Re: [PATCH v6 18/27] files-backend: replace submodule_allowed check in files_downcast()

From: Michael Haggerty <mhagger@alum.mit.edu>
Date: Mar 29, 2017, 10:55 UTC
Message-ID: <a44dfc2e-8196-a782-be69-682f252d2592@alum.mit.edu>
In-Reply-To: <CACsJy8CpUzRjZG7V8a6XvzcjxhwTnAfpwYed3OSLrygUPOK66g@mail.gmail.com>

On 03/26/2017 04:16 AM, Duy Nguyen wrote:

Show 20 quoted lines

> On Mon, Mar 20, 2017 at 4:18 AM, Michael Haggerty <mhagger@alum.mit.edu> wrote:
>>> +/* ref_store_init flags */
>>> +#define REF_STORE_READ               (1 << 0)
>>
>> I asked [1] in reply to v5 whether `REF_STORE_READ` is really necessary
>> but I don't think you replied. Surely a reference store that we can't
>> read would be useless? Can't we just assume that any `ref_store` is
>> readable and drop this constant?
> 
> I deleted it then I realized many lines like these
> 
>     files_downcast(ref_store, REF_STORE_READ, "read_raw_ref");
> 
> would become
> 
>     files_downcast(ref_store, 0, "read_raw_ref");
> 
> I think for newcomers, the former gives a better hint what the second
> argument is than a plain zero in the latter, so I'm inclined to keep
> it.

That's OK with me; I don't feel strongly about it.

Michael