IRC logs for #trustable for Thursday, 2018-09-20

*** iker has joined #trustable06:57
*** paulwaters_ has joined #trustable07:42
*** toscalix has joined #trustable08:01
ikerpaulsherwood, check
ikerthat's the gitect evidence report08:08
ikerYou asked yesterday.08:09
paulsherwoodiker: ok, it's a start :)08:13
paulsherwoodlet's raise an issue on gitect... how to make evidence more user-friendly08:13
paulsherwoodbut i'd be ok to approve now i think08:14
ikergreat, I will write the issue now08:14
ikerI had a job in the pipeline which was stuck and it wasn't running, finally it failed and outputted this:
ikerwhich is the reason for that? how can I fix it?08:48
paulsherwoodiker: i think that's a gitlab infra problem09:06
paulsherwoodi've hit retry09:06
paulsherwoodwhile it's great that gitlab provides so much free capacity for ci, we can't expect it to be 100% reliable/performant09:07
*** Shaun has joined #trustable09:09
ikera ok, good to know09:09
flatmushSTANDUP! Order: flatmush, Shaun, iker, ctgriffiths09:43
flatmush## Ben Brewer (flatmush)09:43
flatmush* Done09:43
flatmush  - Wrestle with SYSLINUX, to no avail.09:43
flatmush  - More updates to freedesktop-sdk junctions branch.09:43
flatmush  - E-mail response to olwen about standards.09:43
flatmush  - Talked with jjardon about moving some sources to git in freedesktop-sdk09:43
flatmush* Doing09:43
flatmush  - Summary e-mail about progress on minimal-distro to trustable distros list09:43
flatmush  - Submit my git sources patch to freedesktop-sdk to update some sources to git.09:43
flatmush  - Update busybox and minimal-distro runtime to use freedesktop-ddk09:43
flatmush  - Help ctgriffiths with looking into using gummiboot/systemd-boot.09:43
flatmush## Shaun Mooney (Shaun)09:43
Shaun* Done09:44
Shaun  - Project catchup09:44
Shaun  - AV STPA losses and hazards09:44
Shaun* Doing09:44
Shaun  - STPA Reading09:44
Shaun    - Re-reading handbook in more detail09:44
Shaun  - Working through primer09:44
Shaun* Next09:44
Shaun  - AV STPA responsibilites and control actions09:44
Shaun* Issues09:44
Shaun  - None09:44
Shaun## Iker Perez ( iker )09:44
iker* Done09:44
iker   - Debug gitect code in av-stpa and minimal-distro repos09:44
iker* Doing09:44
iker    - Change the way in which fredesktop elements are installed09:44
iker# Craig Griffiths ( ctgriffiths )09:45
ctgriffiths  * Done09:45
ctgriffiths    - Looking in to build issues and updated issue report:09:45
ctgriffiths    - Spinning up on buildstream and free desktop SDK.09:45
ctgriffiths  * Doing09:45
ctgriffiths    - Try alternatives to SYS-LIINUX as there are problems building it09:45
ctgriffiths      - systemd-boot looks like the best candidate09:45
ctgriffiths## Discussion09:45
flatmushAny points?09:46
flatmushSTANDUP ENDS!09:46
* Shaun wasn't given much of a chance, but had no points anyway09:47
paulsherwoodShaun: need to keep up :)09:47
flatmushYou snooze, you lose!09:47
paulsherwoodiker, flatmush - coming back to this stdout munging thing. if we can get uname output, why can't we get the test output?09:49
paulsherwoodand if the output doesn't provide anything useful, is that an issue for the test?09:50
flatmushwe do uname and just grep for "Linux", we don't really get the output09:50
flatmushwe need to get the proper output, anything else is a hack09:50
*** iker is now known as ikerperez09:50
flatmushit is technically possible to print some stuff out on the qemu image and then use that to extract the output exactly including an error code, but it's a really messy option09:51
* paulsherwood doesn't understand09:51
flatmushwe run our stuff in an emulator, the only I/O we have is the terminal from that emulator09:51
flatmushcurrent image can't trivially have ssh09:51
flatmushin an ideal world, we'd just run tests via ssh and get their exit-code and i/o09:52
paulsherwoodhow does uname output end up in the terminal?09:52
flatmushthe test effectively "types" "uname -a" into the terminal09:52
paulsherwoodand why can't latency_measure output be treated the same way?09:52
flatmushand then checks to see if the string "Linux" appears at any time withing the next 60 seconds09:52
flatmushif dmesg came up with something that said Linux, then the test would pass but be incorrect09:52
paulsherwoodjust grep for "max latency" or whatever the test reports?09:52
flatmushthe terminal gets all output, not just the program that runs09:53
flatmushthe test reports lots of data09:53
flatmushit would be a very hacky script to ensure we only get the latency test data and not other printouts09:53
flatmushand in the end it would gain us very little09:53
flatmushbecause we'd have to write that for every test09:53
paulsherwoodright. for now i'm just suggesting take the one line that reports max latency. it gives us *something*09:53
flatmushikerperez: Do you want to give that a try? ^^09:54
paulsherwoodi take your point that this is not 'production' or 'scaleable' but i'd rather have some example tests than not09:54
ikerperezI was going to go for the task of changing the freedesktop sources, but I haven't started yet so I don't mind doing that09:56
ikerperezI will do whatever you guys thing is a higher priority09:57
*** paulwaters_ has quit IRC10:01
ikerperezI will go for the max latency report10:06
*** paulwaters_ has joined #trustable10:55
paulsherwoodchange fsdsdk sources11:36
paulsherwoodis higher priority11:36
ikerperezI have already done something to print the latency, is running in the CI right now.11:38
ikerperezIt is not going to use the maximum latency to evaluate the validity of the test but at least we will be able to read it11:39
ikerperezIt's just an approximation11:40
*** paulwaters_ has quit IRC15:59
*** ikerperez has quit IRC16:02
*** Shaun has quit IRC16:20
*** toscalix has quit IRC17:26

Generated by 2.15.3 by Marius Gedminas - find it at!