Subject:
|
Re: tweaking around dlcd_show( )
|
Newsgroups:
|
lugnet.robotics.rcx.legos
|
Date:
|
Mon, 27 Aug 2001 02:37:11 GMT
|
Viewed:
|
1753 times
|
| |
|
|
In lugnet.robotics.rcx.legos, Ross Crawford writes:
> In lugnet.robotics.rcx.legos, Martin Scragg writes:
> > In lugnet.robotics.rcx.legos, Alexandre Bec writes:
> > > I tried to use the "real" value 0xef46,0x4 instead of the define LCD_4_BOTL
> > > dlcd_show(0xef46,0x4); and I got obviously the same error.
> >
> > This should work, unless the preprocesor has done something funny to it, or
> > is expected to do something funny to it.
>
> Nope - bit_set(0xef46,0x4) will work, but not dlcd_show(0xef46,0x4), as it
> needs only 1 parameter.
Since writing this, I have set up a LegOs build environment, so had a play
with it. It doesn't seem very useful the way it is written. See my bit
about it below.
> > > When trying at least to use directly the bit_set function with
> > > bit_set (0xef46,0x4); it works,
> > > but if I put these items in an array and invoking bit_set with
> > > bit_set(progressPattern[ i ][0], progressPattern[ i ][1] );
> > > I obtain a
> > > "inconsistent operand constraints in an `asm' " because my array certainly
> > > does not contain the right type of data for such a low-level operation. (how
> > > can we fit a single bit in an array ?)
> > You would need to define the elements as a structure, and then have an array
> > of that structure.
>
> Choosing the types of the struct elements is the problem - these are passed to
> the h8's bset instruction. First, I tried void* & char, but the compiler didn't
> like it. Then I tried everything else, and nothing worked. When I get a chance
> I'll try & find a h8 guru & see if we can work it out. It may be the assembler
> directives (after the ":") but I don't know enough about those.
I had a bit of a go doing this as well, and since I didn't know what the asm
statement was wanting as paramters, had a hard job convincing it to work.
In reality, I would rewrite the macros, as they seem a little "dodgy" to me.
I think that 'dlcd_show' should take the 2 parameters, that way you can use
it with variables. The way it is now, it can only really be used with the
defines. Also when I normally write macros, I enclose the parameters passed
in parentheses. e.g.
do_stuff(a, b) ((a) + (b))
This way you don't end up with problems in the macro if you pass it values
like this:
do_stuff(5, 7 + 9)
Without the parenteses, this would totaly break the above macro.
Martin
|
|
Message has 1 Reply:
 | | Re: tweaking around dlcd_show( )
|
| (...) LCD_4_BOTL (...) I agree it's a bit bodgy. The one advantage is you only need to pass a single #define when calling it directly (OK not a huge advantage, but if you look at the values in the #defines, you'll see why it was done that way). But (...) (23 years ago, 27-Aug-01, to lugnet.robotics.rcx.legos)
|
Message is in Reply To:
| | Re: tweaking around dlcd_show( )
|
| (...) Nope - bit_set(0xef46,0x4) will work, but not dlcd_show(0xef46,0x4), as it needs only 1 parameter. (...) Choosing the types of the struct elements is the problem - these are passed to the h8's bset instruction. First, I tried void* & char, but (...) (23 years ago, 27-Aug-01, to lugnet.robotics.rcx.legos)
|
10 Messages in This Thread:
  
- Entire Thread on One Page:
- Nested:
All | Brief | Compact | Dots
Linear:
All | Brief | Compact
This Message and its Replies on One Page:
- Nested:
All | Brief | Compact | Dots
Linear:
All | Brief | Compact
|
|
|
|
