fix GXS infinite reload of unprocessed items at startup

[jolavillette]

Problem

GXS Channels service was experiencing a massive CPU and memory spike at every startup. Thousands of "Unprocessed" items were being reloaded in an infinite loop because they were never successfully acknowledged and persisted as "Processed" in the database.

Root Causes

1. Partial Processing: [load_unprocessedPosts] was only handling Posts, leaving Comments and Votes in an UNPROCESSED state forever.
2. Missing Persistence: Status updates were not always flushed to disk immediately, causing them to be lost if the service didn't shut down cleanly.
3. Filter Bypass Bug: In RsGxsDataAccess::getMsgIdList, if a filter resulted in zero matches, it returned an empty ID set. The lower-level DB layer interpreted an empty set as "fetch everything," effectively bypassing the filter and reloading already-processed items.

Proposed Changes

• p3GxsChannels:
  • Updated [load_unprocessedPosts] to iterate through Comments and Votes to mark them as processed.
  • Added a forced call to [processMsgMetaChanges()] to ensure immediate DB persistence.
• RsGxsDataAccess:
  • Fixed the filter bypass by skipping groups with no matching message IDs after filtering.
• RsGenExchange:
  • Exposed [processMsgMetaChanges()] and [processGrpMetaChanges()] to protected visibility to allow services to force DB synchronization.

Impact

• Startup is now clean after a one-time migration of pending items.
• Eliminated redundant DB loads and associated memory/CPU spikes.

[csoler]

Why this return? It blocks the checking of other flags below as it did before.

[csoler]

good catch. One issue to answer/fix.