Re: [PATCH 37/38] refs: make lock generic

From: Michael Haggerty <mhagger@alum.mit.edu>
Date: Jun 9, 2016, 15:21 UTC
Message-ID: <5759897A.7040502@alum.mit.edu>
In-Reply-To: <xmqqlh2griq3.fsf@gitster.mtv.corp.google.com>

On 06/07/2016 07:50 PM, Junio C Hamano wrote:

Show 9 quoted lines

> Michael Haggerty <mhagger@alum.mit.edu> writes:
> 
>> From: David Turner <dturner@twopensource.com>
>>
>> Instead of including a files-backend-specific struct ref_lock, change
>> the generic ref_update struct to include a void pointer that backends
>> can use for their own arbitrary data.
> 
> Hmph.

I don't know what your comment means. This step is a consequence of the design decision to stick with a single ref_transaction class that is used by all ref_stores, which was nice because it avoided the need to virtualize the functions ref_transaction_{begin,update,create,delete,verify}.

Show 18 quoted lines

>> @@ -3591,7 +3590,8 @@ static int lock_ref_for_update(struct files_ref_store *refs,
>>  		for (parent_update = update->parent_update;
>>  		     parent_update;
>>  		     parent_update = parent_update->parent_update) {
>> -			oidcpy(&parent_update->lock->old_oid, &lock->old_oid);
>> +			struct ref_lock *parent_lock = parent_update->backend_data;
>> +			oidcpy(&parent_lock->old_oid, &lock->old_oid);
>>  		}
>> ...
>> @@ -3745,7 +3745,7 @@ static int files_transaction_commit(struct ref_store *ref_store,
>>  	/* Perform updates first so live commits remain referenced */
>>  	for (i = 0; i < transaction->nr; i++) {
>>  		struct ref_update *update = updates[i];
>> -		struct ref_lock *lock = update->lock;
>> +		struct ref_lock *lock = update->backend_data;
> 
> OK, and files_* backend method downcasts it to what it wants, which
> is good.

Michael