[woswasi]

hi,

now I came a bit closer to the mystery of CRC errors in downloaded files.

when downloading with JD2 via zevera from uploaded.to, JD2 always reports "CRC ok", but about 30% of files are definitely broken. the same files downloaded from rapidgater are always OK.

it also looks like if a file is broken on the first try, it is also broken on the 2nd try.

to do: check if uploaded.to files are also broken when downloaded directly (without zevera).

(will add links as soon as avaxhome is back online again)

[pspzockerscene]

Hmm JD should usually check this correctly...

GreeZ pspzockerscene

[woswasi]

new findings:

a big archive downloaded from uploaded.to fails with CRC errors in several .part files. JD2 always reports "CRC ok".

the same big archive downloaded from rapidgator works without problems. I suspect that uploaded.to is the nice guy here.

further investigations needed:

1) does a direct download from uploaded.to (without zevera) produce the same broken part files? (can't test since I have no ul.to premium account)
2) are always the same files broken or are they broken by random?

links to both hosters here:
**External links are only visible to Support Staff****External links are only visible to Support Staff**


UPDATE: JD2 does report "CRC failed", but I couldn't see it initially (column not wide enough) and, misleadingly, the green check mark is displayed. please don't do that in case of a CRC error! better set state to "failed" and leave the broken file as .part file.

[pspzockerscene]

We already have a ticket for the green check mark on errors.

GreeZ psp

[woswasi]

current status: when a broken part is reset then it will be downloaded again without errors on the next try.

my assumption: uploaded.to connections are pretty unstable under high server load. when a download ist aborted and later resumed, JD2 does not continue exactly where it stopped, but some bytes before/after this point, resulting in a wrong CRC.

please check the code which resumes aborted downloads, maybe there's a bug causing an offset of a few bytes.

[pspzockerscene]

There are no known resume bugs and its normal that JD doesn't resume at the exact stoppoint.

GreeZ psp