Re: [RFC/PATCH] builtin/blame: darken redundant line information

[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

 



On Tue, Jun 13, 2017 at 10:33 AM, Junio C Hamano <gitster@xxxxxxxxx> wrote:
> Stefan Beller <sbeller@xxxxxxxxxx> writes:
>
>> But you do not want to (yet)? The goal is not to tell you where the bounds
>> are, but the goal is to point out that extra care is required for review of
>> these particular 3 lines.
>
> And when you _can_ help users in that "extra care" by pointing out
> where the boundary is, what is the justification for hiding that
> information?

It is very complicated and confusing. Consider this:

>          context
>         -C
>          context
>         -B
>         -B
>         -B
>         -A
>         -A
>         -A
>          context
>         +A
>         +A
>         +A
>         +C
>         +B
>         +B
>         +B
>          context

So from your emails I understood you want to markup
block starts and ends, but in this case C is *both* start
and end of a block, and has also different blocks around.

So ideally we could tell the user this

>          context
>         _context C goes to after +A
>         -C
>         _context C goes to before +B
>          context
>         _context -B goes to after +C
>         -B
>         -B
>         -B
>         _context -B goes to before contextB
>         _context -A goes to after contextA
>         -A
>         -A
>         -A
>         _context -A goes to after contextA
>          context
>         _context +A comes from after -B
>         +A
>         +A
>         +A
>         _context +A comes from before contextA
>         _context +C comes from after contextC
>         +C
>         _context +C comes from before contextC
>         _context +B comes from after contextB
>         +B
>         +B
>         +B
>         _context +B comes from after -A
>          context

(show the context of where the move is coming from/going to,
maybe just one or two lines)

And how many colors would be confusing for the user?

I would think we want to start with a simple model and if a
niche is not good (read: people think it can be improved easily
compared to the usefulness they get out of it) enough we fix
it up later.

I thought that adding 4 new colors is already maybe too much,
as Git users were happy with a handful for 10 years, so I am very
opposed to add more than 4 colors unless there is a very good
reason. And I'd think that this is not adding a lot of useful information
for a reviewer?

Thanks,
Stefan



[Index of Archives]     [Linux Kernel Development]     [Gcc Help]     [IETF Annouce]     [DCCP]     [Netdev]     [Networking]     [Security]     [V4L]     [Bugtraq]     [Yosemite]     [MIPS Linux]     [ARM Linux]     [Linux Security]     [Linux RAID]     [Linux SCSI]     [Fedora Users]