From: Junio C Hamano <gitster@pobox.com>
Date: Thu, 11 Feb 2016 20:00:08 GMT
Subject: Re: [PATCHv9 1/6] submodule-config: keep update strategy around
Message-ID: <xmqqpow3vwzb.fsf@gitster.mtv.corp.google.com>
In-Reply-To: <xmqqfux11pm8.fsf@gitster.mtv.corp.google.com>

Junio C Hamano <gitster@pobox.com> writes:

> Stefan Beller <sbeller@google.com> writes:
>
>> +		else {
>> +			submodule->update_command = NULL;
>> +			if (!strcmp(value, "none"))
>> +				submodule->update = SM_UPDATE_NONE;
>> +			else if (!strcmp(value, "checkout"))
>> +				submodule->update = SM_UPDATE_CHECKOUT;
>> +			else if (!strcmp(value, "rebase"))
>> +				submodule->update = SM_UPDATE_REBASE;
>> +			else if (!strcmp(value, "merge"))
>> +				submodule->update = SM_UPDATE_MERGE;
>> +			else if (!skip_prefix(value, "!", &submodule->update_command))
>> +				die(_("invalid value for %s"), var);
>> +		}
>
> I think this "string to enum" parser can become a separate helper
> function in this patch, so that later patch can use it to parse the
> "--update" option in the update_clone() function.
>
> That would solve the slight inconsistency we have seen
>
> +		if ((pp->update && !strcmp(pp->update, "none")) ||
> +		    (!pp->update && sub->update == SM_UPDATE_NONE)) {
>
> in that patch.

In addition, you would eventually want the reverse translation,
i.e. SM_UPDATE_ENUM to string, when you rebase the second patch of
your sb/submodule-init series (I just tried it myself and realized
that a patch to refactor the above part would be the best place to
add such a helper).