Hi, Thanks for helping out. After using codecs rather than frames, I get even weirder errors.
My code now gives me: java.lang.Exception: Cannot evenly divide bytes into sequence of frames. Hard coding the header followed by a terminating :byte works as it should: (decode (compile-frame [tag tstring (header tag (memoize #(compile- frame [(name %) tstring (get-tag %)])) (comp symbol first)) :byte]) data) [:compound "hello world" ["string" "name" "Bananrama"] 0] So the problem seems to be with the repeated. Investigating that, I copied an example from the introduction page, and this is the result: (defcodec t (repeated (string :utf-8 :delimiters ["\n"]) :delimiters ["\0"])) (encode t ["foo" "bar" "baz"]) java.lang.IllegalArgumentException: Don't know how to create ISeq from: java.nio.HeapByteBuffer ((#<HeapByteBuffer java.nio.HeapByteBuffer[pos=0 lim=3 cap=3]> #<HeapByteBuffer java.nio.HeapByteBuffer[pos=0 lim=1 cap=1]>) (#<HeapByteBuffer java.nio.HeapByteBuffer[pos=0 lim=3 cap=3]> #<HeapByteBuffer java.nio.HeapByteBuffer[pos=0 lim=1 cap=1]>) But on the other hand: (decode t (.getBytes "blabla\nhihi\ngrrrr\n\0")) ["blabla" "hihi" "grrrr"] This gives me the same error as my code, but since the header in my code seems correct, I don't see why it has leftover bytes. (decode t (.getBytes "blabla\nhihi\ngrrrr\0")) java.lang.Exception: Cannot evenly divide bytes into sequence of frames. By the way, is there an easy way to get something readable out of encode? Like Unix hexdump, or even just a seq of integers. Debugging is a weak point in Gloss so far, if you ask me. Thanks! Pepijn de Vos On Jan 1, 10:47 pm, Zach Tellman <ztell...@gmail.com> wrote: > The header->body function in (header ...) must return a codec, so you > need to call compile-frame on the vector you're generating. Since you > don't want to call compile-frame every time you decode a frame, you > can memoize the function. A version that does both can be found > athttps://gist.github.com/762031. > > I agree that the way the enumeration and types are blurred in your > code is a little confusing. You could create a stronger distinction > by calling your enumerated types :tag-byte, :tag-int32, etc, and then > defining a map from those tags onto :byte, :int32, and so on. > > Zach > > On Jan 1, 1:01 pm, "pepijn (aka fliebel)" <pepijnde...@gmail.com> > wrote: > > > > > > > > > Hey, > > > I am trying Gloss for reading NBT [1] files. > > > First thing I did like is that it seems to make things real easy. > > First thing I did not like is the weak separation between types > > like :byte and extra data like :foo. > > > I think I'm nearly done with the NBT reader [2], but I ran into a > > problem. Whatever I put in the header form, I get exceptions like > > this: > > > java.lang.IllegalArgumentException: No implementation of > > method: :sizeof of protocol: #'gloss.core.protocols/Writer found for > > class: clojure.lang.PersistentVector > > > Only thing it mentions in the stacktrace [3] is methods on a reify, > > which calls the same method again, or in the most recent case, just > > return nil. > > > [1]http://www.minecraft.net/docs/NBT.txt > > [2]https://gist.github.com/761997 > > [3]http://pastebin.com/AqrsbjuS > > > On Nov 28 2010, 8:14 pm, Zach Tellman <ztell...@gmail.com> wrote: > > > > You're right, that's an omission from the frame syntax. I'll add the > > > ability for all or part of the frame to be scoped as (little- > > > endian ...) and (big-endian ...), with big-endian as the default. > > > > Just as a side-note, though, Calx [1] is already handling little- > > > endian data by using encode-to-buffer, where it's writing to a buffer > > > whose endianness has been preset. This obviously isn't a general > > > solution, but just thought I'd point it out. > > > > Zach > > > > [1]https://github.com/ztellman/calx > > > > On Nov 28, 8:50 am, zoka <ztomi...@gmail.com> wrote: > > > > > If Gloss is to decode incoming packet (byte array) in little-endian > > > > format it is straightforward: > > > > Wrap byte array into ByteBuffer b, invoke b.order(LITTLE_ENDIAN) and > > > > pass b to decode function that will return Clojure map of decoded > > > > values. > > > > > However, when outgoing packet byte array is to be produced from map of > > > > values, encode function will always return ByteBuffer in default big- > > > > endian format, so resulting byte array extracted form ByteBuffer using > > > > get() method will be incorrect. > > > > > If Gloss is to support little-endian frames, it seems that endianness > > > > needs to be part of frame definition. In that case Gloss decode fun > > > > would refuse to accept ByteBuffers with wrong order() and encode fun > > > > will always generate the correct result. > > > > > Zoka > > > > > On Nov 25, 3:00 am, Zach Tellman <ztell...@gmail.com> wrote: > > > > > > ByteBuffers have an order() method which allows you to toggle the > > > > > endianness. I haven't tested this, but since everything is built on > > > > > top of Java's ByteBuffer functionality it should be fine as long as > > > > > the ByteBuffers are correctly set and correctly ordered with respect > > > > > to each other. > > > > > > Zach > > > > > > On Nov 23, 2:52 pm, zoka <ztomi...@gmail.com> wrote: > > > > > > > JVM stores numbers in in big endian format - is there a way to > > > > > > process > > > > > > binary stream containing little endian numbers? > > > > > > > Zoka > > > > > > > On Nov 24, 7:24 am, Zach Tellman <ztell...@gmail.com> wrote: > > > > > > > > Good question. The solution didn't make the cut for my initial > > > > > > > release, but will be added soon. My plan is to have an (ordered- > > > > > > > map ...) frame which encodes and decodes the keys in the given > > > > > > > order. > > > > > > > So for C interop, the frame would be > > > > > > > > (ordered-map :a :int16, :b :float32) > > > > > > > > An alternative would be to just turn any vector which is > > > > > > > alternating > > > > > > > keys and types into an ordered-map, but that seems a bit too > > > > > > > magical. > > > > > > > > Zach > > > > > > > > On Nov 23, 12:12 pm, Chris Perkins <chrisperkin...@gmail.com> > > > > > > > wrote: > > > > > > > > > On Nov 23, 12:03 pm, Zach Tellman <ztell...@gmail.com> wrote: > > > > > > > > > > When writing Calx [1], I discovered it was a huge pain to > > > > > > > > > deal with > > > > > > > > > mixed C datatypes in Java. When writing Aleph [2], I > > > > > > > > > discovered the > > > > > > > > > problem increases by a factor of ten when dealing with > > > > > > > > > streams of > > > > > > > > > bytes. In an attempt to alleviate my own pain, and hopefully > > > > > > > > > help a > > > > > > > > > few other people out, I've written Gloss, which can transform > > > > > > > > > a simple > > > > > > > > > byte-format specification into an encoder and streaming > > > > > > > > > decoder. > > > > > > > > > > A full writeup can be found > > > > > > > > > athttps://github.com/ztellman/gloss/wiki. > > > > > > > > > > A few people have already asked me how this differs from > > > > > > > > > protocol > > > > > > > > > buffers, so I'll preemptively answer that protocol buffers > > > > > > > > > are a fixed > > > > > > > > > format that cannot be used to interface with external > > > > > > > > > systems. Gloss > > > > > > > > > is less performant than protocol buffers, but is also much > > > > > > > > > less picky > > > > > > > > > about formats. > > > > > > > > > > If anyone has any questions, I'd be happy to answer them. > > > > > > > > > Looks very useful, Zach. Thanks. > > > > > > > > > I have a question. > > > > > > > > > I have only taken a quick look, so maybe I'm misunderstanding > > > > > > > > the > > > > > > > > intent, but it's not clear to me how you would use this for > > > > > > > > sending > > > > > > > > and receiving structured data from, say, a C program. > > > > > > > > > Taking your example from the wiki: > > > > > > > > > (def fr (compile-frame {:a :int16, :b :float32})) > > > > > > > > > Let's say I want to talk to a C program that speaks in structs, > > > > > > > > like > > > > > > > > this: > > > > > > > > > struct Foo { short a; float b; } > > > > > > > > > The problem is, the C program cares about order - the short > > > > > > > > comes > > > > > > > > before the float. How does the Clojure program know what order > > > > > > > > I need > > > > > > > > the fields in, since I have specified the format with a map; an > > > > > > > > unordered data structure? Is there another way to specify a > > > > > > > > structure > > > > > > > > where order of the fields matters? If so, why have two ways of > > > > > > > > doing > > > > > > > > it? Or am I just missing something? > > > > > > > > > Thanks, > > > > > > > > > - Chris -- 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
