Re: [PATCH v5 5/7] convert: add 'working-tree-encoding' attribute

From: Junio C Hamano <gitster@pobox.com>
Date: Jan 30, 2018, 21:56 UTC
Message-ID: <xmqqfu6nrowm.fsf@gitster-ct.c.googlers.com>
In-Reply-To: <396FBDFA-606F-41D9-988C-D6886089BC15@gmail.com>

Lars Schneider <larsxschneider@gmail.com> writes:

Show 21 quoted lines

>> On 30 Jan 2018, at 21:05, Junio C Hamano <gitster@pobox.com> wrote:
>> 
>> tboegi@web.de writes:
>> 
>>> +	if ((conv_flags & CONV_WRITE_OBJECT) && !strcmp(enc->name, "SHIFT-JIS")) {
>>> +		char *re_src;
>>> +		int re_src_len;
>> 
>> I think it is a bad idea to 
>> 
>> (1) not check without CONV_WRITE_OBJECT here.
>
> The idea is to perform the roundtrip check *only* if we 
> actually write to Git. In all other cases we don't care
> if the encoding roundtrips.
>
> "git checkout" is such a case where we don't care as 
> noted by Peff here:
> https://public-inbox.org/git/20171215095838.GA3567@sigill.intra.peff.net/
>
> Do you agree?

I am not sure why this is special cased and other codepaths have "if WRITE_OBJECT then die, otherwise error" checks, so no, I do not agree with your reasoning, at least not yet.