Derek Martin wrote: > This strikes me as wrong. It depends on how the caller passes end... > but this seems likely a fence post error to me (obviously I did not > look--I would do so now but I need to step out, so I just have time to > type this quick note). > > If the string is size 4, and starts at address 0, then it occupies > addresses 0-3. 3-0 - 3, not 4. > > Obviously if end is the next character after the end of the string, > this works as expected. If that is the case it should be called out > as such with a comment.
This is using the convention start <= i < end. (Like Python slicing or Ruby's "..." operator.) write_one_header() is using the same convention for its start and end values. I agree it can initially be confusing, but it's used in several places so I'm not sure it particularly merits a comment here. -Kevin
signature.asc
Description: PGP signature
