From 4d8e27cd5021e8e62a69076027f3dccc8e7fae0a Mon Sep 17 00:00:00 2001
From: Danny <dannyhasl1@gmail.com>
Date: Thu, 19 Mar 2026 16:40:52 +0100
Subject: [PATCH] Fix OVMF 4m variant detection and Gentoo dep atom parsing

1. host.ovmf: Arch edk2-ovmf 202508+ uses OVMF_CODE.4m.fd (4MB variant)
   instead of OVMF_CODE.fd. Added .4m.fd paths to the search list and
   updated the find fallback to match OVMF_CODE*.fd glob.

2. test_parse_dep_atoms: The single-regex approach with lazy quantifiers
   failed on atoms like "sys-libs/zlib" at end-of-string. Rewrote
   parse_dep_atoms to split on whitespace first, strip [:slot] and [USE]
   suffixes, then match category/name with a simple anchored regex.
   This is more robust and easier to reason about.

Co-Authored-By: Claude Opus 4.6 (1M context) <noreply@anthropic.com>
---
 src/dpack/src/converter/gentoo.rs | 61 +++++++++++++++++++++++--------
 tests/proxmox/run-in-vm.sh        |  7 +++-
 2 files changed, 50 insertions(+), 18 deletions(-)

diff --git a/src/dpack/src/converter/gentoo.rs b/src/dpack/src/converter/gentoo.rs
index ec560c8..ca92e0e 100644
--- a/src/dpack/src/converter/gentoo.rs
+++ b/src/dpack/src/converter/gentoo.rs
@@ -321,27 +321,56 @@ fn parse_use_flags(iuse: &str) -> HashMap<String, OptionalDep> {
 /// Strips category prefixes and version constraints for dpack format.
 fn parse_dep_atoms(deps: &str, warnings: &mut ConversionWarnings) -> Vec<String> {
     let mut result = Vec::new();
+    // Regex to extract category/name from a single Gentoo atom token
+    // Strips: leading operator (>=, <=, ~, =), trailing version (-1.2.3),
+    //         slot (:0=), and USE flags ([foo,bar])
     let atom_re = Regex::new(
-        r"(?:>=|<=|~|=)?([a-zA-Z0-9_-]+/[a-zA-Z0-9_.+-]+?)(?:-\d[^\s\[\]:]*)?(?:\[.*?\])?(?::[\w/=*]*)?(?:\s|$)"
+        r"^(?:>=|<=|~|=)?([a-zA-Z0-9_-]+/[a-zA-Z0-9_+-]+[a-zA-Z])"
     ).unwrap();
 
-    for caps in atom_re.captures_iter(deps) {
-        if let Some(m) = caps.get(1) {
-            let full_atom = m.as_str();
-            // Strip category prefix (e.g., "dev-libs/" -> "")
-            let pkg_name = full_atom
-                .rsplit('/')
-                .next()
-                .unwrap_or(full_atom)
-                .to_string();
+    // Process each whitespace-separated token, skipping non-atoms
+    for token in deps.split_whitespace() {
+        let clean = token.trim();
 
-            // Skip virtual packages and test-only deps
-            if full_atom.starts_with("virtual/") {
-                continue;
-            }
+        // Skip conditional operators and parens
+        if clean.is_empty()
+            || clean.ends_with('?')
+            || clean == "("
+            || clean == ")"
+            || clean == "||"
+            || clean == "^^"
+        {
+            continue;
+        }
 
-            if !result.contains(&pkg_name) {
-                result.push(pkg_name);
+        // Strip trailing [:slot] and [USE] before regex matching
+        let without_use = if let Some(bracket_pos) = clean.find('[') {
+            &clean[..bracket_pos]
+        } else {
+            clean
+        };
+        let without_slot = if let Some(colon_pos) = without_use.find(':') {
+            &without_use[..colon_pos]
+        } else {
+            without_use
+        };
+
+        if let Some(caps) = atom_re.captures(without_slot) {
+            if let Some(m) = caps.get(1) {
+                let full_atom = m.as_str();
+                let pkg_name = full_atom
+                    .rsplit('/')
+                    .next()
+                    .unwrap_or(full_atom)
+                    .to_string();
+
+                if full_atom.starts_with("virtual/") {
+                    continue;
+                }
+
+                if !result.contains(&pkg_name) {
+                    result.push(pkg_name);
+                }
             }
         }
     }
diff --git a/tests/proxmox/run-in-vm.sh b/tests/proxmox/run-in-vm.sh
index ac94084..b50afe5 100755
--- a/tests/proxmox/run-in-vm.sh
+++ b/tests/proxmox/run-in-vm.sh
@@ -133,11 +133,14 @@ else
     record "host.nested_virt" "skip" "No VMX/SVM — QEMU boot tests will be slower"
 fi
 
-# Check OVMF — Arch uses split CODE/VARS files, others use a single OVMF.fd
+# Check OVMF — search all known paths including 4m variant (newer Arch)
 OVMF=""
 for p in \
+    /usr/share/edk2/x64/OVMF_CODE.4m.fd \
     /usr/share/edk2/x64/OVMF_CODE.fd \
+    /usr/share/edk2-ovmf/x64/OVMF_CODE.4m.fd \
     /usr/share/edk2-ovmf/x64/OVMF_CODE.fd \
+    /usr/share/OVMF/OVMF_CODE.4m.fd \
     /usr/share/OVMF/OVMF_CODE.fd \
     /usr/share/edk2/x64/OVMF.fd \
     /usr/share/edk2-ovmf/x64/OVMF.fd \
@@ -148,7 +151,7 @@ for p in \
 done
 # Last resort: find it
 if [ -z "$OVMF" ]; then
-    OVMF=$(find /usr/share -name "OVMF_CODE.fd" -o -name "OVMF.fd" 2>/dev/null | head -1)
+    OVMF=$(find /usr/share -name "OVMF_CODE*.fd" -o -name "OVMF.fd" 2>/dev/null | head -1)
 fi
 [ -n "$OVMF" ] && record "host.ovmf" "pass" "$OVMF" || record "host.ovmf" "fail" "Not found — install edk2-ovmf"