The delang is using merge instead of rebase/squash

[Vladimir Panteleev via Digitalmars-d]

On Tuesday, 21 March 2017 at 17:58:06 UTC, deadalnix wrote:
> On Tuesday, 21 March 2017 at 12:45:45 UTC, Vladimir Panteleev 
> wrote:
>> On Tuesday, 21 March 2017 at 11:59:42 UTC, deadalnix wrote:
>>> It's not good either. Why would I want to look at a DAG when 
>>> the serie of event is strictly linear to begin with ?
>>
>> Not sure what you mean here. The way it's presented is not a 
>> DAG.
>
> Blue is red, up is down, and the commit graph is not a DAG.

Not sure what you mean. The commit graph is a DAG. The way you 
quoted my post made your remark seem to refer to my attempt to 
reformat it, which is not presented as a DAG.

>>> "Our source control is completely broken, but that's not a 
>>> problem because we developed 3rd party tools to work around 
>>> the brokenness"
>>
>> That's fallacious.
>>
>
> If you can't bissect, it's broken.

By that definition of "broken", all git repositories which use 
branch merging are "broken". That includes some of the biggest 
open-source projects. Frankly, if you want to stick to that 
definition, I have nothing against it.

> Listen, you know it's broken because you wrote tools to work 
> around the brokenness. If it wasn't broken you wouldn't have 
> written these tools as there would be no need to do so. So 
> let's not play pretend.

Digger would probably have existed even if D were a monorepo and 
squashed PRs' commits from the start, because it also knows how 
to satisfy each prior version's build dependencies and how to 
invoke the build scripts. Regardless, D is perfectly suitable for 
automatic bisection, which is unreasonably awkward with git 
itself - Digger makes it much easier. I think there's no shame in 
writing domain-specific tools to enhance some functionality of 
standard ones.