Re: extending the blame callback

Branko Čibej wrote:
> On 14.01.2019 13:36, Julian Foad wrote:
[...]

Thanks, Brane; you gave a good initial summary of the scope of the problem. As you said on IRC there's even more to it, such as the relationship between line splitting and keywords, meaning that all has to be addressed together.

> > Can we re-write this properly?
>
> As I said, this is not limited to blame. The current change in the blame
> API at least makes some sense and is a localised change, that's why I
> support it. It doesn't break anything that wasn't broken before.
>
> Changing the way we handle text-like files for diff, blame and patch, on
> the other hand, is quite a bit more involved and I'm afraid it'll touch
> a lot of code. I wouldn't dream of rejecting svn_client_blame6 just
> because it doesn't solve the larger problem.

Ack. I didn't mean to reject this current upgrade: I agree that providing an svn_string_t is a useful and self-contained upgrade in itself, and I support it. I meant can we next, sometime, address the bigger problem that you discussed above. (We should open an issue in the tracker, and start a new thread for it.)

-- 
- Julian