Lennart Spitzner

> I'm cautiously hopeful that this may, however, be the spirit of the idea for IndentPolicyMultiple? Correct. (And unfortunately for you, it is indeed not implemented to any degree.) An...

See 1c0f8b6811dde8716b2b3842f610c92a2831af89 on branch [indent-policy-multiple](https://github.com/lspitzner/brittany/tree/indent-policy-multiple). This indeed produces (with inline-config we can now write this self-contained) ~~~~.hs -- brittany { lconfig_indentAmount: 4, lconfig_indentPolicy: IndentPolicyLeft } -- brittany { lconfig_columnAlignMode: {...

> Given how IndentPolicyLeft has sort of morphed to be one component of "context independent layout", it seems that the semantics of IndentPolicyMultiple are somewhat orthogonal. Would it make sense...

> Just writing this i notice that this code breaks for no wait, that was stupidity on my part. Pushed another commit to the branch. It seems to work even...

I guess it makes sense to separate the imports-layouting mode from the indentation-policy. Bet this gives rise to some nice design questions of how flags imply other flags .. but...

Ah, yeah i guess for `where` a different BriDoc constructor creates the indentation. I think this can fixed by just disabling the "where special" flag whatever the exact name was....

Great! the remaining work is of course to actually connect the `Multiple` flag instead of overwriting the `Left` behaviour. I fear this will touch some more code (grepping for `IndentPolicyLeft`...

@chreekat my bad, tests were effectively disabled since one commit on master. This is fixed now.

@chreekat ah, good to hear! i was just looking into this as well, because brittany is blocking the ghc-8.4 support for HIE so i want to make a release really...

Right, I will do the rebasing then, one sec.