cc @nhmc @aconley @astrofrog @joergdietrich

Oh, and one point I forgot to mention above: The idea is that we'd deprecate in 0.4, and then wait and see if we get any complaints (might require some actively legwork on our part to actually ask people instead of just seeing if they raise an issue).

Sounds like a good plan to me.

Sounds good to me too.

Alright, @aconley @nhmc, this is now a PR that deprecates the functional interface. I also tried to update the docs appropriately, but some cases may have slipped through.

If we do go ahead with removing them in a future version, some of the tests will have to be re-written, but I figured I'd leave them in so that we're testing until they actually go away.

Thanks for the PR!

The code changes look fine, but I think the documentation will need a bigger overhaul. The 'Getting Started' section should show examples using the built-in cosmologies, and the description of get_current()/set_current() can be moved right to the end of the docs. The nice list of functions at the bottom will also go away once the functional interface is removed, so we should work out a way to list the most commonly used methods in a similar way.

I'm happy to open a separate PR for the documentation changes if you like.

Yeah, I see your point on the docs, @nhmc . If you could take care of it, that would be great (particularly since you have a good idea). Do you want to open it against my branch here, or would you rather wait for the merge and get it from master?

(Also, there was a typo in the "whatsnew" section that I just fixed)

I'll wait for the merge.

@eteq - this will need rebasing following the merging of #2094

Alright, rebased. If the tests pass I'll merge it tomorrow. I also had to make some doc changes to mesh with @2094 and it's change to using default_cosmology.get/set. @nhmc, this may be part of what you had in mind, but of course cleaning it up more is a good idea!

👍

@eteq - I think there is a minor rebase needed, due to #2103 - I just realized, can you remove the function adding in #2103 from the funcs.py file? (since there's no need to deprecate it).

@astrofrog - done. I also realized that the age and comoving_volume convinience functions were not in 0.3, so I removed them as well (previously they were deprecated). If the tests pass, I will merge.

Looks like there was another conflict just merged... Rebasing again.

👍

Alright, merging! @nhmc, you can do the doc changes against master at will.

Great, thanks!

On 8 May 2014 03:32, Erik Tollerud [email protected] wrote:

Merged #2343 #2343.

—
Reply to this email directly or view it on GitHubhttps://github.com//pull/2343#event-118785548
.

Now that it's too late...

Should we warn users somewhere to update their astropy.config information? If you have an old one, every time you import astropy.cosmology you will get a warning about the default_cosmology parameter being deprecated. Easy to fix -- if you happen to know about the astropy config system, which I bet many users don't.

@aconley - I think the warning should only appear if you modified the configuration file, so users that are not familiar with the config file will probably not see the warning. Did you edit your config file?

I have no idea, but it's certainly possible.

As @astrofrog says, by design that warning should only appear for people who edited their config file.