I've written to the authors of the mods with which I have this problem, and one suggested it may be an issue with API, as this was the only common change:

Modded blocks don't drop when broken in survival mode. I tested it with .165, .166, .167 on mods: Cotton resources, Epicurean, Extra Pieces and Blockus.

Blocks don't drop also on Trap Expansion

I am not sure but I think this only happens in production because things work for me in development. (For cotton-resources)

I think this could possibly be happening because of the ByteBuf memory leak fix?? It's the only commit in the timeframe that affected data packs, since data packs are 100% synced.

I reverted to Fabric API [1.14.1] Fabric API 0.3.0-pre build 157a and it works again. As my programming skills are more on mysql/php front I don't know what the issue might be.

For me, I have been using the following json file to add back the dropping blocks. Even though I don't know if this is the best / correct way.
In data/modid/loot_tables/blocks/example_block.json:

{
  "type": "minecraft:block",
  "pools": [
    {
      "rolls": 1,
      "entries": [
        {
          "type": "minecraft:item",
          "name": "modid:example_block"
        }
      ],
      "conditions": [
        {
          "condition": "minecraft:survives_explosion"
        }
      ]
    }
  ]
}

Can confirm that going back to fabric-api-0.3.0-pre+build.157a makes modded blocks drop on breaking again.

Went through the list of Fabric releases on Curseforge and confirmed this is broken since 158 but working on 157. Will investigate further later.

The modules that changed between 157 and 158 are as follows:

containers
events-interaction
item-groups
networking
registry-sync
resource-loader

Only registry-sync and resource-loader have changes which matter, here, and they're big ones. Try to triage one of those two.

The problem is registry-sync, somewhere between commits f161891 and 168284d (comparison). I need to sleep now so I can't dive in and root cause it right now, but if this issue is still open tomorrow evening I'll investigate further then.

Looking at the history of registry-sync, there are only two commits in that timespan: 945f265 and 168284d

That makes the culprit likely 168284d. It's a big commit, so it'll take me time to review.

Nah, now that I know the exact issue I can dive in myself.

This seems to be fixed using fabric-api-0.3.0+build.173

Since this is fixed, can it be closed?

Fixed