On Mon, Aug 4, 2008 at 3:51 PM, Peter TB Brett <[EMAIL PROTECTED]> wrote:
> On Mon, 4 Aug 2008 15:20:39 +0200, Patrick Bernaud <[EMAIL PROTECTED]>
> wrote:
>> Bernd Jendrissek writes:
>>  > [...]
>>  > Press ':', then in command box, type "(file-quit)"  :)
>>
>> Yes, I figured that. BTW the 'Shift :' is impossible to realize on my
>> keyboard. This is a problem that is likely to happen on other
>> combinations but it is the first time I am hit by that.
>
> What about binding it to just ';'?

FWIW I had some trouble getting the key bindings to work properly; it
seemed a bit odd to bind to "shift" + colon when the colon itself is
already shift+semicolon on my keyboard!  I don't know what the
resolution is.

I also notice now that there's already DEFINE_G_KEYS(script_console) and a stub
DEFINE_I_CALLBACK(script_console) that seems like a better name than
"invoke macro".

>>  > Somewhere in the bowels of my slotting mechanism there are changes to
>> enable:
>>  >
>>  > (add-component "opamp.sym")
>>  >
>>  > which initializes the component selector's text entry to that value.
>>  > With a bit more work it can close the dialog when exactly one
>>  > component from the library matches the input.  I realise that doesn't
>>  > answer your question wrt master!
>>
>> Good but avoid adding guile code to i_callback_*(): keep the
>> separation between guile code and callbacks.
>>
>> So in your example the rest arguments "opamp.sym" has to be analyzed
>> by g_keys_add_component() (the function is customized and not provided
>> by the DEFINE_G_KEY macro) that pass it to i_callback_add_component()
>> as a C string (again its prototype is customized).
>
> From what you're saying it sounds like you think I should revert commit
> 61255c9b68aeae2ae845eea82babf445036a72f1. Is that correct?

I think only part of it: if I understood Patrick correctly, only the
g_keys_* functions should have a "rest" argument, which they then
thunk into C-friendly arguments in i_callbacks_*.  Ah, now I see
Patrick has replied too.  'nuff said.

It'll need a way to propagate "unspecified" into C; NULL seems
appropriate enough for char * arguments.  I just hope there won't be
too many optional integer arguments.

I'll fix this sometime this week unless someone beats me to it.


_______________________________________________
geda-dev mailing list
[email protected]
http://www.seul.org/cgi-bin/mailman/listinfo/geda-dev

Reply via email to