[NNTP] Consensus?
Richard Clayton
richard at highwayman.com
Thu Aug 18 06:14:07 PDT 2005
-----BEGIN PGP SIGNED MESSAGE-----
Hash: SHA1
In message <43047C71.1060303 at nntpserver.com>, Steve Walker
<nntp at nntpserver.com> writes
>How many client/servers used
>'unsigned long' as the data type for article numbers?
only dumb ones...
> As soon as
>that client/server is compiled on a EM64 platform it suddenly
>becomes a 64 bit clean.
... and breaks its on-disk file structure.
This type of code will use some sort of generic name (int32, dword etc)
for the numbers and #define it to unsigned long, possibly by using a
configuration testing system to pick the correct set of #defines for the
machine it is being built on.
So as soon as it compiles it stays as it was! (everyone hopes)
> Servers/clients are going to evolve to
>64 bit on their own over the next couple years.
I agree, but it's nothing like as simple as you portray :(
>While a large number are using 32 bit numbers, another large
>group is using 31 bit numbers, and there is a smaller group that
>uses 64 bits or text based math. I do not think you can even
>claim 75% of users are using 32 bit clean software/servers.
10% of posters to Usenet use Google; 71% of distinct sites handling
articles are FIDO nodes ... appealing to numbers is not entirely
illuminating :)
- --
richard Richard Clayton
They that can give up essential liberty to obtain a little temporary
safety deserve neither liberty nor safety. Benjamin Franklin
-----BEGIN PGP SIGNATURE-----
Version: PGPsdk version 1.7.1
iQA/AwUBQwSJn5oAxkTY1oPiEQJPkACfWz2yHgyDIsjzHDaW4ZUS1Hu9jZ0AoIn4
ClXpm0h0C+wDHPmCaTnYFFt7
=qVMx
-----END PGP SIGNATURE-----
More information about the ietf-nntp
mailing list