On Mon, 21 Jun 1999, Bodo Moeller wrote:
> It will work with the current implementation (but only if the
> applications don't do too stupid things, such as point to a buffer
> with different contents); but this may change.
if the application does stupid things (i.e. does not check return value of
write/SSL_write/whatever) then it should be fixed. this check will not fix
them. instead, it just makes properly written programms to fail
undeterministically.
> Think of future protocol versions that may allow records so large that
> we don't want to store everything in a buffer at once; and then the
> encryption or hashing algorithms may use the particular alignment of
> the data as found during the first call to SSL_write.
well, records cannot be very large, because receiver must buffer entire
record in order to check the MAC. and caching pointers to user supplied
buffer between calls to some library function, instead of offsets will
hurt sooner or later.
i don't want to fight very hard to get this change in. i can always
comment out this check when new version arrives, so it's not big deal for
me, if you don't include this change. but there are others who will find
out this bug themselves and waste lot of valuable time. btw, i spent
almost a day using source navigator to be absolutely sure that this is ok
to remove this parameter check.
arne
______________________________________________________________________
OpenSSL Project http://www.openssl.org
Development Mailing List [EMAIL PROTECTED]
Automated List Manager [EMAIL PROTECTED]
