Ready to publish v0.8 today
7 views
Skip to first unread message
David Moreno Montero
Apr 29, 2016, 4:13:01 AM4/29/16
to onion-dev
Hello everyone,
as planned, I want to publish onion v0.8 today.
I could not manage to complete a lot more the documentation, although I will spend a couple of hours now.
Anyway, I wanted to check if nobody has a final request to get into the release. If there is none, I will publish later today.
Regards,
Zachary Grafton
Apr 29, 2016, 4:47:46 AM4/29/16
to David Moreno Montero, onion-dev
Sounds good to me.
--
You received this message because you are subscribed to the Google Groups "onion-dev" group.
To unsubscribe from this group and stop receiving emails from it, send an email to onion-dev+...@coralbits.com.
To post to this group, send email to onio...@coralbits.com.
Visit this group at https://groups.google.com/a/coralbits.com/group/onion-dev/.
For more options, visit https://groups.google.com/a/coralbits.com/d/optout.
Solar Designer
Apr 29, 2016, 5:32:37 AM4/29/16
to David Moreno Montero, onion-dev
Hi David,
On Fri, Apr 29, 2016 at 10:12:41AM +0200, David Moreno Montero wrote:
> as planned, I want to publish onion v0.8 today.
Great. As it happens, I updated my app to your latest code just
yesterday, and tested it for a few hours (mostly at 10k requests/second
from same LAN, but also a similar test from another continent for a
shorter period to see behavior with higher and less stable latency) - no
unexpected issues.
In my testing, I had excluded the poller.c portion of commit
7fb66d1df141fc36c7694954e4fc8cd78c9f3884. You might want to revert this
portion before the release, as I think it only hides real errors now
(not expected errors). In my testing yesterday, that ONION_ERROR was
never triggered, despite of the removed check for "errno!=ENOENT &&
errno!=EBADF". I think EBADF, if it ever occurs there, definitely
indicates real errors: it means the fd was already closed, and thus
could as well have been reused by another thread (in which case an even
worse error would be hidden from us). I don't know why ENOENT would
possibly occur there. Neither actually occurs for me with current code.
I was building onion (as well as my app) with "gcc version 4.4.7
20120313 (Red Hat 4.4.7-11) (GCC)" on Scientific Linux 6.6. I didn't
try building/running your tests (I was using only my app as the test).
I notice your README.md now lists "GCC 4.7" as a minimum requirement.
Maybe this should be relaxed to 4.4 or 4.4.7? It'd be nice for someone
to test building the whole thing (including tests and including other
build options, not just epoll with threads) on a RHEL6'ish distro like
what I had. Unfortunately, I have no time for this.
As a test, I also test-built my app + onion from yesterday on CentOS 7.1
("gcc version 4.8.5 20150623 (Red Hat 4.8.5-4) (GCC)"). It built fine.
I didn't test the resulting binary there.
> Anyway, I wanted to check if nobody has a final request to get into the
> release. If there is none, I will publish later today.
It's just two minor suggestions above from me: on reverting the poller.c
hunk of 7fb66d1df141fc36c7694954e4fc8cd78c9f3884 and on testing and
listing an older version of GCC as minimum requirement. But these are
optional. No showstoppers.
Thank you!
Alexander
On Fri, Apr 29, 2016 at 10:12:41AM +0200, David Moreno Montero wrote:
> as planned, I want to publish onion v0.8 today.
yesterday, and tested it for a few hours (mostly at 10k requests/second
from same LAN, but also a similar test from another continent for a
shorter period to see behavior with higher and less stable latency) - no
unexpected issues.
In my testing, I had excluded the poller.c portion of commit
7fb66d1df141fc36c7694954e4fc8cd78c9f3884. You might want to revert this
portion before the release, as I think it only hides real errors now
(not expected errors). In my testing yesterday, that ONION_ERROR was
never triggered, despite of the removed check for "errno!=ENOENT &&
errno!=EBADF". I think EBADF, if it ever occurs there, definitely
indicates real errors: it means the fd was already closed, and thus
could as well have been reused by another thread (in which case an even
worse error would be hidden from us). I don't know why ENOENT would
possibly occur there. Neither actually occurs for me with current code.
I was building onion (as well as my app) with "gcc version 4.4.7
20120313 (Red Hat 4.4.7-11) (GCC)" on Scientific Linux 6.6. I didn't
try building/running your tests (I was using only my app as the test).
I notice your README.md now lists "GCC 4.7" as a minimum requirement.
Maybe this should be relaxed to 4.4 or 4.4.7? It'd be nice for someone
to test building the whole thing (including tests and including other
build options, not just epoll with threads) on a RHEL6'ish distro like
what I had. Unfortunately, I have no time for this.
As a test, I also test-built my app + onion from yesterday on CentOS 7.1
("gcc version 4.8.5 20150623 (Red Hat 4.8.5-4) (GCC)"). It built fine.
I didn't test the resulting binary there.
> Anyway, I wanted to check if nobody has a final request to get into the
> release. If there is none, I will publish later today.
hunk of 7fb66d1df141fc36c7694954e4fc8cd78c9f3884 and on testing and
listing an older version of GCC as minimum requirement. But these are
optional. No showstoppers.
Thank you!
Alexander
David Moreno Montero
Apr 29, 2016, 6:05:20 AM4/29/16
to Solar Designer, onion-dev
Thanks. I will lower the GCC version, it was just what I could test.
I will check the patch, not sure if I can revert it.
also I just added a new pull request for documentation. Just if somebody wants to comment there. The resulting documentatoin is at http://coralbits.com/static/onion-docs/ .
Regards,
David.
Zachary Grafton
Apr 29, 2016, 9:13:56 AM4/29/16
to David Moreno Montero, Solar Designer, onion-dev
I think 4.7 was the minimum for the c++ bindings, they definitely don't build on 4.4.
Zack
David Moreno Montero
Apr 29, 2016, 9:28:42 AM4/29/16
to Zachary Grafton, Solar Designer, onion-dev
I will sate that into the docs.
David Moreno Montero
Apr 29, 2016, 9:29:01 AM4/29/16
to Zachary Grafton, Solar Designer, onion-dev
I mean: I will state that into the docs.
David Moreno Montero
Apr 29, 2016, 1:17:20 PM4/29/16
to Zachary Grafton, Solar Designer, onion-dev
Zachary Grafton
Apr 29, 2016, 2:00:42 PM4/29/16
to onion-dev
Awesome. David, I know you mentioned splitting off a develop branch and merging that onto master and tagging when doing a release, but have you created the develop branch and if so, should we issue pull requests against that? Also, if that is the plan, should we add some documentation in the Collaborate! section in the README?
Do we also have a list of issues ready to tackle for the next version?
Thanks,
Zack
Reply all
Reply to author
Forward
0 new messages