| Time |
Nick |
Message |
| 07:33 |
|
kworstell-isl joined #evergreen |
| 07:54 |
|
kworstell_isl joined #evergreen |
| 08:03 |
|
kworstell_isl_ joined #evergreen |
| 08:04 |
|
BDorsey joined #evergreen |
| 08:16 |
|
collum joined #evergreen |
| 08:28 |
|
redavis joined #evergreen |
| 08:32 |
|
mmorgan joined #evergreen |
| 08:47 |
|
dguarrac joined #evergreen |
| 09:00 |
|
Dyrcona joined #evergreen |
| 09:47 |
|
sleary joined #evergreen |
| 09:49 |
|
collum joined #evergreen |
| 09:57 |
|
kworstell-isl joined #evergreen |
| 10:24 |
|
sandbergja joined #evergreen |
| 11:16 |
mmorgan |
gmcharlt++ |
| 11:17 |
|
jihpringle joined #evergreen |
| 11:17 |
|
collum joined #evergreen |
| 12:51 |
|
collum joined #evergreen |
| 13:11 |
sandbergja |
Bmagic: would you be able to check the docs server to see if it is re-generating the docs? After merging bug 2036328, I expected to see <meta name="generator" content="Antora 3.1.7"> in the HTML |
| 13:11 |
pinesol |
Launchpad bug 2036328 in Evergreen 3.11 "Upgrade antora to supported version" [Wishlist,Fix committed] https://launchpad.net/bugs/2036328 |
| 13:11 |
sandbergja |
But it still says 2.3.4 :-( |
| 13:12 |
Bmagic |
hmmm |
| 13:13 |
Bmagic |
one of the reasons I didn't upgrade was I thought* we also needed to match/pair the version of the UI. Maybe that's not required |
| 13:14 |
sandbergja |
Hmmm good point. FWIW, it does generate okay on my laptop and in github actions. |
| 13:14 |
Bmagic |
I was about to check that |
| 13:14 |
sandbergja |
Not to be the "but it worked on my laptop" person hahaha |
| 13:16 |
Dyrcona |
sandbergja: We'll just have to give the customer your laptop, then. :) |
| 13:16 |
sandbergja |
heh |
| 13:28 |
jeff |
(and thus, Docker was born) |
| 13:32 |
Dyrcona |
There are some things that Docker can't fix. |
| 13:38 |
Bmagic |
sandbergja: everything looks ok, I'm running it by hand now |
| 13:39 |
sandbergja |
Bmagic++ |
| 13:41 |
Bmagic |
system failed: 32512 at ./generate_docs.pl line 203. |
| 13:41 |
Bmagic |
digging |
| 13:47 |
jeffdavis |
I wonder if we should have a conversation at the dev meeting about barriers to getting stuff committed. |
| 13:48 |
jeffdavis |
Maybe that is a conference/hackfest discussion though. |
| 13:50 |
jeffdavis |
And maybe the recent work on improving our release process has exhausted our appetite for other process discussions. :) |
| 13:50 |
Dyrcona |
jeffdavis: Time... I need a clone. |
| 13:50 |
* sleary |
is always here for process discussions |
| 13:51 |
sleary |
At the last New Devs meeting, a lot of people said they had trouble just setting up test systems to work on. Docker helps, but it requires a ton of resources. |
| 13:52 |
sandbergja |
Bmagic: is node --version 16 or above on that box? |
| 13:56 |
Bmagic |
sandbergja, nope, 12.16.3 |
| 13:56 |
sandbergja |
I noticed that the new antora wants 16 or more: https://www.npmjs.com/package/antora/v/3.1.7?activeTab=code |
| 13:56 |
sandbergja |
in package.json > engines |
| 13:57 |
Bmagic |
the error is: ./node_modules/@antora/cli/bin/antora: not found |
| 13:59 |
Bmagic |
ok, we're up to 16.20, trying again |
| 13:59 |
sandbergja |
Bmagic++ |
| 14:00 |
Bmagic |
(it's ubuntu 18.04, so it couldn't go up to to v20, complaining about glibc 2.28 |
| 14:00 |
sandbergja |
ahh yeah, that sounds familiar |
| 14:01 |
Bmagic |
looking good now, still building, I think we're past the error |
| 14:01 |
sandbergja |
woohoo! |
| 14:01 |
Bmagic |
it does seem like there are a lot* more errors now |
| 14:02 |
sandbergja |
Yeah, I noticed that too. And it is more verbose about each one (saying which branch the error comes from, which the old version did not) |
| 14:03 |
Bmagic |
it's done, site updated. The first thing I'm noticing is the search box moved |
| 14:03 |
sandbergja |
Bmagic++ |
| 14:04 |
Bmagic |
for better or for worse, the search behavior has reverted to it's old self where it results all* versions of Evergreen. I had put in some effort to make it only search latest |
| 14:04 |
|
jvwoolf joined #evergreen |
| 14:05 |
sandbergja |
I can launchpad and work on that, Bmagic. Since I'm the one who broke it :-( |
| 14:05 |
Bmagic |
:) |
| 14:05 |
Bmagic |
not sure it's broken, but I personally think it's cluttery |
| 14:06 |
sandbergja |
Totally, especially since it's so similar between versions |
| 14:07 |
Bmagic |
noting that the Windows patch bug 1930099 has a conflict now :) |
| 14:07 |
pinesol |
Launchpad bug 1930099 in Evergreen "generate_docs.pl should be able to run on Windows" [Medium,Confirmed] https://launchpad.net/bugs/1930099 |
| 14:07 |
sandbergja |
I'm wondering, also, could we maybe make an ansible playbook for that server, and include it in the repo? It would make it easier (for me at least) to try out changes in a production-like environment, without breaking it in prod? |
| 14:08 |
sandbergja |
Classic git :-D |
| 14:09 |
jeffdavis |
sleary: Is the issue that installing EG is difficult and complicated, or is it more that people don't have access to environments where they're able to do so? |
| 14:10 |
Bmagic |
sangbergja: there is a bash script that started pretty simple and has become fairly complex, and is not* revision controlled. It should be exposed somewhere |
| 14:11 |
Bmagic |
including it in the Evergreen repository doesn't seem like the right thing because it references local folder paths |
| 14:11 |
Dyrcona |
Bmagic: Those local folder paths could be made configurable via autotools. |
| 14:12 |
Dyrcona |
Or could be parameterized in the script. |
| 14:14 |
Bmagic |
generate_docs.pl is in the repo and sanitized, and allows the parameters to come in. Which is where the bash scrip comes in, to fill in the parameters for the specifics of the local server |
| 14:15 |
Bmagic |
this: https://pastebin.com/pBcQT6g5 |
| 14:40 |
sandbergja |
jeffdavis++ |
| 14:45 |
sandbergja |
I think it would be a good discussion to have about barriers to committing. |
| 14:46 |
jeffdavis |
I'll add it to the dev meeting agenda |
| 14:46 |
jeffdavis |
I have a section of my notes called "Fear of commitment" |
| 14:46 |
sandbergja |
hahaha |
| 14:46 |
jeffdavis |
:) |
| 14:47 |
Dyrcona |
What do we mean "barriers to committing?" I assume it was about committers not reviewing code for main. |
| 14:47 |
Dyrcona |
My number 1 barrier is time. |
| 14:48 |
sandbergja |
For me, lack of time and access to a system for qa are big. But I'd also say that merge conflicts and semantic conflicts (e.g. an upgrade script changes a stored procedure, but that stored procedure has changed in the meantime, so committing it would cause a regression) also make things harder than they need to be |
| 14:49 |
sandbergja |
Also, I'd add that it takes a while to re-load concerto, or to manually run the automated tests |
| 14:50 |
Dyrcona |
For merge conflicts and semantic issues, just throw back at the original developer and ask for a rebase if you're not comfortable fixing the conflicts yourself. |
| 14:51 |
jeffdavis |
My other big ones are (1) no test plan, and/or tests are complicated to set up; (2) unresolved questions about the fix; (3) extra work to backport and uncertainty about whether to do so. |
| 14:51 |
jeffdavis |
Time is the big one, but the others all make the process even more time-consuming. |
| 14:52 |
Dyrcona |
yeah, testing is often difficult for me. I usually avoid staff client stuff because I never use it, so I often have trouble testing staff client changes. |
| 14:55 |
jeffdavis |
I wonder if some kind-hearted organization would be able to make a limited number of VMs available to committers to help mitigate the lack of access to test environments. |
| 14:56 |
jeffdavis |
(not thinking of any particular organization tbh, actually wondering if this is something my org could help with) |
| 14:59 |
Dyrcona |
My org is trying to get away from being in the hosting business. |
| 15:00 |
jeffdavis |
I know some wonderful folks do already set up test environments pre-loaded with branches for bug squashing weeks, I'm wondering about something a bit more flexible as a supplement. |
| 15:05 |
Dyrcona |
yeah, that's what I thought you meant. |
| 15:43 |
|
jihpringle joined #evergreen |
| 16:49 |
|
jvwoolf joined #evergreen |
| 17:24 |
|
mmorgan left #evergreen |
| 18:36 |
|
jihpringle joined #evergreen |
| 19:04 |
|
jihpringle joined #evergreen |
| 19:46 |
|
jihpringle joined #evergreen |
| 21:06 |
|
sleary joined #evergreen |
| 23:07 |
|
sleary joined #evergreen |
