On Wed, 12 Dec 2012, Dominik George wrote: > > No, CRLF line endings will remain the same. > > > > In any event, if you read the MIME standard as specifying CRLF line > > endings in text/* MIME parts, then that's what the BTS provides, as it > > merely shoves the un-base64 encoded mime content back out to the > > client with the appropriate MIME headers. > > it actually base64-decodes it, so it does conversion on it. > It is *not* appropriate for a HTTP document of type text/* to have > CRLF line breaks if the file transferred does not have them.
How do you read that? If you read 2046 as specifying MIME text/* documents as having CRLF line endings, and HTTP uses MIME, and the document is text/plain, there doesn't seem to be much argument here. Furthermore, there's no way for the BTS to know whether the document should or should not have CRLF in it. > For any user of the BTS, what is hyperlinked from the bottom of a > message if a file is attached is a file, not some obscure MIME > thingy. The file is sent using MIME, as all HTTP transport is. > As I said, *ideally* file_in and file_out are identical to the byte. > But as a matter of fact, this actually *nerver* happens when > bugs.debian.org is involved. Each and every text/* file piped > through the BTS comes out broken. Or at least in the least expected > form, be that as it may. Your argument is that the BTS should send a text/plain MIME document with LF instead of CRLF, which conflicts with RFC 2046. Frankly, I think that's probably more sane, but that's because I think that the MIME standard is wrong is specifying a line ending standard for documents which have been QP or base64 encoded, and pine is wrong in changing the line encoding. Regardless, though, whatever is downloading the document should see that the MIME type is text/plain and that should do the line ending conversion. Don Armstrong -- We have to face the fact that either all of us are going to die together or we are going to learn to live together and if we are to live together we have to talk. -- Eleanor Roosevelt http://www.donarmstrong.com http://rzlab.ucr.edu -- To UNSUBSCRIBE, email to [email protected] with a subject of "unsubscribe". Trouble? Contact [email protected]
