What I was getting at was not that I thought you would require the exact RFC 822 syntax, but that you would need a specific syntax in general, and it would be hard to guarantee that the user would follow that syntax and not screw things up before parsing. The <<EXCERPT idea is more along the lines of what I was thinking, and a snippet could insert the necessary code. Couldn't you just have the header start with a word and a colon, and then the excerpt be terminated by something like ENDEXCERPT?
Do you want to preserve the CR? Do they need to be escaped before posting, or are they stripped in parsing?
On Nov 4, 2006, at 5:15 AM, Allan Odgaard wrote:
On 4. Nov 2006, at 10:38, Brett Terpstra wrote:
If I understand RFC 822 correctly, multi-line headers require specific syntax (CRLF followed by a single LWSP and terminated by a CR). I'll leave the details up to you, and I'm probably overlooking something simple, I just don't quite grasp how you're going to make it simple for the average blogger to add an excerpt which can be easily parsed as separate from the message body if there's nothing to terminate it. I'm just curious.
We wouldn’t follow it strictly, just the idea, so e.g. drop everything about CR.
It boils down to: a header starts with a word, then a colon. If the next line has indent, that indent is removed, and the header continues at that line.
So a simple post:
Title: foo Excerpt: Lorem ipsum dolor sit amet, consectetur adipisicing elit, sed do eiusmod tempor incididunt ut labore et dolore magna aliqua. And here is the actual body.
Or said another way, the header stops with the first non-indented and non-empty line.
A problem though is that we wouldn’t be able to support Markdown for the excerpt, at least not code blocks and hard wrapped list items.
Maybe we should go for the here-doc syntax. So it would be:
Title: foo Excerpt: <<EXCERPT Lorem ipsum dolor sit amet, consectetur adipisicing elit, sed do eiusmod tempor incididunt ut labore et dolore magna aliqua. EXCERPT And here is the actual body.
But then the question is if there should be a way to escape the << used to provide a token. Though probably just better to not allow arbitrary tokens, so the chance of a false positive is highly unlikely.
For new threads USE THIS: textmate@lists.macromates.com (threading gets destroyed and the universe will collapse if you don't) http://lists.macromates.com/mailman/listinfo/textmate
Brett Terpstra : Art Director Circle Six Design, Inc. 111 Riverfront Dr, Suite 204 .................................................. p: 507.459.4398 877.858.4332 f: 1.866.540.3063 e: brett@circlesixdesign.com http://www.circlesixdesign.com ..................................................