Review 2026-04-20 - eMuleAI, Mods, Main, and Web-Demand Backlog Pass¶
Scope¶
Focused comparison of:
analysis\emuleai- current
workspaces\v0.72a\app\eMule-main analysis\mods-archivewith focused Xtreme checks on shared-file and known-file behavioranalysis\stale-v0.72a-experimental-cleanas retired reference-only provenance- a small 2026 web-demand scan for current community signals
The pass stayed constrained to the branch goal from workspace policy:
- minimal drift from stock eMule
- prioritize networking, performance, file-handling, security-adjacent hardening, and bug fixes
- ignore theme and translation work
Local Main Catch-Up¶
Two important backlog-state corrections were needed after rechecking the current local
main history:
FEAT-033 is already landed¶
Local commit e15e9f4 lands a worthwhile low-drift storage hardening slice:
- separate config/temp/incoming disk-space floors
- protected-volume aggregation and stop/save behavior in
DownloadQueue - Tweaks exposure for the separate floors
- retirement of the old Import Parts /
PartFileConvertpath
This should be tracked as Done, not as a future candidate.
REF-032 is partial only¶
Local commits b4206dc, e78ab5a, 3f6994b, 092b88e, and 0fcfade clearly moved the
preferences/layout cleanup forward, but they do not complete the item:
CPreferencesDlgstill derives fromCTreePropSheetTreePropSheet.*is still active- many dialogs still depend on
ResizableLib - the tree still has no live
CMFCPropertySheet/CMFCDynamicLayoutadoption
So REF-032 should move from Open to In Progress, but not to Done.
New Backlog Promotions¶
BUG-036 - atomic save for known.met and cancelled.met¶
Current main still saves both files in place from CKnownFileList::Save(). That leaves
the same class of truncation window that already justified earlier atomic-write hardening
for part.met and ipfilter.dat.
analysis\emuleai already shows the low-drift pattern for known.met by writing
known.met.tmp and promoting it with MoveFileEx(..., MOVEFILE_REPLACE_EXISTING). The
retired stale branch and focused Xtreme archive do not carry that hardening, which makes
this a real remaining delta worth promoting.
BUG-037 - destructive same-hash KnownFile replacement¶
Current main still has the historical SafeAddKFile() same-hash replacement path that
the source itself warns can create unshared files or GUI/logical share mismatches.
analysis\emuleai proves the broader duplicate-tracking problem space, while the Xtreme
archive shows this is also long-standing mod-side knowledge. The branch should not import
the full duplicate-history feature set, but it should track the core hash-only
destructive replacement bug.
FEAT-034 - narrow non-blocking shared-files reload¶
Current main CSharedFileList::Reload() still calls FindSharedFiles(false)
synchronously. analysis\emuleai and the focused Xtreme archive both show background
scan/coalescing approaches that avoid UI stalls.
This is promoted only as a narrow performance candidate:
- manual reload responsiveness
- no always-on watcher
- no broader share-policy drift
- low priority behind correctness fixes
Web-Demand Scan¶
The web scan was used only to filter backlog candidates, not to drive broad product drift.
Strong current signals¶
eMule Qtpositions its March 5, 2026 announcement around features "the community has been asking for", specifically daemon/GUI split, REST/web control, IPv6, and NAT traversal: https://emule-qt.org/2026/03/05/hello-emule-2026/- recent
aMule Web Controller/aMuTorrentposts in January-February 2026 show that web control, automation, and modern remote UX are still active demand surfaces: https://www.reddit.com/r/selfhosted/comments/1q2y41z/amule_web_controller_a_modern_replacement_for_the/ https://www.reddit.com/r/emule/comments/1r4geha/amule_web_controller_is_now_amutorrent_a_unified/
This is inference, not a formal survey, but it is a useful current signal.
Recurring operational pain¶
- VPN / Kad / UDP friendliness remains a recurring user problem: https://www.reddit.com/r/Piracy/comments/14zcygi/emule_kad_and_vpn_no_uploads/
- large queue / large tree sluggishness is a long-running complaint: https://www.reddit.com/r/digitalpiracy/comments/uk68me/issues_with_emule/
These are not enough to justify big new subsystem work on a low-drift branch, but they do support continuing the already-selected networking/performance hardening direction.
Backlog Outcome¶
The filtered result of this pass is:
- mark landed
FEAT-033asDone - correct
REF-032toIn Progress - add
BUG-036 - add
BUG-037 - add low-priority
FEAT-034
Explicit Non-Promotions¶
The pass intentionally did not promote these areas:
- IPv6 support
- NAT traversal / buddy extensions
- source saver / source cache product features
- filesystem watcher behavior
- dark mode, themes, and translation changes
Reason: they are either already represented elsewhere, too feature-drifty for the current branch goal, or too broad to justify on the present evidence.