depends: Fallback mechanism may not be working properly #7627

issue laanwj opened this issue on March 1, 2016
  1. laanwj commented at 4:26 PM on March 1, 2016: member

    Today there was an issue with a corrupted boost download:

    <MarcoFalke> of which: Fetching boost...
    <MarcoFalke>   % Total    % Received % Xferd  Average Speed   Time    Time     Time  Current
    <MarcoFalke>                                  Dload  Upload   Total   Spent    Left  Speed
    <MarcoFalke>   0     0    0     0    0     0      0      0 --:--:-- --:--:-- --:--:--     0
    <MarcoFalke> 100   639  100   639    0     0   2017      0 --:--:-- --:--:-- --:--:--  2017
    <MarcoFalke> sha256sum: WARNING: 1 computed checksum did NOT match
    

    No indication of even trying the fallback download.

    I already suspected this before with the fukushi.org certificate issue with libqrcode. The download was failing due to a certificate issue and it wasn't even trying our fallback. At least it didn't break with the wget -> curl change :-)

  2. laanwj added the label Build system on Mar 1, 2016
  3. theuni commented at 10:49 PM on March 1, 2016: member

    The issue here is that the download apparently succeeded. I'm not sure it makes sense to retry in case the hash doesn't match, seems quite unlikely that a subsequent immediate retry would work. Maybe there's an obvious case I'm missing, though?

    Since it's boost, I'd say it's likely that sf.net had one of its "I'm going down but not throwing 404" moments.

  4. laanwj commented at 12:57 PM on March 2, 2016: member

    I'm not sure it makes sense to retry in case the hash doesn't match, seems quite unlikely that a subsequent immediate retry would work

    Right - It shouldn't retry the same location again, but it can try the fallback URL, which should still have the uncorrupted file.

  5. theuni commented at 6:07 AM on March 3, 2016: member

    Heh right. Not sure where my mind went there. https://github.com/theuni/bitcoin/commit/59dacc0e56ae22b2a3b72d58caeecf082e1a59c6 should fix. I'd like to hold off on merging that until I can fix a few other little things while I'm in there though, since each touch of that file should trigger a rebuild of everything.

    Won't be able to get to it until after MIT. No problem if you'd rather not wait. Though I suspect you'll be headed there soon too :)

  6. laanwj commented at 1:29 PM on March 3, 2016: member

    No hurry.

  7. laanwj closed this on Apr 15, 2016

  8. DrahtBot 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: 2026-04-13 15:15 UTC

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