Wednesday, 2021-10-13

VeeraHi12:39
Veeralkcl: did the problem sort out12:39
lkclVeera, no.13:42
lkcli re-ran the script, it failed.13:42
lkcli did not alter the script13:43
lkclcapnp files are not being found13:43
lkclthey are being installed in /usr/local/share/ (somewhere)13:43
lkclwhere the compile is expecting them to be in /home/lkcl/src/.....13:44
lkcltell you what: i will completely blow away everything and run it entirely again13:44
Veeralkcl: please wait13:46
Veeralkcl: is manually doing steps is painful13:46
lkcli haven't done any manual steps.13:47
lkcli'm running: mk-deb-chroot symbiflow-clean13:47
lkclschroot -c symbiflow-clean13:47
lkclthen sudo bash and install-symbiflow13:47
Veerawhile you were running it, I updated symbiflow-install13:48
Veerathe update fixes exactly the problem13:48
lkcl-cp libs/libvtrcapnproto/gen/rr_graph_uxsdcxx.capnp /usr/local/symbiflow/share/vtr13:50
lkcl+cp ../libs/libvtrcapnproto/gen/rr_graph_uxsdcxx.capnp /usr/local/symbiflow/share/vtr13:50
Veeralkcl: can you look at HDL_workflow/symbiflow page and symbiflow-install script for comments and review13:50
Veerayes13:51
Veerathe way you are running, i have not done that!13:52
VeeraI used debootstrap and manual install13:52
Veeraand then made the script manually13:52
Veeraso it has errors13:52
lkclok that was 3 days ago.  it would have been helpful to have put that as a comment ("i have updated the copying of the capnp file, this is the commit diff link")13:53
lkclokaaay13:53
lkclwell it's good that i'm running the script then :)13:53
lkclok that's underway - completely new schroot.13:54
VeeraIs it that we must not use buster-updates main in apt sources!!!13:57
Veerain test debootstrap13:57
lkclit's fine13:58
lkclhmmm buster-updates, no13:58
lkclthat's a "moving target" (non-reproducible)13:58
lkclbut backports is fine13:59
Veerathe https://bugs.libre-soc.org/show_bug.cgi?id=654#c15 comment is not that much explicitly clear13:59
lkclwhich is probably why i missed it, doh.13:59
Veeraok13:59
lkcledited14:00
lkclwe're fairly low on budget on #384.  can you do python?14:01
VeeraI scripted in python for my net modem and few others14:01
lkclif i give you a "slightly boring but quite straightforward and important repetitive task" would you be willing to give that a shot?14:01
lkclthe unit tests - and i apologise here, there's a *lot* of them - need to be updated to an "expected results" format14:02
VeeraWell reading documentation of python is very clear and easy to understand14:02
Veeracan I know the prerequisite sw for doing that? means system requirements14:03
VeeraI will try14:03
Veeramy mobile is low in power! am using it for net con14:04
lkclhttps://git.libre-soc.org/?p=openpower-isa.git;a=blob;f=src/openpower/test/alu/alu_cases.py;h=13efdd53f9ae699c36eacefa064fa44797c46104;hb=951dd564494af57dde0c439604f529f6b007bb98#l11514:04
lkclok14:04
Veerairc chews my battery power14:04
lkclzowee14:04
lkclit shouldn't.14:04
lkclhttps://libre-soc.org/HDL_workflow/devscripts/14:04
lkclon a (new?) chroot on TALOS-II workstation14:05
lkclif it gets too low you can check irc logs later14:05
Veeraok14:05
Veerai have matrix14:05
lkclsee that "expected" results?14:05
Veerait logs everything14:05
lkclahh14:05
lkclyyeahh...14:05
lkclthat example  case_addis_nonzero_r0_regression()14:06
lkclit was the very first example which added "expected results"14:06
lkclkylel should be able to guide you through it14:06
lkclmaybe work together on this?14:07
lkclall the unit tests in that file need converting.14:07
lkclit means either calculating the answers by hand14:07
lkclor14:07
lkcl(better)14:07
lkcllooking at the debug log output (which i apologise is pretty verbose, but you'll get the hang of it)14:07
Veeraok14:08
lkcland on completion of each test, extract the results that were produced14:08
lkcland literally just put the... well... "expected" results into the unit test.14:08
lkclto save yourself some time, we have a trick14:09
lkcl(don't actually commit this!)14:09
lkclglobal search/replace ":%s/def case_/def nocase_/g"14:09
lkclthen re-edit to enable the *one* test you want to run14:09
lkclonce test/alu/alu_cases.py is done, there are... deep breath... 10 more (!!!)14:09
Veerawant to input more? or that's it?14:11
VeeraI try to understand and do!14:12
lkcl:)14:14
lkclcan you raise a bugreport and cross-ref the IRC log here?14:14
lkcl"adapt ALU test cases to include expected results"14:14
lkclhttps://libre-soc.org/irclog/%23libre-soc.2021-10-13.log.html#t2021-10-13T14:04:2914:15
lkcldo each case_xxx one commit at a time.  make sure to review with "git diff" *before* committing.14:15
ghostmansd[m]lkcl: I just pushed EXTSXL and related changes for extsb/extsh/extsw, test_issuer works flawlessly19:47
ghostmansd[m]So, the next task is ISACaller update?19:47
ghostmansdlkcl: I've raised the task for ISACaller: https://bugs.libre-soc.org/show_bug.cgi?id=72820:01
ghostmansdPlease provide the details and assign the budget when you have time to look at it20:01
lkclghostmansd, brilliant21:10
lkclyes, ISACaller update.21:11
lkclcan you remember the day we talked? irclogs.21:11
lkclgot it21:12
lkclghostmansd, EUR 1000 budget for it. we've room to do a decent job, add a stack of docstrings etc. etc.21:15
ghostmansd[m]lkcl: sure I recall, I even posted link to IRC logs in task :-)21:22
ghostmansd[m]That said, quite likely there be more questions once I dive into the code. call() is quite big and complex, so some questions on how to organize it best may arise in course of this task.21:23
ghostmansd[m]I took a quick look today, but this obviously is not sufficient to make any sort of architecture-wise plans.21:24

Generated by irclog2html.py 2.17.1 by Marius Gedminas - find it at https://mg.pov.lt/irclog2html/!