On Mon, Oct 29, 2001 at 08:36:29AM +0100, Tobias Oetiker wrote: > Today Matt Zimmerman wrote: > > This doesn't sound like too much work, though it would require > > breaking backward compatibility. If the xdr routines are used for > > RPC, they should be pretty quick, and the code in glibc looks > > simple, so the performance impact should not be too great. > > > > Are there other changes already in progress which will break > > backward compatibility? That is, will the rrd_version be > > incremented? If so, I would be willing to do some work on > > implementing an xdr-based format. > > well there are features in the development tree which break backward > compatibility but they are done in a way, so that if they are not > used, still the old format is used ... > > I would imagine an implementation where usage of xdr is optional ... > if rrdtool is used in an embeded system it does not make sense to > require xdr ... even worse it may not be available on all platforms > ... > > I could see xdr even going into the 1.0.x tree if it was done in a way > where the format was recogniced from the magick cookie in the rrd > header ...
How about creating a different cookie for float_cookie to indicate that the RRD is architecture-independent? At configure time, we could check whether the system has xdr routines, and if so, compile in support for architecture-independent RRDs (in which xdr would be used). -- - mdz -- Unsubscribe mailto:[EMAIL PROTECTED] Help mailto:[EMAIL PROTECTED] Archive http://www.ee.ethz.ch/~slist/rrd-developers WebAdmin http://www.ee.ethz.ch/~slist/lsg2.cgi
