refactor: move load block thread into ChainstateManager #21030

1. 
   fanquake commented at 7:27 AM on January 29, 2021: member

   This is a diff of Marcos from #19197, which probably should have just been used at the time. After this change, and #21016, we'll have no more global thread(Group)s hanging out in init.

   Co-authored-by: MarcoFalke falke.marco@gmail.com

2. fanquake added the label Refactoring on Jan 29, 2021
3. fanquake added the label Validation on Jan 29, 2021
4. 
   DrahtBot commented at 12:23 PM on January 29, 2021: member

   <!--e57a25ab6845829454e8d69fc972939a-->

   The following sections might be updated with supplementary metadata relevant to reviewers and maintainers.

   <!--174a7506f384e20aa4161008e828411d-->

   Conflicts

   No conflicts as of last run.

5. DrahtBot added the label Needs rebase on Feb 1, 2021
6. fanquake force-pushed on Feb 2, 2021
7. refactor: move load block thread into ChainstateManager

   This is a diff of Marcos from #19197, which probably should have just
   been used at the time. After this change, and #21016, we'll have no more
   global thread(Group)s hanging out in init.
   
   Co-authored-by: MarcoFalke <falke.marco@gmail.com>

   e3341dfaf7
8. fanquake force-pushed on Feb 2, 2021
9. DrahtBot removed the label Needs rebase on Feb 2, 2021
10. 
    laanwj commented at 5:56 AM on February 2, 2021: member

    Concept ACK. @dongcarl Does moving block loading into the chainstate manager align with your goals ?

    Edit: I personally think it should be factored out, but not necessarily to validation.cpp/h, we might want a separate "block storage handling" component where ThreadImport and such can move.

11. 
    MarcoFalke commented at 9:35 AM on February 2, 2021: member

    review ACK e3341dfaf7146f0f234d1a59007f008742360a7b

    This is removing a "validation/blockimport" global from init.cpp. Happy to re-ACK if this is moved to a separate module as suggested by @laanwj

12. fanquake requested review from dongcarl on Feb 10, 2021
13. 
    dongcarl commented at 9:18 PM on February 18, 2021: member

    @dongcarl Does moving block loading into the chainstate manager align with your goals ?

    I believe so. After a cursory look at this PR, I think the std::thread m_load_block would probably eventually be moved into the ChainContext struct I plan to introduce after g_chainman is deglobalized, however, putting it in ChainstateManager is the appropriate move for now!

14. fanquake closed this on Mar 31, 2021

    ---

15. 
    MarcoFalke commented at 6:27 PM on April 2, 2021: member

    Done in #21575

16. fanquake referenced this in commit 1f14130cb0 on Apr 13, 2021
17. DrahtBot locked this on Aug 16, 2022
18. fanquake deleted the branch on Nov 9, 2022

Review Requested
dongcarl