Re: Preliminary results for proposed new pgindent implementation

Hi,

On 2017-06-16 13:10:50 -0400, Tom Lane wrote:
> I experimented with disabling that logic and just always aligning
> to the paren indentation. That fixes the weird cases with continued
> string literals, but it also makes for a heck of a lot of other changes.
> The full diff is too big to post here, but I've attached a selection
> of diff hunks to give you an idea. I'm not really sure if I like this
> better than pgindent's traditional behavior --- but it's arguably less
> confusing.
>
> An intermediate position that we could consider is to disable the back-off
> logic only when the line starts with a string literal. I haven't actually
> coded this but it looks like it would be easy, if grotty.

I think the current logic is pretty horrible, primarily because it's so
hard to get to manually. I could live with both of these proposed
changes, the selection of the changes you posted looks like it could be
improved by code changes, but that's obviously a large amount of work.
The heuristic also seems to make sense.

At this point however I wonder whether just moving to the new tool on
its own wouldn't be a big enough change - we could just delay that
decision until we've got the rest done at least.

- Andres