[PATCH] D24167: Moving to GitHub - Unified Proposal

Duncan P. N. Exon Smith via llvm-commits llvm-commits at lists.llvm.org
Wed Oct 12 15:42:14 PDT 2016


> On 2016-Oct-12, at 15:24, Mehdi Amini <mehdi.amini at apple.com> wrote:
> 
>> 
>> On Oct 12, 2016, at 3:11 PM, Duncan P. N. Exon Smith <dexonsmith at apple.com> wrote:
>> 
>> 
>>> On 2016-Oct-12, at 14:54, Mehdi Amini <mehdi.amini at apple.com> wrote:
>>> 
>>>>> + * Refactoring across projects is not frienly: taking some functions from clang
>>>> 
>>>> s/frienly/friendly/
>>>> 
>>>>> +   to make it part of a utility in libSupport wouldn't carry the history of the
>>>>> +   code in the llvm repo, preventing recursively applying `git blame` for
>>>>> +   instance.
>>>> 
>>>> Please add: "However, this is no different than the current state.ā€
>>> 
>>> I can compromise with "However, this is not very different than the current state.ā€
>>> 
>>> Although today I can move code across subproject in a single commit and git blame works in the monorepo export.
>>> 
>> 
>> I'm convinced you could set up a monorepo mirror even if the canonical source uses multirepo.
> 
> Sure, but aggregating the individual repo does not bring the possibility to have a single commit in the aggregation repo that moves code across the original sub-repo, while now you can (the canonical repository supports it). 
> 
> Iā€™m skeptical that tooling can really help in practice with this (outside of limited trivial cases).
> 
>> The current state is: multiple subproject refactorings are not friendly (require heavy tooling).  Adopting multirepo doesn't change that much.  It will still be unfriendly, but require heavy tooling.
>> 
>> How about: "However, this is similar to the current state."?

Let me repeat:

How about: "However, this is similar to the current state."?



More information about the llvm-commits mailing list