It is not. Thanks for the information anyway. The real problem consists of a set of tags, with a type identified by a byte. One of those tags is a compound tag which can contain any number of other tags and is terminated by a \0. This means I need parser behavior, rather than a linear regex-alike behavior. I would totally understand if Gloss does not, and will never provide this type of behavior. If it did, that'd mean you could write an XML parser in Gloss. :)
Thanks for your help. While Gloss might not solve this particular problem, I'm sure I'll find lots of other problems to use it for. Pepijn On Jan 10, 11:01 pm, Zach Tellman <ztell...@gmail.com> wrote: > I don't know if your example codec is as simple as your real problem, > but here's a codec that will work for the string you provided: > > (repeated > (string :utf-8 :delimiters ["\n" "\n\0"]) > :delimiters ["\n\0"] :strip-delimiters? false) > > This terminates the whole sequence only on \n\0, but doesn't strip out > the terminator so that it can be used by the last string as well. > Right now this will not properly encode (the last token will be \n\n > \0, so there will be a spurious empty string at the end of the list), > but I plan on fixing that soon. > > Zach > > On Jan 10, 11:47 am, "pepijn (aka fliebel)" <pepijnde...@gmail.com> > wrote: > > > > > > > > > Oh, right. My code does have a start and an end. I'm using header for > > the start. So the only way Gloss could do this is with a fn that > > combines header and repeated into one? > > > On Jan 10, 12:29 pm, Ken Wesson <kwess...@gmail.com> wrote: > > > > On Mon, Jan 10, 2011 at 4:52 AM, pepijn (aka fliebel) > > > > <pepijnde...@gmail.com> wrote: > > > > The later. The string in my example is \n terminated, so it should > > > > read past the \0, and then the outer list should terminate on the last > > > > \0. > > > > > My point is that I need to parse recursive structures, which of course > > > > contain the same terminator as the outer one. > > > > > What if I wanted to parse null terminated lists of null terminated > > > > lists of ... of bytes? Like so: > > > > > abc\0def\0\0 > > > > def\0ghi\0jkl\0\0 > > > > \0 > > > > > Your implementation would just make the outer list read to the first > > > > \0 and call it a day. What I need is that it only checks for a > > > > terminator on the outer list after reading a full inner list, and keep > > > > doing that, until it finds a terminator right after the inner list. > > > > That's not going to work unless all leaves are at the same depth and > > > there are no empty lists at any depth. If the leaves are all at the > > > same depth: > > > > (defn unflatten > > > ([in-seq sentinel] > > > (unflatten (seq in-seq) sentinel [[]] 0)) > > > ([in-seq sentinel stack height] > > > (if in-seq > > > (let [[f & r] in-seq] > > > (if (= f sentinel) > > > (let [h (inc height) > > > l1 (get stack h) > > > l1 (if l1 l1 []) > > > l2 (get stack height) > > > s1 (assoc stack h (conj l1 l2)) > > > s2 (assoc s1 height [])] > > > (recur r sentinel s2 (inc height))) > > > (let [l1 (get stack 0) > > > s1 (assoc stack 0 (conj l1 f))] > > > (recur r sentinel s1 0)))) > > > (peek stack)))) > > > > user=> (unflatten [1 2 0 3 4 0 0 5 6 0 7 8 0 0] 0) > > > [[[1 2] [3 4]] > > > [[5 6] [7 8]]] > > > > Otherwise you're going to need to have a start-of-list delimiter as > > > well as an end-of-list delimiter: > > > > (defn unflatten > > > ([in-seq start-del end-del] > > > (unflatten (seq in-seq) start-del end-del (list []))) > > > ([in-seq s e stack] > > > (if in-seq > > > (let [[f & r] in-seq] > > > (condp = f > > > e (let [[x y & z] stack] > > > (if y > > > (recur r s e (conj z (conj y x))) > > > (throw (Error.)))) > > > s (recur r s e (conj stack [])) > > > (let [[x & y] stack] > > > (recur r s e (conj y (conj x f)))))) > > > (if (> (count stack) 1) > > > (throw (Error.)) > > > (first stack))))) > > > > user=> (unflatten "(((ab)(cd))(ef)(g(hi)))" \( \)) > > > [[[[\a \b] [\c \d]] > > > [\e \f] > > > [\g [\h \i]]]] -- You received this message because you are subscribed to the Google Groups "Clojure" group. To post to this group, send email to clojure@googlegroups.com Note that posts from new members are moderated - please be patient with your first post. To unsubscribe from this group, send email to clojure+unsubscr...@googlegroups.com For more options, visit this group at http://groups.google.com/group/clojure?hl=en
