Re: [PATCH v3 2/2] huawei: add gprs context

[Inaky Perez-Gonzalez <inaky@linux.intel.com>]

[-- Attachment #1: Type: text/plain, Size: 2285 bytes --]

On Fri, 2010-08-06 at 10:26 +0300, Kalle Valo wrote: 
> Inaky Perez-Gonzalez <inaky@linux.intel.com> writes:
> 
> > Hi Kale
> 
> Hola,
> 
> > On Thu, 2010-08-05 at 08:54 +0300, Kalle Valo wrote: 
> >>
> >> This is my setup:
> >> 
> >> ; from linux/Documentation/CodingStyle
> >> (defun c-lineup-arglist-tabs-only (ignored)
> >>   "Line up argument lists by tabs, not spaces"
> >>   (let* ((anchor (c-langelem-pos c-syntactic-element))
> >> 	 (column (c-langelem-2nd-pos c-syntactic-element))
> >> 	 (offset (- (1+ column) anchor))
> >> 	 (steps (floor offset c-basic-offset)))
> >>     (* (max steps 1)
> >>        c-basic-offset)))
> >
> > Thanks! Phew ... this is a handful of lisp! I don't understand anything
> > that is not basic anymore :)
> 
> It sure is. I just copied it from Linux CodingStyle document and it
> seems to work somehow.
> 
> > -- I gave it a whiz and one thing I saw was that
> >
> > struct tx_queue_entry *__ofono_sms_txq_submit(struct ofono_sms *sms,
> > 						GSList *list,
> > 						unsigned int flags, unsigned msg_id,
> > 						ofono_sms_msg_stch_cb_t stch_cb,
> > 						void *data, ofono_destroy_func destroy)
> >
> > is being formatted as:
> >
> > struct tx_queue_entry *__ofono_sms_txq_submit(struct ofono_sms *sms,
> > 					GSList *list,
> > 					unsigned int flags, unsigned msg_id,
> > 					ofono_sms_msg_stch_cb_t stch_cb,
> > 					void *data, ofono_destroy_func destroy)
> >
> > ie: one c-basic-offset less than it should -- I really can't figure out where to poke.
> 
> That's my problem as well. Currently I just manually add one more tab to
> each line with M-i. If you find a way to fix this, please let me know.

I think I found it, seems to cut it; replace:

(steps (floor offset c-basic-offset)))

with

         (steps (floor (+ offset (- c-basic-offset 1)) c-basic-offset)))

Lisp nuts me out, but it is doing the:

steps = floor(offset/c-basic-offset);

with

steps = floor((offset + c-basic-offset - 1) / c-basic-offset);

thus making sure there is always one more tab except when the offset is
right on a tab boundary that we keep the same.

So far, so good -- probably it fails in some case, because this is
totally trial and error as unscientific as it can :)