Tom Lane wrote:
> Not sure what to do here. One idea is to install just the psql-side
> fix, which should break nothing now that version-2 protocol is dead,
> and then wait a few years before introducing the server-side change.
> That seems kind of sad though.
Wouldn't backpatching solve this? Then only the users who don't
apply the minor updates would have non-matching server and psql.
Initially I though that obsoleting the v2 protocol was a recent
move, but reading older messages from the list I've got the
impression that it was more or less in the pipeline since way
before version 10.
Also one of the cases the patch fixes, the one when imported
data are silently truncated at the point of \., is quite nasty
IMO.
I can imagine an app where user-supplied data would be
appended row-by-row into a CSV file, and would be processed
periodically by batch. Under some conditions, in particular
if newlines in the first column are allowed, a malevolent user could
submit a \. sequence to cause the batch to miss the rest of the data
without any error being raised.
[1]
https://www.postgresql.org/message-id/11648.1403147417%40sss.pgh.pa.us
Best regards,
--
Daniel Vérité
https://postgresql.verite.pro/
Twitter: @DanielVerite