wrong guess for py-indent-offset due to backslash continuation

[Georg Brandl]

The indent offset shouldn't be guessed from backslash continuations. I have

from long.pkg.name import long, list, of, \
     class, and, function, names

(note there are five spaces before "class", to match with the start of the pkg name.)

Since the indent of backlash-continued lines has no meaning for code, it should not be considered.

[Andreas Roehler (a-roehler)]

Hi Georg,

need a rule for that.

- a special imports rule?
or
- indent all continuation lines after first whitespace?

BTW we have several indent rules meanwhile, which don't effect execution.

Thanks,

Andreas

[Georg Brandl (gbrandl)]

I wouldn't mind if python-mode didn't indent to column 5 on these statements (although the old python-mode did it).

But it is in any case wrong to guess the indent-offset from this, that's why I opened the bug.

[Andreas Roehler (a-roehler)]

Am 16.09.2011 22:26, schrieb Georg Brandl:
> I wouldn't mind if python-mode didn't indent to column 5 on these
> statements (although the old python-mode did it).
>
> But it is in any case wrong to guess the indent-offset from this, that's
> why I opened the bug.
>

OK, see, it's a little bit tricky.

Understand it that way: if the previous behavior is restored --which
shall be done-- indent guessing will fail.

So two issues to cure.

Andreas

[Andreas Roehler (a-roehler)]

Hi Georg,

could you check again?

thanks,

Andreas

[Georg Brandl (gbrandl)]

Hi Andreas,

offset guessing works now, but the indendation of the second "import" line is sometimes
computed to be sometimes 3, then pressing tab again, 5, spaces.

But sometimes it's changing between 0 and 8 spaces.

(Another comment: my example sadly is invalid syntax, since you cannot import "class":
it's a keyword.)

[Andreas Roehler (a-roehler)]

Am 18.09.2011 07:52, schrieb Georg Brandl:
> Hi Andreas,
>
> offset guessing works now, but the indendation of the second "import" line is sometimes
> computed to be sometimes 3, then pressing tab again, 5, spaces.
>
> But sometimes it's changing between 0 and 8 spaces.
>

could you make a new tracker entry for this?

> (Another comment: my example sadly is invalid syntax, since you cannot import "class":
> it's a keyword.)
>

think it's pseudo-code obviously - saying
  class-and-function-names

Nonetheless, will changes that in test taking the example

thanks,

Andreas