01:59 <re_irc> <@firefrommoonlight:matrix.org> adamgreig: This is actually a great point

01:59 <re_irc> <@firefrommoonlight:matrix.org> That I can't solve with just more HP and may actually break the noise-reducation algo I'm working on

02:04 <re_irc> <@firefrommoonlight:matrix.org> Basically, I'm using something like a Wiener filter to attenuate "mostly noise" freq bands, by considering noise the floor of a given freq band as noise. To make this work well, I need precise frequency bands. I have them set up logarithmically, but filter resolution seems to go linear with freq

02:04 <re_irc> <@firefrommoonlight:matrix.org> So, this means you need lots of taps to get good freq resolution at low freqs

02:27 <re_irc> <@firefrommoonlight:matrix.org> adamgreig: Looked into the FMAC (Filter math accelerator periph). Looks perfect.

02:28 <re_irc> <@firefrommoonlight:matrix.org> I'll gradually start building a lib for high-level support when I get around to it. I'm prototyping on an H743 that doesn't have it though

02:28 <re_irc> <@firefrommoonlight:matrix.org> and I don't think the PAC supports it yet (Maybe SVD isn't out?)

02:28 <re_irc> <@firefrommoonlight:matrix.org> *Actually, PAC has H735

02:30 <re_irc> <@firefrommoonlight:matrix.org> And G4 has it too

05:43 <re_irc> <@9names:matrix.org> martin_lfs:matrix.org: I haven't played with meta-rust, but it sounds like you're either not adding the package to the image, or not building the image.

05:43 <re_irc> <@9names:matrix.org> did you add rust-hello-world to the IMAGE_INSTALL_append line in your local.conf, then build the image again?

09:14 <re_irc> <@emilgardis:matrix.org> therealprof:matrix.org: has this been done?

10:29 <re_irc> <@thejpster:matrix.org> Please can we do a cortex-m-rt patch release (0.7.1?) with the stack unwinding fix? Or I guess my question is *who* can do a cortex-m-rt release?

10:41 <re_irc> <@jamesmunns:beeper.com> thejpster it looks like both of your accounts are on the `cortex-m` team (https://github.com/orgs/rust-embedded/teams/cortex-m), which is authorized as a publisher of the `cortex-m-rt` crate: https://crates.io/teams/github:rust-embedded:cortex-m

10:41 <re_irc> <@jamesmunns:beeper.com> So... you can?

10:42 <re_irc> <@jamesmunns:beeper.com> You'll just need a PR and review from adamgreig, therealprof, or thalesfragoso.

10:43 <re_irc> <@jamesmunns:beeper.com> Jonas isn't a part of the team any more, but I imagine you could ping him privately to walk you through the release process if you have questions, or I'm sure Adam could help with that.

10:43 <re_irc> <@thejpster:matrix.org> Is our release process documented somewhere?

10:44 <re_irc> <@jamesmunns:beeper.com> That: I'm not sure of!

10:44 <re_irc> <@jamesmunns:beeper.com> Adam or Jonas would probably know off the top of their heads.

10:58 <re_irc> <@adamgreig:matrix.org> Check the most recent release pr is probably the most accurate guide... update changelog including link at bottom, update cargo.toml and macros/cargo.toml, PR, once merged git tag, make release on GitHub, publish to crates.io

11:01 <re_irc> <@adamgreig:matrix.org> (https://github.com/rust-embedded/wg/blob/master/ops/publish.md is the written guide but it doesn't cover c-m-rt's specific need to update both cargo.toml)

11:05 <re_irc> <@adamgreig:matrix.org> I guess it's debatable if the macros crate really needs a version bump and publish if nothing changes but then it would get out of sync with the main crate version number...

11:08 <re_irc> <@thejpster:matrix.org> Does macros need bumping always?

11:08 <re_irc> <@thejpster:matrix.org> Oh wait, I need to finish reading before replyhing

11:13 <re_irc> <@adamgreig:matrix.org> Maybe we can only bump it when it changes, but then bump it to the current c-m-rt version number

11:16 <re_irc> <@thejpster:matrix.org> That makes sense. I don't like bumping for bumping's sake.

11:16 <re_irc> <@thejpster:matrix.org> defmt is like five crates, and I have the same questions about that (as it needs a bug-fix in one of them).

11:23 <re_irc> <@jamesmunns:beeper.com> IMO you could "skip" versions instead? I wonder if you publish `1.0.1` and `1.0.4` (but no 2 or 3), if specifying `1.0.3` would still work

11:26 <re_irc> <@thejpster:matrix.org> ^1.0.3 should work? =1.0.3 wouldn't.

11:30 <re_irc> <@jamesmunns:beeper.com> yeah, fair

11:40 <re_irc> <@9names:matrix.org> 9names: oh, they changed the syntax for it - that's fun!

11:40 <re_irc> <@9names:matrix.org> if you're running a newish build of poky you need to use IMAGE_INSTALL:append

11:40 <re_irc> ... long message truncated: https://psion.agg.io/_matrix/media/r0/download/psion.agg.io/3615e44f720ec10a9f867d6365bd6fa64dee2ab2c8d2f9258d177655a92a2433 (12 lines)

11:40 <re_irc> <@9names:matrix.org> IMAGE_INSTALL:append = " rust-hello-world"

11:40 <re_irc> <@9names:matrix.org> tail conf/local.conf -n1

11:45 <re_irc> <@9names:matrix.org> `nice` is technically optional, but if you're like me and prefer your computer to stay usable while it's crunching away at a compile job for hours on end it's worth putting in there

13:43 <re_irc> <@thejpster:matrix.org> Wish I knew the Windows equivalent. Kicking off a `cargo build` in Windows Terminal using Powershell slows down my machine to the point I can see Windows Terminal visibly repaint.

13:50 <re_irc> <@dngrs:matrix.org> I think the Windows equivalent is called WSL2 +runs+

13:50 <re_irc> <@dngrs:matrix.org> on a more serious note, [this](https://stackoverflow.com/a/4332) maybe?

13:51 <re_irc> <@thejpster:matrix.org> START is a cmd.exe built-in unfortunately.

15:02 <re_irc> <@glaeqen:matrix.org> Sorry in case this topic was already discussed before but did anyone manage to get pretty printing to work with embedded ELFs? On x86-64, ELF binaries are generated with appropriate `.debug_gdb_scripts` section containing a list of scripts that are supposed to be auto-loaded when opened with GDB (namely `gdb_load_rust_pretty_printers.py`). Usually GDB will complain because of misconfigured safe-paths but this..

15:02 <re_irc> ... is where `rust-gdb` comes into play and running it instead of vanilla `gdb` makes scripts to load and pretty printing functions properly.

15:02 <re_irc> <@glaeqen:matrix.org> in case of `thumbv7em-..` targets, this section is not generated so these scripts are never loaded (therefore `rust-gdb` with `RUST_GDB` variable set won't help). Moreover, when I tried to manually hack it together via PYTHONPATH set to `$HOME/.rustup/toolchains/<toolchain>/lib/rustlib/etc/` and sourcing the `gdb_load_rust_pretty_printers.py` manually it still fails because it uses a global...

15:02 <re_irc> <@glaeqen:matrix.org> However,

15:02 <re_irc> ... long message truncated: https://psion.agg.io/_matrix/media/r0/download/psion.agg.io/b687b14b0d1a3c69cc18cabf842ddb64cf18df741442c063bede0b58e924f1ef (6 lines)

15:34 <re_irc> <@martin_lfs:matrix.org> 9names: So, I think I figured out the issue. The board has two boot partitions so upgrades are loaded onto the "other" partition. The new image was actually corrupted so the board tried to run the new image multiple times, then flipped back to the old partition. I wasn't putting the package in `image_install:append`, so I'll try that now.

15:42 <re_irc> <@martin_lfs:matrix.org> 9names: Yup, our team just upgraded to honister, so we have the new overload format.

15:59 <re_irc> <@martin_lfs:matrix.org> martin_lfs:matrix.org: KK, retried the install and it produced a corrupted image again, so since I was just trying to do a proof of concept for the RnD staff maybe I'll keep it simple and just make a minimal image on qemu instead of real hardware.

15:59 <re_irc> <@jamesmunns:beeper.com> btw, started writing a newsletter for all my current projects, if y'all are interested, the first issue is up here: http://createsend.com/t/y-7DD5C7B82359F7EB2540EF23F30FEDED

16:00 <re_irc> <@jamesmunns:beeper.com> Also feel free to drop by #onevariable:matrix.org if that's your kind of topic :D