Downgrade ERROR to something less less terrifying? #5794

issue brsq openend this issue on February 14, 2015
  1. brsq commented at 11:31 pm on February 14, 2015: none

    A pretty frequent source of confusion is the “errors” that appear in debug.log that really shouldn’t be, is there any reason that these shouldn’t be changed to warnings or just informational prints? “Error” just comes on a bit too strong for my liking and doesn’t really reflect what the log print is talking about (rejected transaction, not someone hitting the self destruct button).

    Propose:

    0ERROR: AcceptToMemoryPool : non-final
    

    Becomes:

    0INFO: AcceptToMemoryPool : non-final
    
  2. paveljanik commented at 8:25 am on February 17, 2015: contributor

    I haven’t heard any discussion about such confusion myself. But I agree that ERROR is too strong here because this is part of AcceptToMemoryPool standard behaviour and doesn’t need user’s attention. One or two days debug.log contains the following here:

    0ERROR: AcceptToMemoryPool: inputs already spent
    1ERROR: AcceptToMemoryPool: nonstandard transaction: dust
    2ERROR: AcceptToMemoryPool: free transaction rejected by rate limiter
    

    None of these messages is an error, but standard information message about tx not being accepted to mempool. Looking at other AcceptToMemoryPool log messages in debug.log, I’d drop “category/level” prefix (ERROR: here) completely and change the message to contain “not accepted”.

  3. laanwj commented at 8:44 am on February 25, 2015: member
    Agreed, would be better if these are not ERROR. I do remember discussions about this before.
  4. laanwj added the label Docs and Output on Feb 25, 2015
  5. sipa commented at 10:16 am on February 25, 2015: member
    I think we should get rid of all validation failure messages, and only report thing when the final result of validation (in accepttomempool or processblock) is available, and report there based on debug settings.
  6. redpola commented at 12:04 pm on March 31, 2015: none
    +1 from me. Just set up my first node (headless, no wallet) to contribute to the bitcoin movement and saw this in my logs. As someone said, this kind of error is terrifying when I’m investing masses of bandwidth downloading the blockchain and I’m too naive/inexperienced to know whether my install is screwing itself and/or the bitcoin network.
  7. laanwj commented at 3:43 pm on June 3, 2015: member
    @sipa Agreed. Going by the number of complaints about ERROR: AcceptToMemoryPool: nonstandard transaction: dust floods, I think we should stop logging messages for incoming transactions by default, whether accepted or rejected. Then add a debug flag for logging these.
  8. laanwj added this to the milestone 0.12.0 on Jun 3, 2015
  9. laanwj referenced this in commit 72a3bb7af1 on Aug 5, 2015
  10. skarra commented at 4:07 am on August 11, 2015: none
    Even for optional logging, wouldn’t it be more useful to also log some additional information - like which tx we are talking about or some such? For a developer trying to get into Bitcoin these messages make no sense (either as an error or as Info). What does having 2000 of these INFO messages tell you veterans?
  11. dcousens commented at 4:19 am on August 11, 2015: contributor

    +1 for replacing with INFO

    Does this need a PR? Removal is here

  12. laanwj closed this on Aug 11, 2015

  13. laanwj referenced this in commit 44757729a4 on Feb 24, 2016
  14. laanwj referenced this in commit 8fc81e0983 on Feb 24, 2016
  15. makevoid referenced this in commit 596a2f5ad3 on Jun 13, 2016
  16. MarcoFalke locked this on Sep 8, 2021

github-metadata-mirror

This is a metadata mirror of the GitHub repository bitcoin/bitcoin. This site is not affiliated with GitHub. Content is generated from a GitHub metadata backup.
generated: 2024-12-18 21:12 UTC

This site is hosted by @0xB10C
More mirrored repositories can be found on mirror.b10c.me