nagel@wintermute.ICS.UCI.EDU (Mark Nagel) (08/12/90)
I have a fast malloc package I must have saved from some time back. I finally had occassion to try it out, but it doesn't even compile and the author (Piercarlo Grandi) left no contact address. The RCS header on this version is: malloc.c: $Header:/aware0/aware/piercarl/Src./Lib./Block./Block.c,v 3.3 89/03/15 17:36:54 piercarl Exp $ Doesn anyone (particularly Piercarlo, if he's out there listening) know of a newer version that works? It looks very nice except for that it appears not to have been tested on non-segmented architectures.
pcg@cs.aber.ac.uk (Piercarlo Grandi) (08/13/90)
On 11 Aug 90 18:12:16 GMT, nagel@wintermute.ICS.UCI.EDU (Mark Nagel) said: nagel> I have a fast malloc package I must have saved from some time back. nagel> I finally had occassion to try it out, but it doesn't even compile Are you sure? I have compiled it on half a dozen machines with no changes, except for parametrization of course (machines have been 386s, 286s, SUNs, DECs, ...). I have used a set of fairly funny #defines etc, which I normally keep in some header files of mine; but I have provided some equivalents within the source itself. Have you looked at it to enable them? nagel> and the author (Piercarlo Grandi) left no contact address. Where did you get the package? I have always posted it with full details... nagel> The RCS header on this version is: nagel> malloc.c: $Header:/aware0/aware/piercarl/Src./Lib./Block./Block.c,v 3.3 nagel> 89/03/15 17:36:54 piercarl Exp $ Fairly old, but I have not made that many changes in the meantime. I have used it for years (and haven't found a bug for a long long time -- but no warranties!). nagel> Doesn anyone (particularly Piercarlo, if he's out there listening) nagel> know of a newer version that works? I have sent off a new version (very lightly, mostly cosmetically, changed) to Rich $alz for comp.sources.unix, so it will appear there eventually, when he comes back from holidays. nagel> It looks very nice except for that it appears not to have been nagel> tested on non-segmented architectures. This is very strange. I did write the thing on sabi, a 286 with small baaaad segments, but I have now aware that is a 386 machine which runs a 32 bit linear address space, and I have run the code also on a stock Sun3. Are you sure you have enabled the right parameter definitions? As distributed it is configured for a 386 (#ifdef i386). If you want it to run on another system, you have to provide a new parametrization section. If you have a machine similar to a 386 (32 addresses and integers, linear addressing, byte oriented, etc...), you can just copy the 386 section. If this is the source of your problems (not realizing that there is a machine parameters section), than I should perhaps probably emphasize this more in the accompanying notes... Feel free to contact me by mail if you have specific problems... -- Piercarlo "Peter" Grandi | ARPA: pcg%uk.ac.aber.cs@nsfnet-relay.ac.uk Dept of CS, UCW Aberystwyth | UUCP: ...!mcsun!ukc!aber-cs!pcg Penglais, Aberystwyth SY23 3BZ, UK | INET: pcg@cs.aber.ac.uk