2011/5/5 William <[email protected]>
>
>
> Respectfully, may I ask you to explain your argument for modifying a
> working / tested routine that is in use in the field by multiple
> boards?
>

Sure, I'll explain it *again* but I'm happy that you finally ask, it wasn't
easy. I mean, I obviously didn't change this for nothing, there must have
been a reason to do so. And we discussed about this. Of course I'm not
expecting that you read all messages here in "real time" and chime on when
needed, but I sure am expecting you to search messages first, explore few
revision to get an overview of changes, and, at the very least, ask why we
did this, before actually asking to revert without even wondering why we did
this.

Carefully read this thread:
http://groups.google.com/group/jallib/browse_thread/thread/e4862f496ab5014e
Have a look at this:
http://code.google.com/p/jallib/source/detail?r=2586then this
http://code.google.com/p/jallib/source/detail?r=2587

And as a bonus, here's a summary: MSSP2 wasn't supported at all, we decided
to make it so, by duplicating existing library spi_master_hw.jal to
spi_master_hw2.jal. This included using suffixed registers (SSPBUF =>
SSP1BUF, for instance). I did check HEX file sums here, not after, sorry for
the confusion. Once done, I've been able to test it on SD-card. The trouble
is sd card was directly accessing SPI registers to set mode (not the same
layer if you see what I mean). SD card ideally would have to tell spi lib to
set mode, not directly acting on SPI registers. This way spi lib hides
registers to callers, and thus SD card lib is able to work on both MSSP
modules. Same for ethernet module as another example. Setting mode was
already part of the original init procedure, so we decided to split it into
set_speed + set_mode. So far all tests we could perform run well. There are
oviously still room for improvements. If you think "pragma inline" aren't
properly placed, suggest changes and report resource measure. If you think
it should be reverted, give arguments and explain why you're suspicious
about this change, why you think there's a bug and doesn't work.


> When possible, please discuss your proposed changes prior to check-
> in.


This is what we did, but now you know that.


> Consider how you'd feel if I started modifying the minix routines
> without making you aware of it first?
>

Oh well... I would ask why you did this, before claiming to revert it back.
But there's no point here since you was actually aware of our changes thanks
to the group mailing list & SVN.


>
> I'm not sure I can even track down all of the installations that are
> using this routine in the field, much less take time to re-test.


So ?


>  I'm
> not worried about samples, but real applications that utilize JALLIB.
>

We all are.


> Some of these are in locations that are thousands of miles away from
> my location.
>

I understand your concerns, but what is the solution here ? Stick to
revision 1 ? Forbid any future changes, any future release ?

SVN might help here, if you have worries about your applications, you can
stick to a certain revision you give enough confidence to.


Seb


> On May 5, 6:42 am, Sebastien Lelong <[email protected]>
> wrote:
> > Le 5 mai 2011 à 13:09, William <[email protected]> a écrit :
> >
> > > Hi Matt,
> >
> > > On May 5, 12:39 am, mattschinkel <[email protected]> wrote:
> >
> > >> can_mcp2515 does not use the new procedures. The init procedure could
> > >> be left as it was if that is what you prefer.
> >
> > > Yes, that would be my preference -- please restore the init routine to
> > > the original
> >
> > Why? Please give arguments !
> >
> > Seb
> >
> >
> >
> >
> >
> >
> >
> >
> >
> > >> Please leave the new procedures as they are, if you add pragma inline
> > >> to the new set mode procedure we will add code/data space to compiled
> > >> samples.
> >
> > > Here's an idea for you -- add 4 mode routines, and then inline each
> > > one -- will save code space and run faster.
> >
> > > William
>
>

-- 
You received this message because you are subscribed to the Google Groups 
"jallib" group.
To post to this group, send email to [email protected].
To unsubscribe from this group, send email to 
[email protected].
For more options, visit this group at 
http://groups.google.com/group/jallib?hl=en.

Reply via email to