Re: [PATCH 3/8] Doc/gitsubmodules: specify how submodules help in reduced size

[Kaartic Sivaraam <kaartic.sivaraam@gmail.com>]

[-- Attachment #1.1: Type: text/plain, Size: 2678 bytes --]

On Wednesday 10 January 2018 12:31 AM, Stefan Beller wrote:
> On Tue, Jan 9, 2018 at 8:01 AM, Kaartic Sivaraam
> <kaartic.sivaraam@gmail.com> wrote:
>> On Tuesday 09 January 2018 12:08 AM, Stefan Beller wrote:
>>>> diff --git a/Documentation/gitsubmodules.txt b/Documentation/gitsubmodules.txt
>>>> index cb795c6b6..3f73983d5 100644
>>>> --- a/Documentation/gitsubmodules.txt
>>>> +++ b/Documentation/gitsubmodules.txt
>>>> @@ -63,6 +63,9 @@ Submodules can be used for at least two different use cases:
>>>>      * Transfer size:
>>>>        In its current form Git requires the whole working tree present. It
>>>>        does not allow partial trees to be transferred in fetch or clone.
>>>> +      If you have your project as multiple repositories tied together as
>>>> +      submodules in a superproject, you can avoid fetching the working
>>>> +      trees of the repositories you are not interested in.
>>>
>>> You do not fetch a working tree, but a whole repository?
>>>
>>
>> Maybe I misunderstood submodules when I wrote that example. Could you
>> help out with a better and precise replacement?
> 
> If your project consists of multiple repositories tied together, some submodules
> may not be of interest for all users, who do not need to fetch such submodule
> repositories.
> 

OK, now I get why I couldn't get your point. I actually was thinking of
the version of the message I had tweaked for v2 when reading your
message. It doesn't have the confusing meaning. It actually reads,

   ...
   If the project you work on consists of multiple repositories tied
   together as submodules in a superproject, you can avoid fetching the
   working trees of the repositories you are not interested in.

So, my version takes the perspective of the person who gains the
advantage of having to clone unnecessary repos. And yours, the
perspective of the person who gives the consumer of the repo that
advantage. Both sound nice to me. But if mine doesn't sound nice to you,
let me know so that I could replace it with your version.


>> Just putting in some context as to why I did this change, I thought this
>> was the only thing that lacked an example and wanted to make it consistent.
> 
> Oh, sure I like the example; I was just worried about the wording, as a worktree
> is part of a repository, and the repository is the whole thing. In the
> current situation
> you can only fetch all-or-nothing, specifically you cannot fetch "just
> the worktree"
> (a shallow clone/fetch is the closest to that, but that still has the
> same amount of
> information the .git dir, than in the working tree)
> 

I get that!


[-- Attachment #2: OpenPGP digital signature --]
[-- Type: application/pgp-signature, Size: 870 bytes --]