i'm running dovecot 2.3.19.1 + fts_tika for attachment scanning, with
tika-server 2.4.1
submitting a PDF attachment DIRECTLY to tika, via curl, there's no issue --
scan's OK
on mail receipt in dovecot, the attachment is submitted to tika via fts, but
scanning fails with a 'corrupt stream' ERROR.
i've posted the issue, with details here, @ tika ML,
https://lists.apache.org/thread/pztsq8tb8xqz3s4kmjpnt9p3zt07y05k
where it's suggested the message may be getting truncated
since direct-via-curl doesn't truncate, it seems that the submission process --
via fts plugin -- is a possiblity.
can anyone confirm PDF attachment scan in a similar config, with latest
versions?
or suggest what diagnostic to use to dig for further, relevant troubleshooting
info in dovecot+fts?
