[MartinFuzzey]

Currently the partitions used by snagflash DFU have to be the numeric altsettings, let's allow the names defined in u-boot to be used too.

Eg if u-boot is configured with
dfu_alt_info "mmc 0=fsbl raw 0 1000 mmcpart 1;ssbl part 0 1"
writing to the ssbl currently requires the snagflash option
-D 1:binaries/u-boot.img
with this patch the following is also accepted
-D ssbl:u-boot.img

Note: if the name starts with a number it will still be interpreted as a numeric alt setting.

[rgantois]

Thanks a lot! I've added a few comments.

[rgantois]

stray whitespace

[MartinFuzzey]

Hi Romain, Thanks for the quick review

On Fri, 2 Oct 2026 at 11:56, Romain Gantois ***@***.***> wrote: @rgantois requested changes on this pull request. Thanks a lot! I've added a few comments. ________________________________ In docs/snagflash.md: > @@ -166,7 +166,7 @@ snagflash -P ums -s binaries/u-boot.stm32 -d /mnt/u-boot.stm32 ``` **Note:** If you want a static block device path, you can use the following udev -rules to create symlinks when a certain VID:PID pair is detected: +rules to create symlinks when a certain VID:PID pair is detected: stray whitespace ________________________________

It's the reverse. There was previously stray whitespace and my patch removed it (because of my editor settings to automatically do that). Is that a problem?

Regarding this statement in the commit log: Note: if the name starts with a number it will still be interpreted as a numeric alt setting. This doesn't seem to be the case? int(partname) will throw a ValueError if partname has letters in it. And search_partid() will not match on the numerical altsetting. Nevertheless, I'm okay with how the code currently works as I don't want incorrect partname arguments to be interpreted in a lossy way, but I'm just confused about the commit log note.

Yes you are right of course. I must of been thinking in C with atoi() not python when I wrote that comment.

[rgantois]

Hi Romain, Thanks for the quick review
On Fri, 2 Oct 2026 at 11:56, Romain Gantois @.***> wrote: @rgantois requested changes on this pull request. Thanks a lot! I've added a few comments. ________________________________ In docs/snagflash.md: > @@ -166,7 +166,7 @@ snagflash -P ums -s binaries/u-boot.stm32 -d /mnt/u-boot.stm32 ``` Note: If you want a static block device path, you can use the following udev -rules to create symlinks when a certain VID:PID pair is detected: +rules to create symlinks when a certain VID:PID pair is detected: stray whitespace ________________________________
It's the reverse. There was previously stray whitespace and my patch removed it (because of my editor settings to automatically do that). Is that a problem?

Ah yes indeed, then no it's not a problem!

[rgantois]

Thanks!