On 25 February 2016 at 09:26, Colomban Wendling <[email protected]>
wrote:

> An API break is questionable, but a "nice to have" change like this PR is
> not worth an ABI break.
>
> How would this break ABI? It's a typedef rename, not a type change. The
> "nice to have" part is including TM_PARSER_* enum values.
>

​One option was to move it into a GEANY_PRIVATE struct, that would change
the struct and so break the ABI.​



> […] Which one do you prefer?
>
> Hum… I prefer the new type name. I prefer full compatibility. I prefer
> stuff to be defined where it makes sense.
>
> So, I guess what I'd prefer is
>
>    - use the new type name
>    - introduce a deprecated alias of the old name
>    - install *tm_parser.h* (possibly guarding everything but the typedef
>    in GEANY_PRIVATE)
>
> —
> Reply to this email directly or view it on GitHub
> <https://github.com/geany/geany/pull/906#issuecomment-188513167>.
>


---
Reply to this email directly or view it on GitHub:
https://github.com/geany/geany/pull/906#issuecomment-188552606

Reply via email to