Re: [PATCHv10 01/10] receive-pack.c: shorten the execute_commands loop over all commands

From: Stefan Beller <sbeller@google.com>
Date: Jan 5, 2015, 21:07 UTC
Message-ID: <CAGZ79kbRLPYRw+iifigRHqJ5Lc1brQ3qkUV=4YYPSwr72+giPg@mail.gmail.com>
In-Reply-To: <20150105202244.GL29365@google.com>

On Mon, Jan 5, 2015 at 12:22 PM, Jonathan Nieder <jrnieder@gmail.com> wrote:

Show 15 quoted lines

> Stefan Beller wrote:
>
>> --- a/builtin/receive-pack.c
>> +++ b/builtin/receive-pack.c
> [...]
>> @@ -1077,27 +1100,15 @@ static void execute_commands(struct command *commands,
> [...]
>> +     if (shallow_update)
>> +             assure_connectivity_checked(commands, si);
>
> Looking at this code alone, it seems like assure_connectivity_checked()
> is going to ensure that connectivity was checked, so that I can assume
> connectivity going forward.  But the opposite is true --- it is a
> safety check that prints a warning and doesn't affect what I can
> assume.

I disagree on that. Combined with the next patch (s/error/die/) we can assume that the the connectivity is there as if it is not, git is dead.

This is why I choose the word assure. Maybe check_assumption would be better?

Show 6 quoted lines

>
> The factored-out function fails in what it is meant to do, which is to
> save the reader of execute_commands from having to look at the
> implementation of the parts they are not interested in.
>
> Would something like warn_if_skipped_connectivity_check() make sense?

The next patch would then change this to die_if_... ? I'd be ok with that, but in your original email you would still have the last die(...) in the execute_command function which I dislike. So what about:

if (shallow_update)
       (warn|die)_on_skipped_connectivity_check()

>
> Jonathan