Remove state-saving by default

[cpfiffer]

Hotfix for #1091 and reports here. This PR makes it so that state is not automatically saved in a Chains object when sampling is completed -- people who want to resume the chain will have to explicitly plan ahead of time to resume sampling:

chain = sample(model, spl, N; save_state = true)

[mohamed82008]

Please change prob"x = 1.0, y = 1.0 | chain = chain" to prob"x = 1.0, y = 1.0 | chain = chain, model = gdemo" in the docs, the guide.md file.

[mohamed82008]

Or add both saying a word about the save_state kwarg.

[yebai]

@cpfiffer can you take a look at Mohamed's comment, and see whether we can fix them before merging?

[yebai]

Hi @yebai, I just got a benchmark command from you,
But I can't determine on which branches to run,
Please specify the branches in your command.

@KDr2 Maybe reduce unnecessary replies (e.g. only reply to comments containing @BayesBot benchmark() from the Benchmark bot?

[cpfiffer]

This should be good to merge once the tests pass.

[KDr2]

Hi @yebai, I just got a benchmark command from you,
But I can't determine on which branches to run,
Please specify the branches in your command.

@KDr2 Maybe reduce unnecessary replies (e.g. only reply to comments containing @BayesBot benchmark() from the Benchmark bot?

OK :D