#define history

bkamen · Posted: Sat Feb 06, 2010 3:56 pm

When I look at CCS code vs. MCHP code, I see they use a lot of

#define SOMETHING (10u)

It was easy for me to determine what that means -- but I've looked around for the history of that syntax and was googled to death with non-answers.

I don't see it used at all in CCS examples and was wondering why.

Anyone wanna toss in their $0.02? I'd love to hear it.

-Ben
_________________
Dazed and confused? I don't think so. Just "plain lost" will do. :D

mbradley · Joined: 11 Jul 2009 Posts: 118 Location: California, USA

mah 2 cents....make that 1...

I am a little confused, are you referring to #define in general? it is used alot in CCS C. pretty much all that is in a chip def file, such as 18f2525.h

We use stuff like, #define LED PIN_C1, and in the some_chip.h file, PIN_C1 is defined something like #define PIN_C1 0x1200

CCS has done a wonderful job of putting all the defines in the .h files for us, as related to general chip definitions.

I may be lost, is this what you were referring to?
_________________
Michael Bradley
www.mculabs.com
Open Drivers and Projects

bkamen · Posted: Sat Feb 06, 2010 5:55 pm

my bad...

I'm talking about the (10u) where it's shorthand for 10unsigned

or something like 10ul (unsigned long) -- if I'm reading this correctly.

I never use it.. but I've seen it a lot and thought, "when did this come about?"

-Ben
_________________
Dazed and confused? I don't think so. Just "plain lost" will do. :D

mbradley · Joined: 11 Jul 2009 Posts: 118 Location: California, USA

Oh, the infamous 10u, I though I misunderstood your question (I have seen you around on the forum so I was like, no way is he asking this)

ok the 10u, I have no idea.

Perhaps just a MCHP variable, that is akin to the int8 that is not used by them, perhaps each compiler has there own way of doing it? actualy, I think I lost myself here. ok, it apears I have been of no help.
_________________
Michael Bradley
www.mculabs.com
Open Drivers and Projects

bkamen · Posted: Sat Feb 06, 2010 6:57 pm

mbradley · Joined: 11 Jul 2009 Posts: 118 Location: California, USA

OMG, now I feel dumb, I just saw this in another post, and I was like, ok, I learned something new..

http://www.ccsinfo.com/forum/viewtopic.php?p=127645&highlight=#127645

they talk about an L, as in 200L
_________________
Michael Bradley
www.mculabs.com
Open Drivers and Projects

bkamen · Posted: Sat Feb 06, 2010 7:37 pm

Right.

And still, I understand *what* is does.

Where I've seen it in MCHP's tcp stack, it's apparent the author wants to make absolute that a value like 0xf is in a 32bit int because the #define says to shift the bits over some amount greater than 4 (where f would get slid off the left end).

So there it makes sense.

In the case you linked where someone was setting duty cycle, I would think that would be unnecessary as the incoming value should get type converted properly. (well, I would HOPE.)

But if the case is programmers always wondering how a called function would convert a passed argument's type is in question: why don't I see more constant suffix usage than I do.

The plot thickens.
_________________
Dazed and confused? I don't think so. Just "plain lost" will do. :D

Ttelmah · Guest

The syntax is standard. Part of C itself.

The 'reason', is where they are reworking/writing examples, to be more portable between the PCD, ANSI, and 'normal' CCS syntax. Problem is that data sizes change with the versions, so where they want to force particular interpretation, they have started using this.
Older examples, generally don't use this, it is reserved for a very few later examples, where the size 'matters'.

Best Wishes

Ttelmah · Guest

Ongoing, the point about 'duty cycle', is that CCS has different behaviour for an 8bit value, and a long value. If a value is 8bit (it will be assumed to be so, if it is <256), the compiler puts it into the high 8bits of the 10bit duty cycle, leaving the two low bits unchanged. If instead the value is 16bit, it'll be written to the whole 10bits. Hence the size is 'critical' in this case.

Best Wishes

bkamen · Posted: Sun Feb 07, 2010 9:13 pm

I see what you're saying... I just found it interested that "suddenly" a different syntax was seen is if another programmer was writing the code with a different style.

Certainly makes sense now that something like the TCPIP stack runs on several different families of CPU's with different bit-sizes.

(shrug)

Anyway - just wondering.

Cheers,

-Ben
_________________
Dazed and confused? I don't think so. Just "plain lost" will do. :D

Ttelmah · Guest

Yes.
I'd suspect that one of the programmers has been 'caught' by the size problems, and has got into the habit of being explicit. Not a bad habit.... Smile

Best Wishes