web interfaces vs "git blame"

[Jan Beulich]

On 05.09.2026 21:10, Mark Wielaard wrote:
> On Thu, 2026-09-03 at 23:41 -0700, Collin Funk wrote:
>> Jan Beulich <jbeulich@suse.com> writes:
>>
>>> gitweb (https://sourceware.org/git/gitweb.cgi?p=binutils-gdb.git) had
>>> its "blame" links removed long ago, yet now cgit
>>> (https://sourceware.org/cgit/binutils-gdb) followed this unhelpful
>>> behavior. Can this please either be restored, or another easy to use
>>> (not involving typing in git commands by hand) mechanism be pointed
>>> out?
>>
>> I wouldn't be surprised if 'git blame' was expensive to render for large
>> files with lots of history. Perhaps it got turned off to avoid bots
>> using it.
> 
> Yes, that was the reason. The recent AI scraperbot attack was the last
> straw :{ Really sorry we had to disable that functionality.
> 
> git annotate locally of course gives the same information.
> And as Collin pointed out the forge mirror does have that functionality
> (we are lucky the bots apparently haven't found the forge yet).

It's only a matter of time, I think. Really all these anti-AI measures
(everywhere, not just for sourceware) look extremely bogus to me. Leaving
aside that such shouldn't be necessary in the first place if those
controlling the various models were acting responsibly, whatever is done
should never affect what humans might sensibly do. Yet of course - I know
I'm dreaming.

Jan