!(super::size_after_name(r#"Show [01/10] - "a.jpg" - 28"#, "a.jpg"), r#"Show [01/10] - "a.jpg" - 28"#);
         assert_eq!(super::size_after_name("Show - Episode 05", "x.mkv"), "Show - Episode 05");
+        for (n, cut) in [
+            ("lois 1258 Billy Joel - Extremes.part01.rar lois 1258 Billy Joel - Extremes ", "lois 1258 Billy Joel - Extremes.part01.rar"),
+            ("GrooveCoverage.part26.rarGrooveCoveragebestof", "GrooveCoverage.part26.rar"),
+            ("Goga Sekulic.part134.rar - Goga Sekulic - Diskografija (2000-2014).", "Goga Sekulic.part134.rar"),
+            ("HASf293jLAIWUELvhaG3.part091.rar HASf293jLAIWUELvhaG3", "HASf293jLAIWUELvhaG3.part091.rar"),
+            ("Movie.rar.Collection.mkv", "Movie.rar.Collection.mkv"),
+            ("x.part01.rar", "x.part01.rar"),
+        ] {
+            assert_eq!(super::run_on(n), cut, "{n}");
+        }
+        let lois = |i: &str, n: &str| super::release_identity("p", &format!(r#"(thnx to original poster) [{i}/92] - "{n}" yEnc"#), n);
+        assert_eq!(
+            lois("02", "lois 1258 Billy Joel - Extremes.part01.rar lois 1258 Billy Joel - Extremes ").key,
+            lois("01", "lois 1258 Billy Joel - Extremes.par2 lois 1258 Billy Joel - Extremes ").key
+        );
         let milf = |i: &str, n: &str, kb: &str| super::release_identity("p", &format!("(Hot MILF getting banged hard)[{i}/70] - {n} - {kb}"), n);
         assert_eq!(milf("27", "2407.jpg", "28").key, milf("28", "2408.jpg", "31").key);
         assert_eq!(super::file_stem_re("rld-gtre.r13.002"), "rld-gtre");

lib/release/identity.ts

@@ -51,7 +51,8 @@ export function fileStem(name: string): string {
   // Old-style rar volumes after .r99 ("BD_NoTHing018.s41", "….s00", ".t91", ".v81"): only as the file's own extension, in
   // lower case, so a season ("show.s01.rar", "Show.S01") stays. Split zip volumes (".z18"), and .xNN but not the builds
   // ".x32", ".x64", ".x86" ("Licence 700790900887990.z18", "….x31"; "….v1.0.142.x64" stays).
-  let stem = name.trim().replace(/\.(?:[s-vz]\d{2}|x(?!32$|64$|86$)\d{2})$/, '')
+  // Also three digits ("Reply 987986177631730.x004"), but not ".x264", ".x265".
+  let stem = name.trim().replace(/\.(?:[s-v]\d{2}|z\d{2,3}|x(?!(?:32|64|86|264|265)$)\d{2,3})$/, '')
   let changed = true
   while (changed) {
     changed = false
@@ -291,6 +292,26 @@ export function randomPoster(from: string): boolean {
   return looksRandom((at === -1 ? address : address.slice(0, at)).replace(/[-_.]/g, ''))
 }
 
+/** An archive's or par2's extension, where a name that runs on into a title is cut (runOn). */
+const RUN_ON = /\.(?:part\d+\.rar|vol\d+[+-]\d+\.par2|par2|rar|r\d{2,3}|7z(?:\.\d{3})?|zip|nzb|nfo|sfv)/gi
+
+/**
+ * A quoted name that runs on past its extension into a title ("lois 1258 Billy Joel - Extremes.part01.rar lois 1258 Billy
+ * Joel - Extremes ", "GrooveCoverage.part26.rarGrooveCoveragebestof", "Goga Sekulic.part134.rar - Goga Sekulic - …": each
+ * part a release of its own): the name up to that extension, when a space or a dash follows it, or the name's own start.
+ */
+export function runOn(name: string): string {
+  const head = name.slice(0, 4).toLowerCase()
+  const ascii = /^[\x00-\x7f]{4}$/.test(head)
+  for (const m of name.matchAll(RUN_ON)) {
+    const end = m.index + m[0].length
+    const rest = name.slice(end)
+    if (rest.length === 0) return name
+    if (/^\s|^\s*-/.test(rest) || (ascii && rest.slice(0, 4).toLowerCase() === head)) return name.slice(0, end)
+  }
+  return name
+}
+
 /** The subject without a number after the file's unquoted name at its end (" - 28": a size in kB). */
 export function sizeAfterName(subject: string, name: string): string {
   const end = subject.trimEnd()
@@ -305,7 +326,7 @@ export function releaseIdentity(input: ReleaseInput): ReleaseIdentity | null {
   const counter = /\[(\d+)\/(\d+)\]/.exec(input.subject)
   // The post rule's key takes the [i/M] total (as when the catalog started); the release's file total any counter's.
   const postTotal = counter && !brokenTotal(counter[1], counter[2]) ? Number.parseInt(counter[2], 10) : undefined
-  const stem = fileStem(input.name)
+  const stem = fileStem(runOn(input.name))
 
   if (input.weak) {
     // Obfuscating posters randomise the poster per article; a real filename

lib/release/index.ts

@@ -1,5 +1,5 @@
 /**
  * Grouping of files into releases (downloads), independent of AdonisJS.
  */
-export { releaseIdentity, par2ReleaseIdentity, par2ReleaseName, rarReleaseIdentity, isGenericName, fileStem, sizeAfterName, subjectRemainder, namesRelease, isPostToken, type ReleaseIdentity, type ReleaseInput } from './identity.js'
+export { releaseIdentity, par2ReleaseIdentity, par2ReleaseName, rarReleaseIdentity, isGenericName, fileStem, sizeAfterName, runOn, subjectRemainder, namesRelease, isPostToken, type ReleaseIdentity, type ReleaseInput } from './identity.js'
 export { looksRandom } from './random.js'

tests/unit/release/identity.spec.ts

@@ -1,5 +1,5 @@
 import { test } from '@japa/runner'
-import { fileStem, releaseIdentity, sizeAfterName, subjectRemainder } from '#lib/release/index'
+import { fileStem, releaseIdentity, runOn, sizeAfterName, subjectRemainder } from '#lib/release/index'
 import { looksLikeFilename, weakFileKey } from '#lib/headers/index'
 
 test.group('release / fileStem', () => {
@@ -308,6 +308,16 @@ test.group('release / par2 names', () => {
     assert.equal(fileStem('App.x86'), 'App.x86')
     assert.equal(fileStem('Licence 700790900887990.z18'), 'Licence 700790900887990')
     assert.equal(fileStem('Licence 077887994400668.x12'), 'Licence 077887994400668')
+    assert.equal(fileStem('Reply 987986177631730.x004'), 'Reply 987986177631730')
+    assert.equal(fileStem('Some.Movie.2010.x264'), 'Some.Movie.2010.x264')
+    assert.equal(runOn('lois 1258 Billy Joel - Extremes.part01.rar lois 1258 Billy Joel - Extremes '), 'lois 1258 Billy Joel - Extremes.part01.rar')
+    assert.equal(runOn('GrooveCoverage.part26.rarGrooveCoveragebestof'), 'GrooveCoverage.part26.rar')
+    assert.equal(runOn('Goga Sekulic.part134.rar - Goga Sekulic - Diskografija (2000-2014).'), 'Goga Sekulic.part134.rar')
+    assert.equal(runOn('HASf293jLAIWUELvhaG3.part091.rar HASf293jLAIWUELvhaG3'), 'HASf293jLAIWUELvhaG3.part091.rar')
+    assert.equal(runOn('Movie.rar.Collection.mkv'), 'Movie.rar.Collection.mkv')
+    assert.equal(runOn('x.part01.rar'), 'x.part01.rar')
+    const lois = (i: string, n: string) => releaseIdentity({ poster: 'p', subject: `(thnx to original poster) [${i}/92] - "${n}" yEnc`, name: n, weak: false })!
+    assert.equal(lois('02', 'lois 1258 Billy Joel - Extremes.part01.rar lois 1258 Billy Joel - Extremes ').key, lois('01', 'lois 1258 Billy Joel - Extremes.par2 lois 1258 Billy Joel - Extremes ').key)
     assert.equal(fileStem('chappelle1.part02.rev'), 'chappelle1')
     assert.equal(fileStem('chappelle1.rev'), 'chappelle1')
     assert.equal(fileStem('VwE3i90MVX9dqMFuDfzdS7Q6.part1.par2'), 'VwE3i90MVX9dqMFuDfzdS7Q6')