<borneoa___>
PaulFertser: hello. Today's merge triggered clang build 3 times instead in 1. The first patch triggered it immediately, then time to build it and it found the others and builds again. And then a final build with no changes. I think the first two are spurious and only the last should be triggered.
<PaulFertser>
borneoa___: hey! So you expect some coalescing to happen, i.e. it should wait for a while after getting the information that master was updated to see if anything was added on top and build only after that? Makes sense. I'll take a look.
<PaulFertser>
borneoa___: how is the responsiveness so far?
<borneoa___>
PaulFertser: before there was a delay of maybe 30 minutes. It reduces the load on the server.
<borneoa___>
PaulFertser: it worked flawless today
d_olex has quit [Ping timeout: 272 seconds]
d_olex has joined #openocd
d_olex_ has joined #openocd
d_olex has quit [Ping timeout: 276 seconds]
renrelkha has quit [Ping timeout: 252 seconds]
renrelkha has joined #openocd
<PaulFertser>
borneoa___: OK, that's interesting. Now I see that the builds are supposed to be working in a way I wasn't aware of: "openocd" builds on trigger from merge hook on Gerrit and it has a "quiet" period of one hour. If it succeeds then openocd-clang and openocd-build (all test subtargets) are run. The week before the trigger from Gerrit wasn't working at all due to changes that made security token
<PaulFertser>
mandatory but I also added a change to openocd-clang itself not knowing it would be auto-triggered. So it was built twice by the changes and then once again after "openocd" succeeded. The configuration fixed now, thank you!