Re: pfNodeData Standard Approach

New Message Reply Date view Thread view Subject view Author view

Randy Stiles (stiles++at++aic.lockheed.com)
Wed, 23 Aug 95 12:04:03 PDT


   Date: Wed, 23 Aug 95 19:16 MET DST
   From: matomira++at++lig.di.epfl.ch (Fernando D. Mato Mira)

   I was just thinking that we need some guidelines for pfNodeData use,
   so that it's possible to merge different applications together.

   (eg: nodedata is a struct storing an array of void* (null entries allowed),
   and its length. The real nodedatas pointed to are structs whose first
   field is a long storing a tag (vendor, user, or application id). Tag numbers
   can be registered with SGI.
   )

Yes I agree. That is the approach I take, where there is an integer tag
that tells you what the struct pointer holds. This can still result in problems
when merging, but those problems can be resolved by redefining what the integer
tags mean, rather than rewriting alot of code, etc.

Another, perhaps more long-term approach would be to have user data be
a pfList, with these integer/pointer combos as elements...

-Randy

// Randy Stiles Office: 415.354.5256 Orgn 9620 Bldg 255
// stiles++at++aic.lockheed.com Fax: 415.354.5235 3251 Hanover Street
// Lockheed AI Center Lab: 415.424.2690 Palo Alto, CA 94304-1191
// http://hitchhiker.space.lockheed.com/~stiles/HOME.html


New Message Reply Date view Thread view Subject view Author view

This archive was generated by hypermail 2.0b2 on Mon Aug 10 1998 - 17:51:49 PDT

This message has been cleansed for anti-spam protection. Replace '++at++' in any mail addresses with the '@' symbol.