| @stekern | juliusb: I just found an excellent example where one of your "finish assertments" are needed | 05:56 |
|---|---|---|
| @stekern | I wrote a simple lsu test, that just stores a value and then load it in the next instruction, in the trace the register ends up as 'x', generating flag = 'x' when comparing that register with a known value | 05:58 |
| @stekern | and apperantly l.bf and l.bnf will never branch when flag == 'x' | 05:59 |
| @stekern | so my test falsely passes | 05:59 |
| @stekern | so a sanity check for 'x' on the flag in the is needed | 05:59 |
| -!- ServerMode/#openrisc [+o ChanServ] by leguin.freenode.net | 12:16 | |
| -!- Netsplit *.net <-> *.split quits: heroux, O01eg | 12:20 | |
| -!- Netsplit over, joins: O01eg, heroux | 12:22 | |
| -!- Netsplit *.net <-> *.split quits: bradfa | 12:23 | |
| -!- Netsplit over, joins: bradfa | 12:24 | |
| -!- Netsplit *.net <-> *.split quits: heroux, O01eg | 12:26 | |
| -!- Netsplit over, joins: O01eg, heroux | 12:27 | |
| @juliusb | stekern: yeah we should have that kind of stuff throughout the model | 12:36 |
| @stekern | I've added that assert to my local repo, now the test fails ;) | 12:40 |
| @juliusb | :) | 12:51 |
| @juliusb | this is exactly the thing kind of thing which is useful | 12:51 |
| @juliusb | because, putting in things into the RTL to check is better than doing a corner case software test | 12:52 |
| @juliusb | it's nice to have that corner case software test, but the much more useful thing is the assert | 12:52 |
| @stekern | I agree, they should complement each other | 13:43 |
| @olofk | Hey, all weekend I was complaining about how insanely complicated some things are in git, but everyone was telling me that it was the simplest thing ever to use | 19:22 |
| @juliusb | haha yep | 20:22 |
| @juliusb | I still don't understand what I did yesterday, haven't reflected on it too much though | 20:22 |
| @olofk | whoops... how do I remove a commit that has been pushed to another repo? | 21:30 |
| @olofk | github in this case | 21:30 |
| @olofk | pheww... fixed it | 21:35 |
| _franck_ | on my openOCD repo, I didn't keep the git log history | 23:46 |
| _franck_ | so Marek Czerski helped me to restart from julius' repo | 23:46 |
| _franck_ | then include my changes | 23:47 |
| _franck_ | now I merged the last openOCD git into a new github repo | 23:47 |
| _franck_ | now, my question: | 23:47 |
| _franck_ | can I rename the Franck79/openOCD repo to Franck79/openOCD_old safely (as it's forked at openrisc github) | 23:48 |
| _franck_ | then I will rename the new repo (up to date with full log history) Franck79/openOCD | 23:49 |
| _franck_ | and stefan could delete the github openrisc/openOCD repo dans fork it again from my Franck79/openOCD repo | 23:50 |
| _franck_ | are we ok ? | 23:50 |
Generated by irclog2html.py 2.15.2 by Marius Gedminas - find it at mg.pov.lt!