[FIX] via source forge updates

John M McIntosh johnmci at smalltalkconsulting.com
Tue Feb 5 02:47:47 UTC 2002


>"Andreas Raab" <Andreas.Raab at gmx.de> is claimed by the authorities 
>to have written:
>
>>  Actually, trying to do this kind of meta-programming is rather pointless
>  > too. squeakInt64 and squeakFileOffsetType are completely unrelated, a
>>  system that has 64bit file support is not required to have intrinsic
>>  64bit integers; it could always define a file offset like:
>>  typedef struct {
>>	unsigned int lo;
>>	unsigned int hi;
>>  } file_offset_t;
>>  and vice versa; a system that has 64bit integers may not have 64bit file
>>  support. Mixing up the two definition is just a bad idea.
>Err, probably. I did actually suggest to John that squeak64Int be called
>something that tied it more closely to the use in the squeak 64 bit
>file offset arena so as to make it explicit. Using a structure as above
>within squeak would probably make some of this a bit simpler.
>
>tim

Actually squeakFileOffsetType needs to map to whatever the operating 
system uses for 64 bit file offsets. If there is no 64 bit offsets, 
then it needs to map to whatever is used for 32bit offsets.

squeakInt64 needs to map to what we want to use for 64 bit integers, 
right now that is long long int, or long int. It could be a 
structure, but then we start to need platform support. Right now 
mapping it to long long int, or long int so fine until some brave 
fellow wants to make it work on non-64bit compiler platforms.

-- 
--
===========================================================================
John M. McIntosh <johnmci at smalltalkconsulting.com> 1-800-477-2659
Corporate Smalltalk Consulting Ltd.  http://www.smalltalkconsulting.com
===========================================================================



More information about the Squeak-dev mailing list