Fred's firmware development diary comments thread

[sry_not4sale]

Any chance of getting another "state of the onion" announcement Fred?

[MotoFab]

Any chance of getting another "state of the onion" announcement Fred?

"Development continues under current fief."

[Fred]

Any chance of getting another "state of the onion" announcement Fred?

Yeah, it's overdue, along with diyefi.org site update, documentation updates, and various other things. I'm planning to write up the USA trip story at some point very soon (hoping that donations will feed me a little longer) and with that, I'll go and hammer all the forums that I'm a member of with updates, including that state of the onion thread.

JHarvey, work is continuing almost every day, it's just not necessarily obvious. SeanK has been doing heaps of compiler dev work, and I've been chatting to him regularly about that, and supporting him with changes on my side where required. Plus other little bits of work.

Marcos, I'd roll with them with less hesitation if they were free of the _t. We'll see, I'll have a think about it and make a decision some time. In the mean time, as noted, search and replace will solve this issue very quickly if ever required anyway.

Fred.

[EssEss]

more of my 2c:

the _t is there to avoid collision since they usually reside in a .h. The .h is only file scope, but usually gets inserted across many translation units .. thus they end up more 'global' than intended most of the time.

From some real code:

you're more likely to get a collision from

Code: Select all

typedef unsigned long timer;

than:

Code: Select all

typedef unsigned long timer_t;

It's highly likely that someone is going to name one of their vars 'timer' and then they waste 3 days tracking down why their stuff breaks. Like I said before, it's usually the new guy that gets bit by this, so I do it for him . Whatever you decide, staying consistent throughout the codebase is more important than haggling over type names.

In case you haven't noticed (I think you already have), typedef's don't get you any real static safety:

Code: Select all

timer_t
timerAddTo( timer_t t, 
            unsigned long val )
{
    return t + val;
}

You could probably figure out what can happen there. Someone goes off and calls it w/the args reversed. The compiler would never detect it, and your code would actually appear to work correctly (in this simplified example). But, to be pedantic, it is being called incorrectly.

Instead, I lean on typedefs for abstracting out an interface most of the time. Which is why the various int[xx]_t's serve as a nice platform independent abstraction for me even if I never have the goal of platform independence. Its pure habit at this point based on previous headaches.

As you already know, if someone codes to the interface you can always change the implementation in the future with minimal fuss.

[Fred]

Sure, but the chances of someone naming their variable uint8 are slim to none, and slim just copped one from some n*****s 45! :-)

Thanks for the static type checking lesson, though. Good to know.

[sry_not4sale]

I just built FreeEMS with XGATE code in it! I estimate that we will have an XGATE release inside the next month. That would be cool All credit to SeanK who has been working his arse off on this. See the Advanced section for more detail. Good work young padawan.

Excuse the ignorance, but what difference does XGATE make to the release?

[Fred]

It's a milestone, that's all. And, I found out after I posted, that I was lying anyway. It was linked with the XGATE tools, but didn't include the xgate code. I should have RTFM (Sean's email) more thoroughly... I'll try to make sure that happens today, though. I' love to see that LED flashing by XGATE power. XGATE will be key to doing as much as we have dreamed of, well.

Fred.

[sry_not4sale]

This whiny bitch likes to stay informed!!! (thanks for the public update)

[Fred]

Seriously, I try to put everything vaguely important up, but you seem to want to know if I've sneezed on the board ;-) I'll try to keep enough stuff happening to justify posting and keep you happy ;-)

[sry_not4sale]

Seriously, I try to put everything vaguely important up, but you seem to want to know if I've sneezed on the board I'll try to keep enough stuff happening to justify posting and keep you happy

Even if you don't get much done, I think it's good for the community to see some sort of movement - even if it's "read some docs"