[rrd-developers] Re: rrd/cricket 64-bit clean ?

[Jeff R. Allen]

Peter Marquardt wrote:
> - On True64, DEC-Alpha sizeof(long) is 8, which affects mainly the
>   size of the rrd_format.h - structures (They rest as *.rrd on disk,
>   so the files might not be readable by other architectures).
> 
>   Is there any hope to make the datafiles kinda 'binary-portable' ?
>   (this would really speed up debugging: "hey, byte 14 must be
>   0x43!")

This is something you'd have to lobby Tobias to do for a future RRD
version. It seems like a good idea to me... since I have to maintain
RRD::Format as it is right now.

> - Perl's 'pack'-template mentions 'L':
> 
>   "  L   An unsigned long value.
>           (This 'long' is _exactly_ 32 bits, which may differ from
>            what a local C compiler calls 'long'.)  "

Yes, this would seem to be the problem. I think if you replace the
places where the current format uses 'L' with 'I', the right thing
might happen. To protect people from this mistake in the future,
getFormat.c would need to return formats with 'I' in them instead.

I'm a little scared about changing all the 'L's in Format.pm to 'I's
right now. Let's do that in version 1.1.x, so that it gets tested a
bit before we inflict it on people.

> - Nice examples, but I don't know if the output I get is, what it really
>   should look like!
> 
>   Any chance to put in some kind of "this_is_how_it_should_look_like.gif"
>   in the example-directory ?

I have just (a few minutes ago) proposed a canonical test config tree
which would allow all developers to be talking to the same
"routers". (It would actually be a program on the local machine that
would make data points.) Would that help?

>   - Missing Y-scales in rrdtool graph output (rrdtool-examples)

Perhaps this is because the data isn't getting in correctly. If RRD
thinks that there is nothing to graph, then it will not try to paint a 
Y axis.

-- 
Jeff R. Allen            |   jeff.allen at acm.org (personal)
HMC Alum, Class of '96   |   http://www.munitions.com/~jra

--
Unsubscribe mailto:rrd-developers-request at list.ee.ethz.ch?subject=unsubscribe
Help        mailto:rrd-developers-request at list.ee.ethz.ch?subject=help
Archive     http://www.ee.ethz.ch/~slist/rrd-developers