RM: install - remove --nosave/save=False

[yarikoptic]

As it is of no effect for awhile -- I have not done investigation when it
started to be of no effect, but it was reported on the matrix forum that
it has no effect in 0.13.5-0.13.7

Since 0.14.0 seems to be not "revolutionary" enough , and option is of no effect, I think it would be ok to just kill it

[kyleam]

As it is of no effect for awhile -- I have not done investigation when it
started to be of no effect

Perhaps 21f64ea (RF+BK: Non-functional install RF, documents general approach, 2017-04-23)?

Anyway, sounds good to me. Thanks.

[kyleam]

Anyway, sounds good to me.

Though perhaps someone has scripts that call install with save=True or save=False, and they didn't notice it had no effect. Rather than crashing their scripts, it might be nice to use a dummy parameter along the lines of _NoAnnexDefault and issue a warning saying the parameter has been a noop for a long time and will be removed in the next release.

[yarikoptic]

Anyway, sounds good to me.

Though perhaps someone has scripts that call install with save=True or save=False, and they didn't notice it had no effect. Rather than crashing their scripts, it might be nice to use a dummy parameter along the lines of _NoAnnexDefault and issue a warning saying the parameter has been a noop for a long time and will be removed in the next release.

what kind of a revolution would that be? ;) I would better wait for corpses in the basement (#3496, and #3496 (comment)) to come alive, i.e. not going to bother for this unlikely to be heavily used "feature". Feel welcome to take over

[kyleam]

not going to bother for this unlikely to be heavily used "feature"

It's not about it being a feature. It's about avoiding potential script breakage when we have an easy way to.

But sure, I agree that perhaps nobody has every done install(..., save=True) without thinking, and I'm fine with your judgment call that this isn't worth worrying about.