00:48 <rqou> found myself an excuse to finally play with Rust

01:12 <cr1901_modern> I've been using it for small command line applications just to get a feel. But I'm def not using it to its full potential

01:35 <whitequark> rqou: yeah?

01:47 <jedb> eh, I find Ada to be much more readable

04:12 <rqou> why does rust still use the "curl | sh" installation method?

04:49 <lain> why not?

04:50 <lain> you're just going to wget && ./thing it anyway

04:50 <lain> well, ok, there should probably be a signature check in there :P

04:50 <lain> but /other than that/

05:12 <cyrozap> lain: First, because it encourages bad practices. Second, because what happens if the curl download dies in the middle of the script? Then you'll have a half-installed something-or-other, or you'll run a partial "rm -r ..." command.

05:12 <cyrozap> Also, to some people, it may not be obvious that piping to a shell is going to execute a bunch of commands.

05:14 <cyrozap> And if it's done over HTTP, you might as well just give a telnet-accessible root shell to everyone between you and the script server, including whoever happens to be poisoning the BGP route that day.

05:18 <cyrozap> The ideal situation is, you either `wget` or `curl -O` the script and its signature, `gpg --verify` the signature, `less` the script to make sure it isn't doing anything sneaky, and then run it.

07:23 <cyrozap> pointfree: I'm pretty much "done" with the KitProg driver at this point. Can you please try out the latest version and let me know if anything needs fixing?

07:23 <cyrozap> ^ That goes for anyone else with KitProgs as well.

07:28 <pointfree> cyrozap: Eagerly updating my openocd-kitprog-git Arch Linux PKGBUILD as we speak!

07:29 <pointfree> Are there any areas I should pay special attention to?

07:37 <cyrozap> pointfree: Well, if you haven't updated in a while, all the reset issues I'm aware of are now fixed, so OpenOCD will now reconnect if you hard-reset the target (toggling SRST/XRES/RST). Mostly I just want to make sure there aren't any regressions, and that gdb debugging, memory read/write, and flashing (using the unmerged flash driver) still work.

07:38 <pointfree> I use all of those things.

07:40 <cr1901_modern> azonenberg: Not ready to go into details, but I have concluded that industry inertia prevents me from using Ethernet as my data xfer medium :'(.

07:41 <azonenberg> :(

08:00 <pointfree> cyrozap: Is "openocd -c "interface kitprog; kitprog_acquire_psoc" -f target/psoc5lp.cfg ..." still the syntax?

08:02 <cyrozap> pointfree: Sorry, it's "kitprog_init_acquire_psoc" now, because there's also now a run-time "kitprog acquire_psoc" command (note the space).

08:04 <pointfree> Ah.

08:05 <pointfree> Also looks like "-f target/psoc5.cfg" instead of "-f target/psoc5lp.cfg"

08:08 <cyrozap> "-f target/psoc5lp.cfg" should still be correct, since that's the config in OpenOCD master.

08:11 <cyrozap> Oh, wait, are you using the stuff from my GitHub and not from OpenOCD Gerrit? Because it should all be from Gerrit.

08:12 <pointfree> Oh whoops.

08:17 <cyrozap> I just use the GitHub repo to backup my reasoning behind each individual commit and as a sort of staging area for the Gerrit stuff, and the PSoC driver branch is deprecated in favor of that other guy's stuff in Gerrit.

08:17 <pointfree> ah

08:20 <cyrozap> Also, if you want to do a hard reset, you don't do "kitprog reset_target" any more, you just configure reset with "reset_config srst_only separate srst_gates_jtag srst_open_drain connect_deassert_srst". Then, any "reset" call will do a hard reset instead of a soft reset.

08:24 <pointfree> good to know

08:27 <pointfree> cyrozap: There is a merge conflict with http://openocd.zylin.com/#/c/3432/ (flash/nor).

08:31 <cyrozap> Also also, please let me know what your KitProg's firmware version is when you try the driver. So far, I haven't seen any differences in behavior between 2.12, 2.14, and 2.16 (besides the CDC-ACM descriptor fix that happened in either 2.15 or 2.16), but it's still good to know just in case there are any firmware version-specific bugs.

08:34 <cyrozap> pointfree: What file did it conflict on?

08:38 <cyrozap> Ah, I see it

08:38 <pointfree> cyrozap: src/flash/nor/Makefile.am

08:39 <cyrozap> Yup

08:39 <pointfree> I have "KitProg v2.16" and "Hardware version: 17"

08:40 <pointfree> At some point I updated the KitProg firmware.

08:44 <cyrozap> That's fine. Anyways, to resolve the merge conflict, you just need to add "%D%/psoc5lp.c \" below "%D%/psoc4.c \" in the version of that file from the git HEAD. The error happened because the flash driver hasn't been rebased on master yet.

08:50 <cyrozap> pointfree: Here's the fixed patch you can apply with `patch -p1 < patchy_the.patch`: https://paste.debian.net/hidden/8fe8a8c5/

08:56 <pointfree> Thanks cyrozap.

08:56 <cyrozap> np

09:06 <pointfree> Everything seems to work great: flashing, setting the blue led on and off, gdb.

09:07 <pointfree> I just cut power to the PSoC board while keeping the KitProg plugged in...

09:07 <pointfree> ..."Polling again in 6300ms"

09:09 <pointfree> Oh right...

09:10 <pointfree> I need to power the PSoC before powering the KitProg.

09:11 <pointfree> (I was working on USB while using openocd, so I snapped off the KitProg and I only power through USB)

09:14 <cyrozap> Ah, yeah, you need to do another acquisition if you power off the PSoC

09:14 <pointfree> "OpenOCD will now reconnect if you hard-reset the target (toggling SRST/XRES/RST)"

09:14 <cyrozap> But now, you can just run "kitprog acquire_psoc" and it'll work again

09:15 <cyrozap> pointfree: Yeah, hard-reset using the reset pin, not powering off :P

09:15 <cyrozap> But it should still work if you run "kitprog acquire_psoc; reset"

09:15 <pointfree> cyrozap: The reset button on the KitProg works great. I can toggle it and it reconnects.

09:16 <pointfree> Oh-yeah!

09:17 <cyrozap> I just checked the "power off, re-acquire" and that works if you run those commands

09:19 <cyrozap> Actually, I don't even think you need the "reset" part

09:21 <cyrozap> Yup, you just need to re-acquire

09:21 <pointfree> awesome.

09:22 <pointfree> dmesg is getting spammed because my laptop is polling for setup with the nascent usb driver. I guess the only way forward is to finish it.

09:22 <cyrozap> I was really hoping to get rid of the separate kitprog_init_acquire_psoc, but OpenOCD has no way to run commands post-adapter-init, pre-target-communication.

09:24 <pointfree> should it?

09:32 <cyrozap> Maybe?

15:25 <pie_> you cannever haveenough hooks

22:11 <jrernst> @Marex: hi! are you listening?