Re: imagemmagick patch 5

[joakim]

address@hidden writes:

> Stefan Monnier <address@hidden> writes:
>
>>> The Imagemagick patch allows Emacs to use the Imagemagick libraries to
>>> load images. This support can be used in parallell with the existing
>>> image loading support. There is support for asking your imagemagick
>>> installation which image types it supports, and registering them in
>>> Emacs selectively.
>>
>> Thanks.  Looks pretty good.  Feel free to install it in the `pending'
>> branch, but please see the comments below first, and don't forget to add
>> a good NEWS entry when you install this change.
>
> Ok. I will need to read up on bzr before committing. 

I've done a local bzr branch now, called "imagemagick". I would now like
to publish this branch.

http://www.emacswiki.org/emacs/BzrForEmacsDevs
mentions ways of publishing a branch to launchpad etc, but I cant find
how to publish it on savannah. Could some kind soul enlighten me?




> Could you possibly also comment on the interface I chose for scaling and
> rotation? If they are ok, I will document them in the NEWS entry. I
> assume the "index" spec is ok since that is also used for gifs already.
>
> Also please comment if the interface to select which image types
> imagemagick gets to handle is ok.
>
>>
>>> +HAVE_IMAGEMAGICK=no
>>> +if test "${HAVE_X11}" = "yes" ; then
>>
>> Do I understand it right that this X11-only restriction could be lifted
>> at some point in the future?
>
> Yes. Theres nothing completely X specific in the patch I think, just the
> same type of bitmap manipulations that goes on for the other image
> libraries.
>
>>> diff --git a/src/dbusbind.c b/src/dbusbind.c
>>> index 7c0be49..7a47730 100644
>>> --- a/src/dbusbind.c
>>> +++ b/src/dbusbind.c
>>> @@ -773,6 +773,7 @@ xd_add_watch (watch, data)
>>>        if (fd == -1)
>>>     return FALSE;
>>>  
>>> +
>>>        /* Add the file descriptor to input_wait_mask.  */
>>>        add_keyboard_wait_descriptor (fd);
>>>      }
>>
>> Please drop this gratuitous change.
>
> Hmm, I have no idea how this crept in. Odd.
>
>
>>> +/***********************************************************************
>>> +                            imagemagick
>>> + ***********************************************************************/
>>> +#if defined (HAVE_IMAGEMAGICK)
>>> +Lisp_Object Vimagemagick_render_type;
>>> +/* Function prototypes.  */
>>> +
>>> +static int imagemagick_image_p P_ ((Lisp_Object object));
>>> +static int imagemagick_load P_ ((struct frame *f, struct image *img));
>>> +
>>> +static int imagemagick_load_image P_ ((struct frame *, struct image *,
>>> +                                       unsigned char *, unsigned int, 
>>> unsigned char *));
>>
>> Please don't use the P_ macro in new code: just use prototypes.
>
> Ok. Like this?
>
> static int imagemagick_image_p (Lisp_Object object);
> static int imagemagick_load (struct frame *f, struct image *img);
>
> static int imagemagick_load_image (struct frame *, struct image *,
>                                        unsigned char *, unsigned int, 
> unsigned char *);
>
>
>>> +//#include 
>>> "/home/joakim/current/unison/data/ImageMagick-6.5.4-7/magick/xwindow-private.h"
>>
>> This should be removed as well.
>
> Ok.
>
>>> +    //try if magicexportimage is any faster than pixelpushing
>> [...]
>>> +    //oddly, the below code doesnt seem to work:
>>
>> We currently only use /*...*/ comments, so I'd rather we don't introduce
>> // comments for now.
>>
>
> Sorry, I keep forgetting this, it seems to be in my muscle memory. 
>
>>         Stefan
>>
-- 
Joakim Verona