10:57 <JamesMunns[m]> Before I open an issue, is this a bug?... (full message at <https://catircservices.org/_matrix/media/v3/download/catircservices.org/WtESAHxNGOJeMcOHjZjlnVbB>)

10:58 <JamesMunns[m]> so __RESET_VECTOR is at offset 4 (correct!), but __reset_vector is at offset 8 (not correct!)

11:06 <JamesMunns[m]> the layout itself is right, I think the __reset_vector symbol is just wrong for some reason.

11:07 <JamesMunns[m]> I tried adding an NMI (the vector table goes SP, Reset, NMI), and it was in the right place.

11:14 <JamesMunns[m]> Opened https://github.com/rust-embedded/cortex-m/issues/495

11:23 <JamesMunns[m]> There's even an assert for the (IMO wrong) value 🙃... (full message at <https://catircservices.org/_matrix/media/v3/download/catircservices.org/VCAPYUGpOodToKJEvmaBVCiP>)

11:25 <JamesMunns[m]> git blame says it's been like that since Aug 7th, 2018 lol

11:25 <dirbaio[m]> lol

11:25 <dirbaio[m]> is __reset_vector actually used from any core?

11:25 <dirbaio[m]> * is __reset_vector actually used from any code?

11:25 <dirbaio[m]> perhaps it's just a bad name

11:25 <JamesMunns[m]> nope, it's just that assert AFAICT

11:26 <JamesMunns[m]> and that assert is really only checking that the vector table exists

11:27 <JamesMunns[m]> the __ symbols are considered private, and the lowercase ones aren't even discussed: https://docs.rs/cortex-m-rt/latest/cortex_m_rt/#symbols-objdump-nm

11:27 <dirbaio[m]> perhaps that's why the confusing name? "I need a symbol to check the reset vector exists, so I'll name it __reset_vector"

11:27 <JamesMunns[m]> yeah, I don't think this is likely to cause any real problems, it's mostly confusing IMO

11:27 <dirbaio[m]> perhaps rename it __exceptions

11:28 <JamesMunns[m]> esp since we have both __RESET_VECTOR and __reset_vector and they have different values lol

11:28 <JamesMunns[m]> dirbaio[m]: I'll add that as a comment

11:28 <dirbaio[m]> or perhaps remove it and make the assert ASSERT(__EXCEPTIONS == ADDR(.vector_table) + 0x8, "

11:29 <JamesMunns[m]> or move it to above the KEEP, so it actually points there

11:30 <dirbaio[m]> but then you're not checking that there's a reset vector with the assert

11:32 <JamesMunns[m]> Yeah, probably good to discuss on Tuesday in the meeting and pick something, I think there are lots of "right" options, the goal is to just make it not confusing

11:32 <JamesMunns[m]> if it hasn't bothered anyone in 5 years tho, I doubt it's urgent lol

12:29 <AdamHott[m]> hey all, just wanted to introduce myself. I'm Adam Hott. I've been programming web and cloud (AWS) for quite some time. Started studying Rust about 6 months ago and really love it. I'm down the rabbit hole now and decided I want to focus on embedded systems. I'm not sure I'll have much to contribute at this point, I'd like to help if I can.

12:31 <AdamHott[m]> If there are any small tasks that could be helpful, or research needed, I'd be happy to assist.

12:32 <AdamHott[m]> You can learn more about me here: https://www.adamhott.io/

13:03 <adamgreig[m]> <JamesMunns[m]> "Opened https://github.com/rust-..."; <- lol, talk about a scary issue title

16:36 <firefrommoonligh> Hey - oddball question. I recently reinstalled my PC; clean Windows install. I can no longer compile CMSIS-DSP using bindgen; I receive this error:... (full message at <https://catircservices.org/_matrix/media/v3/download/catircservices.org/ffpOsZKqhosLlSncjOuMOxcz>)

16:36 <firefrommoonligh> In general, for me, FFI and linker errors are at the extreme ends of difficulty and inscrutability

16:37 <firefrommoonligh> s/ends of/-/, s/and inscrutability/level/

16:38 <thejpster[m]> I would go to where you installed arm gcc and grep for those functions

16:39 <firefrommoonligh> Thank you. I don't think I did that, to my knowledge

16:41 <firefrommoonligh> My understanding is, on Linux, the install is apt install build-essential, and on Windows, it's install VisualStudioCommunity 2022, and enable Windows11SDK + v143 C/C++ build tools. And maybe LLVM, with the LIBCLANG_PATH environment variable pointing to its bin folder

16:41 <firefrommoonligh> * My understanding is, on Linux, the install is apt install build-essential, and on Windows, it's install VisualStudioCommunity 2022, and enable Windows11SDK + v143 C/C++ build tools. And maybe download/install LLVM, with the LIBCLANG_PATH environment variable pointing to its bin folder

16:47 <firefrommoonligh> Ok, solved. (Coincidence?) after comparing my tablet where it compiles to Desktop: LLVM needs to be version=15. 17 doesn't work. I don't know about 16

16:55 <firefrommoonligh> Hey so, moment to contrast the biggest thing I love about Rust (When pure rust, I guess). Great user experience. cargo build just works. C/C++/linker errors always seem to be long and inscrutable.

16:55 <firefrommoonligh> Like, why can't it just say "Incompatible LLVM version installed. Please downgrade to LLVM v15.x"

16:56 <firefrommoonligh> Adding an error about sqrt is orthogonal to the problem

16:56 <firefrommoonligh> s/Adding/Displaying/, s//`/, s//`/

16:57 <firefrommoonligh> And don't get me started in blocks and blocks of text followed by LINK ERROR 69

23:28 <JonatanMagnusson> Hi! I wonder if anyone is familiar with STM32F030 using the stm32f0xx-hal crate? I have a device with SPI, but the device is MOSI only - there's no MISO pin. There was a PR that added support for `NoMiso`, as there are in some other HAL's, but it was rejected.

23:28 <JonatanMagnusson> Any idea how to initialize SPI with MOSI only with this HAL?