Re: [dmidecode] [PATCH v2 3/4] dmioem: Decode HPE OEM Record 203

[Jean Delvare]

Hi Jerry,

On Tue, 12 Jan 2021 11:35:44 -0700, Jerry Hoemann wrote:
> Do you in general have a preference on the capitialization of hex numbers?

The source code I inherited from dmidecode version 1 originally used
a upper-case letters in most cases, with a few exceptions. PCI device
IDs and topology used lower-case letters. The other exceptions looked
more like inconsistencies than an educated choice.

Personally I tended to prefer upper-case letters for hexadecimal,
because they occupy the same space as digits do.

However, several industry or defacto standards went for lower-case
letters: PCI and USB device IDs and topology, but also UUIDs (per RFC
4122), MD5/SHA1, git commit IDs, etc. I just can't go against
all that by myself ;-) so going forward I'm probably going to use
lower-case letters.

For consistency and stability, I'm however going to stick to upper-case
letters in dmidecode - with the exception of PCI device IDs and
topology. The only time I changed something, it was for UUIDs to comply
with RFC 4122, and it caused more trouble than was worth.

> My prior submit was inconsistent, so I made version 2 of the patches to
> use capital X for hex formatting which seems to be more prevalent in
> other parts of dmidecode.

That's right.

> Or is PCI device like above an exception?

Yes, PCI device IDs are an exception.

-- 
Jean Delvare
SUSE L3 Support