Re: [PATCH] push: add config option to --force-with-lease by default.

["Ævar Arnfjörð Bjarmason" <avarab@gmail.com>]

On Wed, Jul 05 2017, Junio C. Hamano jotted:

> On Tue, Jul 4, 2017 at 11:34 PM, Francesco Mazzoli <f@mazzo.li> wrote:
>>
>> Could you clarify the danger you're referring to? E.g. give an example
>> of surprising --force-with-lease behavior that we do not want to
>> encourage?
>
> https://public-inbox.org/git/1491617750.2149.10.camel@mattmccutchen.net/

In the context of this patch I don't understand why you're concerned
that making --force mean --force-with-lease makes things worse.

See my
https://public-inbox.org/git/CACBZZX48RanjHsv1UsnxkbxRtqKRGgMcgmtVqQmR84H5j8awqQ@mail.gmail.com/
(follow-up to the E-Mail you posted):

    To me the *main* feature of --force-with-lease is that it's less
    shitty than --force, without imposing too much UI overhead. We have to
    be really careful not to make --force-with-lease so complex by default
    that people just give u and go back to using --force, which would be
    worse than either whatever current problems there are with the
    current --force-with-lease behavior, or anything we replace it with.

I.e. yes there are workflows with some background auto-update that will
make it less safe, which I documented in f17d642d3b ("push: document &
test --force-with-lease with multiple remotes", 2017-04-19).

But it is still the case that --force-with-lease is categorically a more
safer option than simply --force, which has none of the safety
--force-with-lease has. It would still wipe away history in this
scenario you're pointing out *and others*.

Surely the point of having an option like this is to have a net
reduction in complexity.

I think it can be argued that it's bad UI design though to have --force
mean different things depending on the config, and we'd be better off
with a patch that disables --force.