While we may be tempted to go with the
reason that memcpy is faster (simply because it knows the no of
bytes to be copied in advance and doesn't need to compare each
byte-to-be-copied with a '\0', it is also implementation dependent
and performance could vary depending on if assembler optimizations
to copy several bytes at once are used or not), please note that
these are *functionally* different. While strcpy() will stop
copying on encountering the null terminator, memcpy will copy the
number of bytes passed in past the termination character. They are
meant for different things IMHO, depending on what you want :) In
general its probably better practice to stick to strcpy (or even
better, strncpy) when dealing with strings.
Anand
On 03/24/2014 12:35 PM, Atin Mukherjee wrote:
Anand
On 03/24/2014 12:35 PM, Atin Mukherjee wrote:
Hi List,
In GlusterFS codebase, strcpy() calls have been used in many places. IMO, if we can replace strcpy calls with memcpy() we would get a performance bonus as memcpy directly interacts with memory area. I am posting this suggestion based on my earlier project experience where we have seen a significant difference of performance with this replace.
Please feel free to add if you think otherwise.
Regards,
Atin Mukherjee
Senior Software Engineer
E-9, Extn : 73196, Direct : 08039245196, Mobile: +919739491377Email - address@hidden
_______________________________________________ Gluster-devel mailing list address@hidden https://lists.nongnu.org/mailman/listinfo/gluster-devel