Re: [Chicken-users] Lowdown sxml-serializer not working together
Hi, Jim-- Thanks for all your work on this! On Sun, Aug 25, 2013 at 1:57 AM, Jim Ursetto zbignie...@gmail.com wrote: On Aug 24, 2013, at 5:25 PM, Jim Ursetto zbignie...@gmail.com wrote: Unfortunately it doesn't declare the default namespace for html. You can work around this by providing a default namespace prefix *and* using the actual prefix (below, xhtml:) on your elements. This is kind of ugly. However, the stock serializer didn't even support default namespaces (!) so I may have simply overlooked this case when adding support. I'll see if I can look into it further. I've added support for the default (unqualified) namespace ID in the source SXML now; the code is available at https://github.com/ursetto/sxml-serializer while I run a couple more tests. Formerly, you could render to unqualified names but not use them properly in the source. I'm not sure what you mean by use them properly in the source. Actually, other than the problem with lowdown, I was getting just the results I wanted before. Specifically, I'm working with my civet templating system, which I created a few months ago, and up to version 0.3 of sxml-serializer, I got output with a default namespace declaration and unqualified element names ... the default namespace was represented in my SXML as '(#f ns-uri) ... with sxml-serializer 0.4, they are getting an arbitrary prefix applied to them. So that's a step backwards for me. [Note: below, unqualified names are rendering to the xhtml: namespace, because there is a list of default namespace URI - ID mappings, called conventional-ns-prefixes. To override this and render them without qualification, add *default* as the 3rd arg of the namespace association (as the original prefix). I'm not following that. Where is there a 3rd argument? The word original is used as if the XML document was the original source and you had parsed it to SXML; ... which is in fact the case. the SXML prefix and XML prefix are permitted to differ.] I know, and it seems like an unnecessary complication to me ... but I suppose it must be a useful feature for some users. -- Matt Gushee ___ Chicken-users mailing list Chicken-users@nongnu.org https://lists.nongnu.org/mailman/listinfo/chicken-users
Re: [Chicken-users] Lowdown sxml-serializer not working together
Oh, wait ... On Fri, Aug 30, 2013 at 1:44 AM, Matt Gushee m...@gushee.net wrote: results I wanted before. Specifically, I'm working with my civet templating system, which I created a few months ago, and up to version 0.3 of sxml-serializer, I got output with a default namespace declaration and unqualified element names ... the default namespace was represented in my SXML as '(#f ns-uri) ... with sxml-serializer 0.4, they are getting an arbitrary prefix applied to them. So that's a step backwards for me. Oops. Disregard that. I need to read my own stuff better. What actually happened is that I just bumped the version number on the NSURI for the template vocabulary, but I forgot to update it in the templates I'm working with, so the template processor didn't recognize them. And just an hour ago I was reminding myself of the need to document that very thing! However, I see I'm not getting the default namespace declaration. I thought it was there before, but I'm not sure, and I seem to have deleted all the older output files. -- Matt Gushee ___ Chicken-users mailing list Chicken-users@nongnu.org https://lists.nongnu.org/mailman/listinfo/chicken-users
Re: [Chicken-users] Lowdown sxml-serializer not working together
On Aug 24, 2013, at 5:25 PM, Jim Ursetto zbignie...@gmail.com wrote: Unfortunately it doesn't declare the default namespace for html. You can work around this by providing a default namespace prefix *and* using the actual prefix (below, xhtml:) on your elements. This is kind of ugly. However, the stock serializer didn't even support default namespaces (!) so I may have simply overlooked this case when adding support. I'll see if I can look into it further. I've added support for the default (unqualified) namespace ID in the source SXML now; the code is available at https://github.com/ursetto/sxml-serializer while I run a couple more tests. Formerly, you could render to unqualified names but not use them properly in the source. [Note: below, unqualified names are rendering to the xhtml: namespace, because there is a list of default namespace URI - ID mappings, called conventional-ns-prefixes. To override this and render them without qualification, add *default* as the 3rd arg of the namespace association (as the original prefix). The word original is used as if the XML document was the original source and you had parsed it to SXML; the SXML prefix and XML prefix are permitted to differ.] (use sxml-serializer) (print (serialize-sxml '(*TOP* (@ (*NAMESPACES* (*default* http://www.w3.org/1999/xhtml;) (cvt http://xmlns.therebetygers.net/civet/0.1;))) (*PI* xml version=\1.0\ encoding=\utf-8\) (html (@ (xml:lang en) (lang en)) (cvt:template (cvt:head (cvt:locale (@ (lang en) (country US) (encoding utf-8 (cvt:block (@ (name headerContent)) #\ Single #\space (#\a #\r #\t #\i #\c #\l #\e) #\space page )) ?xml version=1.0 encoding=utf-8? xhtml:html xmlns:xhtml=http://www.w3.org/1999/xhtml; xml:lang=en lang=en cvt:template xmlns:cvt=http://xmlns.therebetygers.net/civet/0.1; cvt:head cvt:locale lang=en country=US encoding=utf-8 / /cvt:head cvt:block name=headerContentlt;Single article pagegt;/cvt:block /cvt:template /xhtml:html ___ Chicken-users mailing list Chicken-users@nongnu.org https://lists.nongnu.org/mailman/listinfo/chicken-users
Re: [Chicken-users] Lowdown sxml-serializer not working together
Matt, On Aug 22, 2013, at 3:28 PM, Matt Gushee m...@gushee.net wrote: Sounds good. I was actually hoping to be able to blame lowdown for this, since I don't really like the way it makes every word and every space a separate text node. But maybe that's more efficient? And actually, I'm wondering if that character-in-list construction: '(#\space) is really allowed by the SXML spec. The narrative documentation doesn't address that, and I'm a bit out of practice reading EBNF, but maybe I can puzzle it out. I've modified sxml-serializer in 0.3 to accept chars, symbols and nulls as text nodes, and to accept text nodes in list head position. This should allow it to parse lowdown's variant of SXML. I note that lowdown outputs the '' and '' in string form, whereas the '' becomes #\. That seems like a peculiar inconsistency, but I suppose there's some reason for it. Probably lowdown is using (abusing?) the fact that the universal-conversion-rules only escape string text nodes, whereas characters etc. are output verbatim; it's generally not necessary to escape the sign. Jim ___ Chicken-users mailing list Chicken-users@nongnu.org https://lists.nongnu.org/mailman/listinfo/chicken-users
Re: [Chicken-users] Lowdown sxml-serializer not working together
On Aug 23, 2013, at 11:24 PM, Matt Gushee m...@gushee.net wrote: Hello again-- Well, I'm running into a new problem with SRV:send-reply. By following the documentation, I managed to create a rule set that escaped problem characters i the manner I want. But now, working with the complete document, I discover that some of the non-element nodes are not handled properly. The SXML form of my document starts like this: (*TOP* (@ (*NAMESPACES* (#f http://www.w3.org/1999/xhtml;) (cvt http://xmlns.therebetygers.net/civet/0.1;))) (*PI* xml version=\1.0\ encoding=\utf-8\) (html (@ (xml:lang en) (lang en)) But when I serialize it with (SRV:send-reply (pre-post-order tree rules)), the output begins like this: *TOP* *NAMESPACES*=http://www.w3.org/1999/xhtml cvt=http://xmlns.therebetygers.net/civet/0.1; *PI*xmlversion=quot;1.0quot; encoding=quot;utf-8quot;/*PI* html xml:lang=en lang=en This should serialize reasonably correctly with sxml-serializer. (use sxml-serializer) (print (serialize-sxml '(*TOP* (@ (*NAMESPACES* (#f http://www.w3.org/1999/xhtml;) (cvt http://xmlns.therebetygers.net/civet/0.1;))) (*PI* xml version=\1.0\ encoding=\utf-8\) (html (@ (xml:lang en) (lang en)) (cvt:template (cvt:head (cvt:locale (@ (lang en) (country US) (encoding utf-8 (cvt:block (@ (name headerContent)) #\ Single #\space (#\a #\r #\t #\i #\c #\l #\e) #\space page)) ?xml version=1.0 encoding=utf-8? html xml:lang=en lang=en cvt:template xmlns:cvt=http://xmlns.therebetygers.net/civet/0.1; cvt:head cvt:locale lang=en country=US encoding=utf-8 / /cvt:head cvt:block name=headerContentlt;Single article pagegt;/cvt:block /cvt:template /html Unfortunately it doesn't declare the default namespace for html. You can work around this by providing a default namespace prefix *and* using the actual prefix (below, xhtml:) on your elements. This is kind of ugly. However, the stock serializer didn't even support default namespaces (!) so I may have simply overlooked this case when adding support. I'll see if I can look into it further. (print (serialize-sxml '(*TOP* (@ (*NAMESPACES* (xh http://www.w3.org/1999/xhtml;) (cvt http://xmlns.therebetygers.net/civet/0.1;))) (*PI* xml version=\1.0\ encoding=\utf-8\) (xh:html (@ (xml:lang en) (lang en)) (cvt:template (cvt:head (cvt:locale (@ (lang en) (country US) (encoding utf-8 (cvt:block (@ (name headerContent)) #\ Single #\space (#\a #\r #\t #\i #\c #\l #\e) #\space page ns-prefixes: '((*default* . http://www.w3.org/1999/xhtml;)) )) ?xml version=1.0 encoding=utf-8? html xmlns=http://www.w3.org/1999/xhtml; xml:lang=en lang=en cvt:template xmlns:cvt=http://xmlns.therebetygers.net/civet/0.1; cvt:head cvt:locale lang=en country=US encoding=utf-8 / /cvt:head cvt:block name=headerContentlt;Single article pagegt;/cvt:block /cvt:template /html Namespaces can be tricky. Check my blog posts on the subject and feel free to ask for help. http://3e8.org/blog/2010/07/30/namespaces-in-sxml-part-1/ http://3e8.org/blog/2010/07/31/namespaces-in-sxml-part-2/ http://3e8.org/blog/2010/08/01/default-namespaces-in-sxml/ Jim ___ Chicken-users mailing list Chicken-users@nongnu.org https://lists.nongnu.org/mailman/listinfo/chicken-users
Re: [Chicken-users] Lowdown sxml-serializer not working together
Hello again-- Well, I'm running into a new problem with SRV:send-reply. By following the documentation, I managed to create a rule set that escaped problem characters i the manner I want. But now, working with the complete document, I discover that some of the non-element nodes are not handled properly. The SXML form of my document starts like this: (*TOP* (@ (*NAMESPACES* (#f http://www.w3.org/1999/xhtml;) (cvt http://xmlns.therebetygers.net/civet/0.1;))) (*PI* xml version=\1.0\ encoding=\utf-8\) (html (@ (xml:lang en) (lang en)) But when I serialize it with (SRV:send-reply (pre-post-order tree rules)), the output begins like this: *TOP* *NAMESPACES*=http://www.w3.org/1999/xhtml cvt=http://xmlns.therebetygers.net/civet/0.1; *PI*xmlversion=quot;1.0quot; encoding=quot;utf-8quot;/*PI* html xml:lang=en lang=en ... which is obviously not quite right. I suppose it should be possible to create conversion rules for the *TOP*, *NAMESPACES*, and *PI* nodes, but I don't understand the syntax well enough. Are there any examples out there of how to do this? -- Matt Gushee On Thu, Aug 22, 2013 at 3:01 PM, Moritz Heidkamp mor...@twoticketsplease.de wrote: Hi Matt, Matt Gushee m...@gushee.net writes: markdown-sxml converts whitespace in the input to character objects, or lists of character objects, e.g.: '(#\space) #\newline ... but serialize-sxml can't handle these: evidently it wants either strings or symbols. I haven't found any arguments or parameters in either egg that seem to be able to affect this behavior. yeah, Lowdown is cheating a bit here. Those individual characters fall right out of the parser and it would require an additional walk of the result to turn them into strings. As you already found out, this is not really part of the SXML spec but most tools allow characters just fine (sxml-serializer is an exception here and is particularly strict). Same applies to the additional list wrapping. I usually serialize SXML like this: (use sxml-transforms) (define some-data '(some #\d (#\a ta))) (SRV:send-reply (pre-post-order some-data) universal-conversion-rules)) Which outputs: somedata/some As you can see, universal-conversion-rules and SRV:send-reply (what's the deal with that name?) don't care much about the structure. Note that characters will not be entity escaped though, they are just passed through as-is. However, I do take your point and I think if I advertise Lowdown as producing SXML I should better change it to produce SXML according to spec :-) I hope I will get around to changing this soon but can't promise when. I also hope the workaround above will serve you until then! Moritz ___ Chicken-users mailing list Chicken-users@nongnu.org https://lists.nongnu.org/mailman/listinfo/chicken-users
Re: [Chicken-users] Lowdown sxml-serializer not working together
Hi, Jim-- On Wed, Aug 21, 2013 at 10:12 PM, Jim Ursetto zbignie...@gmail.com wrote: I don't remember offhand what the rationale was for this in sxml-serializer -- I can take a look but it might not be until this weekend. Sounds good. I was actually hoping to be able to blame lowdown for this, since I don't really like the way it makes every word and every space a separate text node. But maybe that's more efficient? And actually, I'm wondering if that character-in-list construction: '(#\space) is really allowed by the SXML spec. The narrative documentation doesn't address that, and I'm a bit out of practice reading EBNF, but maybe I can puzzle it out. Anyway, thanks. Guess I'll just work around it for now. -- Matt Gushee ___ Chicken-users mailing list Chicken-users@nongnu.org https://lists.nongnu.org/mailman/listinfo/chicken-users
Re: [Chicken-users] Lowdown sxml-serializer not working together
Hi Matt, Matt Gushee m...@gushee.net writes: markdown-sxml converts whitespace in the input to character objects, or lists of character objects, e.g.: '(#\space) #\newline ... but serialize-sxml can't handle these: evidently it wants either strings or symbols. I haven't found any arguments or parameters in either egg that seem to be able to affect this behavior. yeah, Lowdown is cheating a bit here. Those individual characters fall right out of the parser and it would require an additional walk of the result to turn them into strings. As you already found out, this is not really part of the SXML spec but most tools allow characters just fine (sxml-serializer is an exception here and is particularly strict). Same applies to the additional list wrapping. I usually serialize SXML like this: (use sxml-transforms) (define some-data '(some #\d (#\a ta))) (SRV:send-reply (pre-post-order some-data) universal-conversion-rules)) Which outputs: somedata/some As you can see, universal-conversion-rules and SRV:send-reply (what's the deal with that name?) don't care much about the structure. Note that characters will not be entity escaped though, they are just passed through as-is. However, I do take your point and I think if I advertise Lowdown as producing SXML I should better change it to produce SXML according to spec :-) I hope I will get around to changing this soon but can't promise when. I also hope the workaround above will serve you until then! Moritz ___ Chicken-users mailing list Chicken-users@nongnu.org https://lists.nongnu.org/mailman/listinfo/chicken-users
Re: [Chicken-users] Lowdown sxml-serializer not working together
Hi, Moritz-- On Thu, Aug 22, 2013 at 3:01 PM, Moritz Heidkamp mor...@twoticketsplease.de wrote: (use sxml-transforms) (define some-data '(some #\d (#\a ta))) (SRV:send-reply (pre-post-order some-data) universal-conversion-rules)) Thanks for the suggestion. I'm now working on this. the deal with that name?) don't care much about the structure. Note that characters will not be entity escaped though, they are just passed through as-is. That could be a problem, since I am outputting web pages ... I just did an experiment where I inserted '', '', '' . I note that lowdown outputs the '' and '' in string form, whereas the '' becomes #\. That seems like a peculiar inconsistency, but I suppose there's some reason for it. Anyway, am I right in thinking that I can get escaped characters by using something other than universal-conversion-rules? -- Matt Gushee ___ Chicken-users mailing list Chicken-users@nongnu.org https://lists.nongnu.org/mailman/listinfo/chicken-users
Re: [Chicken-users] Lowdown sxml-serializer not working together
I don't remember offhand what the rationale was for this in sxml-serializer -- I can take a look but it might not be until this weekend. On Aug 21, 2013, at 22:38, Matt Gushee m...@gushee.net wrote: Hi, chickeneers-- I am working on an application that uses the lowdown egg to process *some* of the input, and sxml-serializer to render the output. I'm running into an issue ... I'm not sure if it's really a bug, but it's at least an compatibility problem: markdown-sxml converts whitespace in the input to character objects, or lists of character objects, e.g.: '(#\space) #\newline ... but serialize-sxml can't handle these: evidently it wants either strings or symbols. I haven't found any arguments or parameters in either egg that seem to be able to affect this behavior. So, is there something that needs to be fixed in one of these eggs, or do I just need to convert the markdown-sxml output to something that serialize-sxml can handle? Please advise. -- Matt Gushee ___ Chicken-users mailing list Chicken-users@nongnu.org https://lists.nongnu.org/mailman/listinfo/chicken-users ___ Chicken-users mailing list Chicken-users@nongnu.org https://lists.nongnu.org/mailman/listinfo/chicken-users
