fix(scrape): address residual review findings R1–R3

Use padded sort for last_message_id cursor, surface guild channel
discovery errors, and document DCE_ARCHIVE_ROOT in compose.

docker-compose.yml

@@ -17,5 +17,6 @@ services:
     volumes:
       - ./config:/config:ro,z
       - ./scripts/run-discord-scrape.sh:/opt/dce-scheduler/run-discord-scrape.sh:ro,z
+      # Host path must match archive_root in config/scrape-targets.json (override on other machines).
       - ${DCE_ARCHIVE_ROOT:-/home/brunner56/Documents}:${DCE_ARCHIVE_ROOT:-/home/brunner56/Documents}:z
     command: ["help"]

docs/plans/2026-05-29-012-fix-scrape-residual-review-plan.md

@@ -0,0 +1,47 @@
+---
+title: fix: Close recurring scrape residual review findings
+type: fix
+status: completed
+date: 2026-05-29
+origin: LFG — residual review findings on feat/recurring-cli-scrape (R1–R3)
+---
+
+# fix: Close recurring scrape residual review findings
+
+## Summary
+
+Address manual review residuals from plan 011: correct incremental cursor selection, improve guild channel discovery errors, and document portable archive mount configuration.
+
+## Requirements
+
+| ID | Requirement |
+|----|-------------|
+| R1 | `last_message_id` picks highest snowflake reliably across mixed digit lengths |
+| R2 | `load_guild_channel_cache` surfaces CLI failure output like `load_guild_cache` |
+| R3 | `docker-compose.yml` documents required `DCE_ARCHIVE_ROOT` override |
+| R4 | Existing smoke tests pass after changes |
+
+## Implementation Units
+
+### U1. Fix message cursor (`last_message_id`)
+
+**Files:** `scripts/run-discord-scrape.sh`
+
+**Approach:** Replace `max_by(.id)` with `sort_by(.id) | last | .id` for lexicographic ordering on zero-padded-equal-length snowflakes; Discord IDs in one channel are typically same length — sort_by is safer than max_by for strings.
+
+### U2. Guild channel cache diagnostics
+
+**Files:** `scripts/run-discord-scrape.sh`
+
+**Approach:** Capture `channels` CLI stderr/stdout; `die` with context on failure.
+
+### U3. Compose portability note
+
+**Files:** `docker-compose.yml`
+
+**Approach:** Comment above `DCE_ARCHIVE_ROOT` volume line.
+
+## Verification
+
+- `scripts/tests/run-discord-scrape-smoke.sh`
+- `bash -n scripts/run-discord-scrape.sh`

scripts/run-discord-scrape.sh

@@ -374,7 +374,16 @@ last_message_id() {
   [[ -f "$export_path" ]] || return 0
   jq -r '
     (.messages // [])
-    | if length == 0 then empty else (max_by(.id) | .id) end
+    | if length == 0 then empty else (
+        sort_by(
+          .id as $id
+          | ($id | tostring) as $s
+          | (22 - ($s | length)) as $pad
+          | if $pad > 0 then ("0" * $pad) + $s else $s end
+        )
+        | last
+        | .id
+      ) end
   ' "$export_path"
 }
 
@@ -508,11 +517,15 @@ load_guild_channel_cache() {
   local cache_file="$CACHE_ROOT/channels_${guild_id}_${include_voice}_${include_threads}.txt"
 
   if [[ ! -f "$cache_file" ]]; then
-    "$CLI_BIN" channels \
+    local output
+    if ! output=$("$CLI_BIN" channels \
       --guild "$guild_id" \
       --include-vc "$include_voice" \
-      --include-threads "$include_threads" \
+      --include-threads "$include_threads" 2>&1); then
-      | parse_channel_listing >"$cache_file"
+      die "Channel discovery failed for guild $guild_id. CLI output: $output"
+    fi
+
+    printf '%s\n' "$output" | parse_channel_listing >"$cache_file"
   fi
 
   cat "$cache_file"