On Wed, 14 Jul 2010, Giuseppe Scrivano wrote:
Thanks for your report. I am going to apply this patch instead, it seems safer than consider just one value between RETR and SIZE.
I consider the size from SIZE to be much more reliable than the size you need to "guess" from the RETR response - based on the fact that SIZE has a documented way to return the exact size while RETR has not. No matter which happens to be the biggest...
-- / daniel.haxx.se
