[1:41]<tlrobinson1> what would the best way to read a stream until a certain byte (e.x. EOT) be? [1:42]<ashb> in terms of API? [1:42]<tlrobinson1> sitting in a JavaScript while loop reading one character at a time seems inefficient [1:42]<tlrobinson1> given the APIs proposed so far [1:42]<ashb> ah. kind of does [1:42]<ashb> http://www.boost.org/doc/libs/1_37_0/doc/html/boost_asio/reference/async_read_until/overload1.html [1:43]<ashb> (jump to the end. there is am atching sync version) [1:43]<ashb> something like that perhaps? [1:43]<tlrobinson1> yea [1:43]<tlrobinson1> that would be good for readLine, etc [1:43]<tlrobinson1> though i guess there are multiple line endings :-/ [1:43]<ashb> eof is a special case of that i guess [1:44]<ashb> could make it a regexp without adding too much complexity [1:45]<ashb> also actualy reading a byte at time would be silly - it would probably make sense to buffer it even if you return it byte at a time to JS [1:45]<tlrobinson1> EOT = 0x04 and is different than "EOF" [1:46]<ashb> EOT? that one of those wierd control chars? [1:46]<tlrobinson1> yeah apparently [1:46]<tlrobinson1> i don't know much about it [1:46]<ashb> ah ASCII - what a wonderful legacy [1:46]<ashb> i'd have some special fale for EOF that isn't a number [1:46]<ashb> stream.readUntil(-1) etc or something [1:47]<tlrobinson1> basically i want to delimit a stream of text [1:47]<ashb> read it in 4k at a time and split? [1:47]<tlrobinson1> i don't want to close the stream between each [1:49]<tlrobinson1> yeah i guess that would be better [1:54]<tlrobinson1> i suppose i should really just use a length prefixed string