F-Droid Devs

116 Members
F-Droid development discussion only | Use #fdroid:f-droid.org for general, app- and repo-related matters | Meeting every Thursday at 11:30 UTC | This channel is publicly logged at https://matrix.f-droid.org/alias/%23fdroid-dev:f-droid.org 20 Servers

Load older messages


SenderMessageTime
5 Jun 2020
@eighthave:matrix.org_hcoh you mean downloads18:28:42
@bubu:bubu1.euBubuyeah18:28:45
@bubu:bubu1.euBubutried a few mirror. one worked with 2 mb/s finally18:28:59
@eighthave:matrix.org_hchmm, strange, I've seen 40MB/s on gitlab-ci from deb.debian.org using apt18:29:15
@bubu:bubu1.euBubuAll pending jobs got canceled but the running jobs are starting to finish: https://gitlab.com/fdroid/fdroiddata/-/pipelines/15340372318:31:59
@bubu:bubu1.euBubuactually faster than I expected after that start18:32:26
@bubu:bubu1.euBubuthe network right now seems much faster then 5 hours ago18:33:57
@freenode_cdesai:matrix.orgcdesaiupdating the ci-images-server could also give a slight speedup, not having to install all those apt updates18:36:37
@eighthave:matrix.org_hcI donn't think checkupdates needs that big image18:37:05
@eighthave:matrix.org_hcdoes it need the Android SDK/gradle/etc?18:37:13
@bubu:bubu1.euBubuI don't think so!18:39:53
@bubu:bubu1.euBubulike I said the image size isn't a problem as it's cached on the runner18:40:10
@bubu:bubu1.euBubuonly the initial package install time matters really18:40:20
@bubu:bubu1.euBubuwhich currently takes about 8 minutes on that runner18:40:37
* @bubu:bubu1.euBubu is very confused why all the jobs got canceled18:40:55
@bubu:bubu1.euBubuthey should have waited quite awhile longer for the runner to pick them up before getting failed18:41:21
@bubu:bubu1.euBubuno idea why they got canceled18:41:27
@freenode_cdesai:matrix.orgcdesaia new ci-images-something with the packages installed would save those 8 minutes then18:43:08
@eighthave:matrix.org_hcI've seen huge time differences due to image size on both gitlab.com runners and ours18:46:36
@eighthave:matrix.org_hclike 20 seconds to set up a light images vs 3-4 minutes on a heavy one18:46:59
@eighthave:matrix.org_hc Bubu: if you are going to run it on gcc148, why use gitlab-ci at all? you could use straight docker 18:48:28
@eighthave:matrix.org_hcprep the image once each time locally, then run multiple jobs18:48:56
@bubu:bubu1.euBubu
In reply to @eighthave:matrix.org
Bubu: if you are going to run it on gcc148, why use gitlab-ci at all? you could use straight docker
because then everyone can see the run logs
18:49:01
@bubu:bubu1.euBubuthe image is cached locally it isn't redownloaded18:49:30
@eighthave:matrix.org_hcI mean prep it with all the installs, updates, etc.18:50:10
@bubu:bubu1.euBubuYes, that's on the todo list18:50:48
@eighthave:matrix.org_hcyou could do a cron'ed run with straight docker that rsyncs the build logs somewhere public18:53:35
@bubu:bubu1.euBububut why?18:53:47
@bubu:bubu1.euBubuif we can just use gitlab CI18:53:57
@eighthave:matrix.org_hcbecause it'll simplify things a lot18:54:04

Show newer messages


Back to Room ListRoom Version: 1