From: Junio C Hamano <gitster@pobox.com>
Date: Tue, 03 Jul 2012 17:19:14 GMT
Subject: Re: [PATCH v4 03/18] sha1_name.c: get rid of ugly get_sha1_with_mode_1()
Message-ID: <7vk3ykwynh.fsf@alter.siamese.dyndns.org>
In-Reply-To: <vpqsjd9nuid.fsf@bauges.imag.fr>

Matthieu Moy <Matthieu.Moy@grenoble-inp.fr> writes:

> Junio C Hamano <gitster@pobox.com> writes:
>
>> +/*
>> + * Call this function when you know "name" given by the end user must
>> + * name an object but it doesn't; the function _may_ die with a better
>> + * diagnostic message than "no such object 'name'", e.g. "Path 'doc' does not
>> + * exist in 'HEAD'" when given "HEAD:doc", or it may return in which case
>> + * you have a chance to diagnose the error further.
>> + */
>> +void die_on_misspelt_object_name(const char *name, const char *prefix)
>
> It seems unusual to have a function named die_* that is not a noreturn
> function. I'd call it die_*_maybe, or diagnose_* instead.
>
> (but as the comment above documents the behavior, it's not terribly
> important, you may ignore my comment if you whish)

I was hoping "on" may imply "if not misspelled, ignore and keep
going", but apparently that failed.  I am not good at names.

die_if_misspelt?  maybe_die_on_misspelt?

I am fairly negative on "diagnose" as it does not say much about
what would happen after diagnosis (namely, we _die_).