Stream: implementers
Topic: fhir-svn
Github Notifications (Nov 20 2015 at 10:53):
jmandel pushed to branch master
- 9363f16: Updates for overview, added conformance resources
Github Notifications (Nov 20 2015 at 15:13):
jmandel pushed to branch master
- 1fb174a: Test zulip notifications on build
Github Notifications (Nov 21 2015 at 15:40):
jmandel pushed to branch master
- 4632213: * Fixed 4 local paths like C:/work/org.hl7.fhir/build/tools/java/imports/ to be relative paths like the other 3.
Josh Mandel (Nov 21 2015 at 19:23):
Hrmn, these commits are all pushed by jmandel (because it's my github account running the sync process), but we're not seeing who authors them. Would be good to update zulip's github receiver plugin to display authors too (or instead), but for now at least the messages are appearing.
nicola (RIO/SS) (Nov 21 2015 at 19:25):
One more reason to move main repository to github :)
nicola (RIO/SS) (Nov 21 2015 at 19:26):
Working standard should live in a working environment
Josh Mandel (Nov 21 2015 at 19:26):
Amen to that :-) But this point always appears to meet a brick wall.
Travis CI (Nov 21 2015 at 21:43):
Author: fake-author@fake.com
Build status: Passed 
Travis CI (Nov 21 2015 at 21:44):
Author: fake-author@fake.com
Build status: Passed
Travis CI (Nov 21 2015 at 21:45):
Author: fake-author@fake.com
Build status: Passed
Travis CI (Nov 21 2015 at 22:20):
Author: fake-author@fake.com
Build status: Passed
Github Notifications (Nov 21 2015 at 22:27):
jmandel pushed to branch master
- 6dccb98: Use webhook for travis
Travis CI (Nov 21 2015 at 22:50):
Author: fake-author@fake.com
Build status: Passed
Travis CI (Nov 21 2015 at 22:56):
Author: fake-author@fake.com
Build status: Passed
Travis CI (Nov 21 2015 at 22:58):
Author: fake-author@fake.com
Build status: Passed
Travis CI (Nov 21 2015 at 22:59):
Author: fake-author@fake.com
Build status: Passed
Travis CI (Nov 21 2015 at 23:01):
Author: fake-author@fake.com
Build status: Passed
Grahame Grieve (Nov 21 2015 at 23:08):
moving to github will remain a brick wall as long as it's me supporting it, since I have zero confidence in my ability to support it, and it's always me who needs to support it
Josh Mandel (Nov 21 2015 at 23:09):
I don't know why so little confidence -- the rest of the world manages. It's a learning curve, but less steep than plenty of crap you're willing to put up with :-)
Grahame Grieve (Nov 21 2015 at 23:10):
maybe so. What I know is that with just 2 people, I get in a mess, I have no idea why it thinks things are a mess, and the only way to fix it is to clear it out and start again.
Josh Mandel (Nov 21 2015 at 23:10):
A huge part of the sanity comes from not having 30 people committing directly to your repo (or at least, not to the main branch of your code).
Grahame Grieve (Nov 21 2015 at 23:11):
so it's not good for actual use, that's what I hear you saying
Josh Mandel (Nov 21 2015 at 23:11):
Not saying that at all. I'm saying it gives you options by which you can prevent people from randomly committing stuff that breaks.
Josh Mandel (Nov 21 2015 at 23:12):
A common practice is: someone submits a pull request, tests run automatically indicating if the PR builds OK, and only then (after it's determined to be safe) someone decides to merge it into the main code branch.
Grahame Grieve (Nov 21 2015 at 23:13):
it's the last bit that's terrifying. because maybe someone doesn't. So you get this long list of possible things that might work. Or might not. And who the hell knows?
Grahame Grieve (Nov 21 2015 at 23:13):
and it'll be more sorting it out always
Josh Mandel (Nov 21 2015 at 23:13):
This "safety" notion isn't about the version control breaking things. It's about the new code/changes breaking things.
Grahame Grieve (Nov 21 2015 at 23:13):
it'll be me sorting it out, I meant
Josh Mandel (Nov 21 2015 at 23:13):
Wait, deciding to merge only code that *works* is terrifying?
Grahame Grieve (Nov 21 2015 at 23:14):
that's the one thing that annoys me about Zulip - it's too easy to write something in the wrong stream
Josh Mandel (Nov 21 2015 at 23:14):
Right stream, wrong topic in this case.
Josh Mandel (Nov 21 2015 at 23:14):
But easily fixed if you care to.
Grahame Grieve (Nov 21 2015 at 23:14):
well, ok, wrong topic
Josh Mandel (Nov 21 2015 at 23:14):
The workflow really is "hit r"
Josh Mandel (Nov 21 2015 at 23:14):
To reply to a message.
Travis CI (Nov 21 2015 at 23:25):
Author: fake-author@fake.com
Build status: Passed
Github Notifications (Nov 21 2015 at 23:28):
jmandel pushed to branch master
- 6c457d3: Debug build hooks
Grahame Grieve (Nov 21 2015 at 23:30):
what's' debug build hooks'?
Josh Mandel (Nov 21 2015 at 23:31):
Oh, I'm trying to get the Travis CI notifications to feed into Zulip. Right now they're... not. They're generating a 404 even though they appear (in the logs) to be hitting the same URL I'm hitting when I test with curl. So, I added a debugging destination to see what happens.
I should probably debug this somewhere other than the FHIR build, come to think of it ;-)
Josh Mandel (Nov 21 2015 at 23:32):
(That would certainly eliminate the 20-min lag time.)
Travis CI (Nov 21 2015 at 23:45):
Author: j
Build status: Passed
Josh Mandel (Nov 21 2015 at 23:45):
Ah well, apparently adding a debugging check fixed it.
Github Notifications (Nov 22 2015 at 00:00):
jmandel pushed to branch master
- 03f7987: Fixed build hooks; removed debugging line
Travis CI (Nov 22 2015 at 00:04):
Author: fake-author
Commit @1234
Build status: Passed
Grahame Grieve (Nov 22 2015 at 00:06):
And here was me hoping you were debugging the build in terms of how it randomly fails often...
Josh Mandel (Nov 22 2015 at 00:07):
Oh, that would have been nice. How often? I didn't realize this was still an issue.
It's not clear from https://travis-ci.org/hl7-fhir/fhir-svn/builds how many of the failures are for a reason, vs. for some flakiness in Travis CI. Do we have a test that can be applied to the logs to categorize?
Grahame Grieve (Nov 22 2015 at 00:11):
Well, I look at the logs when it fails. Usually, at the moment, it fails because it can't read a file somewhere. But the only way to know is to know that this isn't a rational failure point.
Josh Mandel (Nov 22 2015 at 00:11):
Sure, but if it's always like file.read[] FAILED or something, then we could at least count those up.
Grahame Grieve (Nov 22 2015 at 00:13):
I'll look back through the logs and see later
Travis CI (Nov 22 2015 at 00:19):
Author: josh_mandel
Commit @7324
Build status: Passed
James Agnew (Nov 22 2015 at 05:48):
Could we put fhir-svn into a different stream so it is easier to mute?
Josh Mandel (Nov 22 2015 at 07:37):
Yes! Will do.
Josh Mandel (Nov 22 2015 at 07:52):
I've created a setup a bit more like Skype, with commit + build notifications happening on the "committers" stream. But they could also be moved to separate streams entirely (like "auto-notifications"). Would that be better @James Agnew ?
Josh Mandel (Nov 22 2015 at 08:18):
Also, that said: @James Agnew why are you finding it easier to mute a stream than a topic? They're both mute-able in a very similar way.
James Agnew (Nov 23 2015 at 16:33):
Ahhh you're right @Josh Mandel, you can mute topics. I found the mute button in the "manage streams" dialog and that one doesn't show how to mute a topic. On the left you canthough
Last updated: Apr 12 2022 at 19:14 UTC
