Where communities thrive


  • Join over 1.5M+ people
  • Join over 100K+ communities
  • Free without limits
  • Create your own community
People
Activity
Lucas Di Pentima
@ldipenti
Yeah, that’s something that smells a little fishy to me. Removing whitespaces/newlines on a file to be able to compare it doesn’t seem a great approach to me.
Tom Clegg
@tomclegg
@ldipenti aha, thanks
Nico Cesar
@nicocesar
@tomclegg mmmg check 123e39f45192118a1e84629bf94a3a7db2bd5014 and ba87b0cb04e4a7e8e837567d9a6e317f3d5fa0b2
I don't think adding it to Call() (or CallRaw as a matter of fact) is what we want
so I separated it in 2 commits and you can see the different results
ba87b0cb04e4a7e8e837567d9a6e317f3d5fa0b2 still has the 60s delay, because of retries
oopps somthign in sdk/go/keepclient/keepclient.go was not commited. give me asec
Tom Clegg
@tomclegg
Nico Cesar
@nicocesar
I think c41f3b2cb 17530-arvados-client-fastfail-take2 meets both criterias
Tom Clegg
@tomclegg
yeah
Nico Cesar
@nicocesar
@tomclegg I think we can merge https://dev.arvados.org/issues/17530
(tests passed)
Tom Clegg
@tomclegg
The commit message says CallRaw() will take 60s to fail... that is referring to the old behavior which has now been fixed, right?
Nico Cesar
@nicocesar
@tomclegg if we only have CallRaw() that will be the behaviour but since we have both changes, from arvados-client wont happen.
but if rawcall() is called will show that warning
does it make sense?
Tom Clegg
@tomclegg
@nicocesar thanks!
Tom Clegg
@tomclegg
@nicocesar on 17530 -- the code makes sense, just not the commit message. Not worth worrying about. You should merge it!
Nico Cesar
@nicocesar
ok will do
Nico Cesar
@nicocesar
Tom Clegg
@tomclegg
Tom Clegg
@tomclegg
@nicocesar https://dev.arvados.org/issues/17437#note-21 ... add Content-Length to "dropHeaders" doesn't cause any failures. Reading through the ticket history I'm not sure you ever tested whether that makes WEBrick work, but either way I think we should do it, because there's no need to interfere with Go's automatic handling of Content-Length, encoding, etc... WDYT?
Nico Cesar
@nicocesar
Maybe is not clear in the tickets, from what I remember.... and Explicit value in Content-Length (we took it from the request, but might not be the right thing to do) Plus TransferEncoding: []string{"identity"}, made it work wirh WEBrick
Tom Clegg
@tomclegg
ok, I'll do that. (I also edited the ticket description so it's more obvious that it's only an issue with webrick)
Nico Cesar
@nicocesar
thanks
Daniel Kutyła
@DannyNs
it required upgrade of cypress to latest as this issue was fixed recently
Lucas Di Pentima
@ldipenti
Great!, on it. Related note: while working on the cypress tests flakyness I discovered a bug in the cy.loginAs() call: when called with 2 different users on the same test, the 1st session leaked some data into the 2nd session (this may be also a sign that a bug exists on the /token?api_token=xxx handling code), and one of the favorites tests was relying on that bug to sometimes pass :)
Lucas Di Pentima
@ldipenti
@DannyNs You may want to fix this and re-run the tests :)
[lucas:~/Devel/git/arvados-workbench2] 17415-Mountainduck-Bookmark-files+* 3s ± git grep 'it\.only'
cypress/integration/collection.spec.js:35:    it.only('allows to download mountain duck config for a collection', () => {
[lucas:~/Devel/git/arvados-workbench2] 17415-Mountainduck-Bookmark-files+* 3s ±
Daniel Kutyła
@DannyNs
omg
sorry
Lucas Di Pentima
@ldipenti
np :D
Lucas Di Pentima
@ldipenti
Lucas Di Pentima
@ldipenti
@tetron 17500 is ready for review
@tomclegg I realized I have a pending review from you, sorry.
Lucas Di Pentima
@ldipenti
@tomclegg 17346 LGTM
Tom Clegg
@tomclegg
@ldipenti thanks! https://dev.arvados.org/issues/17343 is up for review too
Lucas Di Pentima
@ldipenti
Ok! on it.
Tom Clegg
@tomclegg
@nicocesar https://dev.arvados.org/issues/17204 is up for review... and replied on https://dev.arvados.org/issues/17566
Peter Amstutz
@tetron
@ldipenti 17500-cypress-flakyness-fixes LGTM
Lucas Di Pentima
@ldipenti
awesome, thanks!
Nico Cesar
@nicocesar
Tom Clegg
@tomclegg
thanks
Lucas Di Pentima
@ldipenti
Good morning @tomclegg , 17343 LGTM. Have you made some timing metrics? I tried making just one run of the Cypress tests on Jenkins and didn’t noticed time savings, I guess RailsAPI and Workbench1 share the majority of their gems deps?
Tom Clegg
@tomclegg
@ldipenti looks like the difference between builds 391 and 392 in "launching arvados in test mode..." is 3m33s - 1m50s = 1m43s ... there's still nearly 8 minutes of setup before any of the wb2 tests actually start, but it's something...
Lucas Di Pentima
@ldipenti
Right, didn’t see that. almost 2 minutes in ~17 total is not insignificant :)