Add unstored orphans with rejected parents to recentRejects #9261

1. 
   morcos commented at 3:18 AM on December 2, 2016: member

   Orphan txs that are not stored as orphans because of rejected parents, will not be found by AlreadyHave and so will be requested from each peer that inv's them to us. This will prevent those unnecessary re-requests and when the parent is cleared from recentRejects on a new block, the orphan will have been cleared as well.

   Future improvements may be to intelligently save rejected txs (for compact block reconstruction) or rerequest rejected parents (in case of CPFP) but for now this is an improvement.

2. 
   gmaxwell commented at 3:19 AM on December 2, 2016: contributor

   utACK. Will test.

3. 
   TheBlueMatt commented at 6:48 AM on December 2, 2016: member

   You might want to rebase on #9260.

4. Add unstored orphans with rejected parents to recentRejects dfbe0d530f
5. morcos force-pushed on Dec 3, 2016
6. 
   morcos commented at 1:42 PM on December 3, 2016: member

   rebased on the grave of main.cpp

7. 
   gmaxwell commented at 8:47 AM on December 4, 2016: contributor

   ACK.

8. fanquake added the label P2P on Dec 14, 2016
9. 
   sipa commented at 6:46 AM on December 14, 2016: member

   utACK dfbe0d530f75ca3a9121ed1a391165c90f1068c4

10. 
    morcos commented at 8:51 PM on January 10, 2017: member

    Please tag 0.14

    Not important but simple change which will cut down on some completely unnecessary network requests

11. jonasschnelli added this to the milestone 0.14.0 on Jan 10, 2017
12. 
    sdaftuar commented at 1:19 AM on January 12, 2017: member

    So, post-segwit, a parent tx sent with malleated witness would (after this patch) cause an orphan descendant tx to be added to the reject filter, even though the parent txid is not added to the reject filter. I'm not sure whether the orphan case is common enough for this to be a significant concern, however?

    EDIT: Actually -- I guess the condition here is that the parents are all in the reject filter, so what I wrote is incorrect? Never mind! utACK dfbe0d5

13. 
    laanwj commented at 11:35 AM on January 12, 2017: member

    utACK dfbe0d5

14. laanwj merged this on Jan 12, 2017
15. laanwj closed this on Jan 12, 2017

    ---

16. laanwj referenced this in commit 2742568a00 on Jan 12, 2017
17. codablock referenced this in commit 3cf5d66764 on Jan 19, 2018
18. codablock referenced this in commit 458f684b4e on Jan 20, 2018
19. codablock referenced this in commit 74bfe0b223 on Jan 21, 2018
20. andvgal referenced this in commit 0f45f7487d on Jan 6, 2019
21. CryptoCentric referenced this in commit 07ec015332 on Feb 27, 2019
22. CryptoCentric referenced this in commit e11d1ae195 on Feb 27, 2019
23. MarcoFalke locked this on Sep 8, 2021

Milestone
0.14.0