Closed tiyash-basu-sociomantic closed 5 years ago
Updated to return DRIZZLE_RETURN_COULD_NOT_CONNECT
while reading when addrinfo_next == nullptr
Closed as problem was resolved at the user's side
Closed as problem was resolved at the user's side
This is not resolved with [redacted] yet.
@tiyash-basu-sociomantic, did you get more details about this from your side?
@tiyash-basu-sociomantic, did you get more details about this from your side?
I have been using this fix in my tests and have not encountered any segfault so far.
@tiyash-basu-sociomantic
I have been using this fix in my tests and have not encountered any segfault so far.
Great. I suggest that we do a v6.1.6
patch release as soon as possible. Then subsequently v6.2.0
should soon be ready for release
Great. I suggest that we do a
v6.1.6
patch release as soon as possible. Then subsequentlyv6.2.0
should soon be ready for release
Absolutely! Please let me know if can help with anything.
@tiyash-basu-sociomantic could you rebase on v6.1.x
?
@tiyash-basu-sociomantic could you rebase on
v6.1.x
?
Sure! I'll change the target branch and rebase (the target branch is v6.x.x
atm).
Edit: Done!
@tiyash-basu-sociomantic try to rebase locally on the updated v6.1.x
and force push
@tiyash-basu-sociomantic try to rebase locally on the updated
v6.1.x
andforce push
It's already on v.6.1.x
branch, and the diff confirms that. Not sure why this failure happens.
libdrizzle-redux-container | make: *** No rule to make target 'api-sanity-checker'. Stop.
libdrizzle-redux-container exited with code 2
It's already on
v.6.1.x
branch, and the diff confirms that. Not sure why this failure happens.
No a bit strange. I tried to clear the build cache on travis.org and manually start a new build. Hopefully that solves the issue
Nope. travis is for some reason still using the config file with the api-sanity-checker build. Could you try to do a force-push, just some random whitespace change, to force a new build on travis?
Yep. Now it is working. Just revert if you did any changes and it is good to go :)
Yep. Now it is working. Just revert if you did any changes and it is good to go :)
Nice! :) I just made a harmless updated to a comment:
Try the next address, if available.
=> Try the next address, if it is available.
Thanks!
Previously,
addrinfo_next
was not checked to be null before de-referencing to getai_next
. This would often result in segfaults.In this PR, upon running out of addresses (i.e.,
addrinfo_next == nullptr
),getaddrinfo
is be called again (viadrizzle_state_addrinfo
) to refresh the list of addresses.