On Saturday, November 4, 2023 at 11:49:41 PM UTC-4 Carl Gay wrote:
Hi. I'm implementing <https://github.com/cgay/protocol-buffers> protocol
buffers for a case-insensitive language (Dylan) and I'm wondering if
there's prior art that would give me an idea how to handle the following
case:
message M {
string S = 1;
oneof O {
string s = 2;
}
}
Since fields M.S and M.s are siblings, the names to access those fields
will collide. I can generate the name "s*" for lowercase "s". Since "*" is
invalid in proto field names that's future proof.
So I think I have a way forward but I'm curious if anyone knows of other
languages that have already addressed this, and what they did.
The Common Lisp implementation ignores the problem and relies on it not
happening. The docs for the PHP implementation don't mention this
particular problem. (They do mention reserved names such as Empty being
renamed to PBEmpty, but that's not a problem for Dylan.)
Thoughts? Prayers?
Thanks.
-Carl
A couple more thoughts:
Auto-assigning a different name for the field, such as "s*", would have to
be resilient to movement of the fields textually within the file. If the
oneof were moved textually above field S the name "s*" could suddenly name
a different field.
Another possibility: a custom field option: dylan_name = "s*". This assumes
access to modify the .proto file, or the need to maintain a copy of it.
Maintain a custom name mapping file to pass to the proto parser / code
generator.
(This is largely an academic exercise. This scenario seems unlikely to
happen in practice, especially considering the number of Dylan users on the
planet.)
--
You received this message because you are subscribed to the Google Groups
"Protocol Buffers" group.
To unsubscribe from this group and stop receiving emails from it, send an email
to [email protected].
To view this discussion on the web visit
https://groups.google.com/d/msgid/protobuf/eb524684-d548-4b95-8411-77f07a9b28f9n%40googlegroups.com.
