fractalnavel : metablog

tech notes: twitter aggregator refactoring #1

fractalnavel — Sat, 21 Mar 2009 21:26:31 GMT

it usually takes me three rounds of coding to get something somewhat satisfactory - the original "just get it to work" phase, and two refactorings. the first refactoring follows, with some notes in the comments for the next round:

fractalnavel.CS.PostTwitter.cs

using System;
using System.Xml;

using CommunityServer.Blogs.Components;

using CookComputing.XmlRpc;

using fractalnavel.CS.components;

namespace fractalnavel.CS
{

    // TODO: 
    //  - log processing status; 
    //  - loop through multiple service/settings nodes (optionally post all together?)
    class CMain
    {
        [STAThread]
        static void Main(string[] args)
        {
            XmlDocument xmlSettings = new XmlDocument();
            xmlSettings.Load( args[0] ); 
            GetAndPost gap = new GetAndPost( 
                xmlSettings.DocumentElement, 
                (XmlRpcProxyGen.Create<IMetaWebLogNewPost>()).newPost 
                );

            if ( gap.Get() ) 
                gap.Post();
            else if ( gap.PostNone )
                gap.Post( "nothin'" );
        }
    }

    [XmlRpcUrl("*** wherever it was put *** /metablog.ashx")] 
    public interface IMetaWebLogNewPost
    {
        [XmlRpcMethod("metaWeblog.newPost")]
        string newPost(
            string blogid,
            string username,
            string password,
            MetaWeblog.Post post,
            bool publish
            );
    }
    
}

fractalnavel.CS.components.cs

using System;
using System.Xml;
using System.Xml.Serialization;

using CommunityServer.Blogs.Components;

using fractalnavel.components;

namespace fractalnavel.CS.components
{

    // get xml, transform to post, post to blog.    
    public class GetAndPost
    {
        private MakeNewPost _doPost;
        private Settings _settings;
        private XmlDocument _xmldocAll = new XmlDocument();
        private string _strPostBody = string.Empty;

        public delegate string MakeNewPost(
            string blogid,
            string username,
            string password,
            MetaWeblog.Post post,
            bool publish
            );

        public GetAndPost( XmlNode nodSettings )
        {
            GetAndPost( nodSettings, (new MetaWeblog()).newPost );
        }

        public GetAndPost( XmlNode nodSettings, MakeNewPost DoPost )
        {
            _doPost = DoPost;
            _settings = (Settings) XmlUtils.Deserialize <Settings> ( nodSettings );
        }
        
        public bool PostNone {
            get{ return _settings.bPostNone; }
        }

        public string PostBody
        {
            set {_strPostBody = value;}
            
            // TODO: add xsl parameter handling (will pass various settings)
            get 
            {
                if ( _strPostBody == string.Empty )
                    _strPostBody = XmlUtils.Transform( _xmldocAll, _settings.hrefXsl );
                return _strPostBody;
            } 
        }

        /* 
        returns true if results contain status nodes
        ***twitter specific at the moment.***
        TODO: 
            - factor the href & success into a separate service interface, 
            - implement for each service, 
            - specify implementation class/assembly in config element, 
            - instantiate here, 
            - pass settings to service interface, 
            - and consume.
         */
        public bool Get() 
        {
            // Tue%2C+27+Mar+2007+22%3A55%3A48+GMT
            string href = _settings.hrefGet + "?count=200&since=" + ((DateTime.UtcNow).AddHours( - _settings.hours )).ToString("r");
            _xmldocAll.LoadXml( XmlUtils.GetTextFromHref( href, _settings.useridGet, _settings.pwdGet ) );
            return (_xmldocAll.SelectNodes("//status")).Count > 0;
        }
        
        public string Post()
        {
            return Post( this.PostBody );
        }
        
        // returns postid
        public string Post( string strPostBody )
        {
            bool bPublish = true;
            DateTime dtNow = DateTime.Now;
            MetaWeblog.Post post = new MetaWeblog.Post();
            post.title = _settings.titleBlog + " " + (dtNow).ToString("yyyy.MM.dd");
            post.dateCreated = dtNow;
            post.description = strPostBody;
            return _doPost( _settings.idBlog, _settings.useridBlog, _settings.pwdBlog, post, bPublish);
        }
    
        /*
        Settings element attributes:
            hrefget - get: string (must return results as xml; bascially assuming a rest interface)
            useridGet, pwdget - get: authentication = userid, password (yeah, not very secure, need to do something different)
            hrefXsl - transform reference (url; this xslt is applied to the "get" xml, and will result in the post body)
            idBlog - post: destination blog
            useridBlog, pwdBlog - post: authentication (userid, password) (again, security issue)
            titleBlog - post title (appended with date)
            hours - how far back to look
            bPostNone - post when nothing was retrieved ?
        */
        [Serializable()]
        public class Settings
        {
            private string _hrefGet;
            private string _useridGet;
            private string _pwdGet;
            private string _hrefXsl;
            private string _idBlog;
            private string _useridBlog;
            private string _pwdBlog;
            private string _titleBlog;
            private int _intHours;
            private bool _bPostNone;

            public Settings() {}

            [XmlAttribute("hrefGet")]
            public string hrefGet{get {return this._hrefGet;} set {this._hrefGet = value;}}

            [XmlAttribute("useridGet")]
            public string useridGet{get {return this._useridGet;} set {this._useridGet = value;}}

            [XmlAttribute("pwdGet")]
            public string pwdGet{get {return this._pwdGet;} set {this._pwdGet = value;}}

            [XmlAttribute("hrefXsl")]
            public string hrefXsl{get {return this._hrefXsl;} set {this._hrefXsl = value;}}

            [XmlAttribute("idBlog")]
            public string idBlog{get {return this._idBlog;} set {this._idBlog = value;}}

            [XmlAttribute("useridBlog")]
            public string useridBlog{get {return this._useridBlog;} set {this._useridBlog = value;}}

            [XmlAttribute("pwdBlog")]
            public string pwdBlog{get {return this._pwdBlog;} set {this._pwdBlog = value;}}

            [XmlAttribute("titleBlog")]
            public string titleBlog{get {return this._titleBlog;} set {this._titleBlog = value;}}

            [XmlAttribute("hours")]
            public int hours{get {return this._intHours;} set {this._intHours = value;}}

            [XmlAttribute("bPostNone")]
            public bool bPostNone{get {return this._bPostNone;} set {this._bPostNone = value;}}
        }

    }
    
}

fractalnavel.components.cs

using System;
using System.IO;
using System.Net;
using System.Text;
using System.Xml;
using System.Xml.Serialization;
using System.Xml.Xsl;

namespace fractalnavel.components
{

    public class XmlUtils
    {
        public static string GetTextFromHref(string url, string user, string password)
        {
            HttpWebRequest request = (HttpWebRequest)HttpWebRequest.Create(url);
            request.Credentials = new NetworkCredential(user, password);
            WebResponse response = request.GetResponse();
            StreamReader reader = new StreamReader(response.GetResponseStream());
            string responseString = reader.ReadToEnd();
            reader.Close();
            return responseString;
        }
        
        public static string Transform( string strXmlHref, string strXmlUserId, string strXmlPassword, string strXslHref )
        {
            XmlDocument xmlDoc = new XmlDocument();
            xmlDoc.LoadXml( GetTextFromHref( strXmlHref, strXmlUserId, strXmlPassword ) );
            return Transform( xmlDoc, strXslHref );
        }
        
        // TODO: add xsl parameter handling
        public static string Transform( XmlDocument xmlDoc, string strXslHref )
        {
            XslTransform xsl = new XslTransform();
            xsl.Load( strXslHref, new XmlUrlResolver() );
            StringBuilder sbResult = new StringBuilder();
            StringWriter swResult = new StringWriter( sbResult );
            xsl.Transform( xmlDoc, null, swResult, null );
            return sbResult.ToString();
        }   

        public static Object Deserialize <T> (XmlNode node) 
        {
            XmlSerializer ser = new XmlSerializer( typeof(T) );
            return ser.Deserialize(new XmlNodeReader(node));
        }

    }

}

Settings.xml

<?xml version="1.0" encoding="utf-8" ?>
<!--
    TODO: 
        - use an array of settings nodes for separate services; 
        - specify service handler classs/assembly for items that need programmatic suport (eg, href) 
-->
<Settings 
    hrefGet="http://twitter.com/statuses/user_timeline.xml" 
    useridGet="***" 
    pwdGet="***" 
    hrefXsl="*** wherever it was put *** /GetAndPost.xsl" 
    idBlog="*** destination weblog name ***" 
    useridBlog="***" 
    pwdBlog="***" 
    titleBlog="a day in a life" 
    hours="24" 
    bPostNone="false" 
    />

GetAndPost.xsl

<?xml version="1.0" ?>
<xsl:stylesheet 
    version="1.0" 
    xmlns:xsl="http://www.w3.org/1999/XSL/Transform"
    xmlns:user="http://blogs.no-ip.org/fractalnavel"
    xmlns:msxsl="urn:schemas-microsoft-com:xslt"
    exclude-result-prefixes="xsl user msxsl" 
    >

    <xsl:output 
        method='html' 
        omit-xml-declaration="yes"
        version="1.0"
        encoding="UTF-8"
        indent="yes"
        cdata-section-elements=""
        /> 
    
    <!-- TODO: add parameters handling for settings from GetAndPost.Settings class -->
    <xsl:template match="/statuses">

        <div class="divLTAll">

            <div class="divLTHead">
                via @<a href="http://www.twitter.com/*** who you are ***" target="_blank" 
                title="twitter!">*** who you are ***</a>: in the last twenty-four hours:
            </div>
        
            <div class="divLTBody">
                <ul>
                    <xsl:apply-templates select="status" >
                        <xsl:sort select="position()" order="descending" data-type="number"/>
                    </xsl:apply-templates>
                </ul>
            </div>

            <div class="divLTFoot">
                (pulled direct from twitter via custom job)
            </div>
            
        </div>

    </xsl:template>

    <xsl:template match="status">
        <li><span class="spnLTItemDate"><xsl:value-of select="user:fnNormalizeDate(string(created_at))" /></span>: 
        <xsl:value-of select="user:fnSetLinks(string(text))" disable-output-escaping="yes" /></li>
    </xsl:template>

    <msxsl:script language="JScript" implements-prefix="user">
    //<![CDATA[
    
    function fnSetLinks( strText )
    {
        // embedded hrefs
        strText = strText.replace( /(http:\/\/\S*)/g, "<a href=\"$1\" class=\"aLTLink\" target=\"_blank\">$1</a>" );
        // twitter ids
        strText = strText.replace( /@(\S*)/g, "@<a href=\"http://www.twitter.com/$1\" class=\"aLTUser\" target=\"_blank\">$1</a>" );
        // hash tags
        strText = strText.replace( /#(\S*)/g, "<a href=\"http://search.twitter.com/search?q=%23$1\" class=\"aLTHash\" target=\"_blank\">$1</a>" );
        return strText;
    }
    
    function fnNormalizeDate(strDate) {
        try {
            // "Mon Mar 16 18:16:59 +0000 2009" to "2009.03.16 18:16": 
            var rxMatch = /.{3} (.{3}) (\d{2}) (\d{2}:\d{2}):\d{2} \+\d{4} (\d{4})/;
            strDate = strDate.replace( rxMatch , 
                function(strMatch, strMonth, strDay, strTime, strYear)
                {
                        var dat = new Date( strMonth + " " + strDay + " " + strYear + " " + strTime );
                        // months number is zero based ?!?  wtf
                        strMonth = (dat.getUTCMonth()+1).toString().replace( /^(\d)$/, "0$1" );
                        return strYear + "." + strMonth + "." + strDay + " " + strTime; 
                });
        } catch(e) {
            strDate = e.message;
        }
        return strDate ;
    }

    //]]>
    </msxsl:script>

</xsl:stylesheet>

i also need to reorganize the naming a bit. i'll reserve commentary until the last revisions are complete.

guano machine

fractalnavel — Tue, 17 Mar 2009 21:32:35 GMT

i poked around twitter a bit recently while I've been playing, and i noticed that there were basically two styles: the originally intended open-loop micro-blogging, and the chat. people don't seem to have figured out how to really use micro-blogging, they either fell back into chat patterns, or just compulsively post trivia. the news channels aren't so bad, but that's what used to be professionally provided via rss from the news websites.

so, i wonder: how has im traffic been faring lately ? i bet it's down by an amount corresponding to twitter use. not sure what im'ing in public gets you. and twitter makes a lousy chat room.

as for the micro blogging, what proportion of the updates out there are bodily function related ? "went to the gym" "got back from the gym" "at the gym" "tired / cold / hot / sleeping / stupid / omg, what's this infection, my nuts have just dropped off!". does anyone really want to know that ? does anyone really want others to know that ?

then there are those projecting "image". professional, cool, wacky, dark - all that self-indulgent cvrap.

here's what i had in mind, and why i haven't really used it, and why I've been tinkering with daily accumulation and posting of micro-posts: sometimes i have quick thoughts, or a stream of thinking, that doesn't quite go together in a longer coherent post yet. but i want to capture that stuff that's been getting lost to date. stuff that i don't email, im or talk to anyone about. in other words, i'm trying to catch another source of information. there are times when i'm just thinking about economics, or cosmogony, or metaphysics, or whatever. a string of what-ifs, conjectures, speculations, maybe an aha or two. and then they're gone.

so, maybe this new lighter weight tool might be of help there. i'll see. now that i have my tools built, i'll try it out, when i'm having one of those times. like anything else, it'll take practice. and maybe it just won't float at all. expectations of the lowest order would be of help here.

as for the twitter-world-as-it-is, as with any popular phenomena, there are a lot of critics popping up - and they're right.

so sure, there's unprecedented openness, but no one seems to know what do with it, on either end of the pipe. it's like i noted a long time ago, privacy is generally no big deal, outside of personal or institutional maliciousness. ever seen someone else's email ? once you get over the voyeurism, it's just boring crap. but maybe we never get over the voyeurism.

once upon a time i wanted to open up all exchanges anyway, making it able to delve into causality in new ways. and it's slowly been happening. global mind ? not yet. a demonstration of illusory free will / consciousness ? no one will pay attention (proving the point).

so - i have a rock. i wonder what i can bang with it.

tech notes: touching up

fractalnavel — Tue, 17 Mar 2009 19:30:40 GMT

re my recent playing, "final" xsl:

    
<?xml version="1.0" ?>
<xsl:stylesheet 
    version="1.0" 
    xmlns:xsl="http://www.w3.org/1999/XSL/Transform"
    xmlns:user="http://blogs.no-ip.org/fractalnavel"
    xmlns:msxsl="urn:schemas-microsoft-com:xslt"
    exclude-result-prefixes="xsl user msxsl" 
    >

    <xsl:output 
        method='html' 
        omit-xml-declaration="yes"
        version="1.0"
        encoding="UTF-8"
        indent="yes"
        cdata-section-elements=""
        /> 

    <xsl:template match="/statuses">

        <div class="divLTAll">

            <div class="divLTHead">
                via @<a href="http://www.twitter.com/fractalnavel" target="_blank" 
                title="twitter!">fractalnavel</a>: in the last twenty-four hours:
            </div>

            <div class="divLTBody">
                <ul>
                    <xsl:apply-templates select="status" >
                        <xsl:sort select="position()" order="descending" data-type="number"/>
                    </xsl:apply-templates>
                </ul>
            </div>

            <div class="divLTFoot">
                (pulled direct from twitter via custom job)
            </div>

        </div>

    </xsl:template>

    <xsl:template match="status">
        <li><span class="spnLTItemDate"><xsl:value-of select="user:fnNormalizeDate(string(created_at))" /></span>: 
        <xsl:value-of select="user:fnSetLinks(string(text))" disable-output-escaping="yes" /></li>
    </xsl:template>

    <msxsl:script language="JScript" implements-prefix="user">
    //<![CDATA[

    function fnSetLinks( strText )
    {
        // embedded hrefs
        strText = strText.replace( /(http:\/\/\S*)/g, "<a href=\"$1\" class=\"aLTLink\" target=\"_blank\">$1</a>" );
        // twitter ids
        strText = strText.replace( /@(\S*)/g, "@<a href=\"http://www.twitter.com/$1\" class=\"aLTUser\" target=\"_blank\">$1</a>" );
        // hash tags
        strText = strText.replace( /(#\S*)/g, "<a href=\"http://search.twitter.com/search?q=$1\" class=\"aLTHash\" target=\"_blank\">$1</a>" );
        return strText;
    }

    function fnNormalizeDate(strDate) {
        try {
            // "Mon Mar 16 18:16:59 +0000 2009" to "2009.03.16 18:16": 
            var rxMatch = /.{3} (.{3}) (\d{2}) (\d{2}:\d{2}):\d{2} \+\d{4} (\d{4})/;
            strDate = strDate.replace( rxMatch , 
                function($0, $1, $2, $3, $4)
                {
                        var strMonth = $1;
                        var strDay = $2;
                        var strYear = $4;
                        var strTime = $3;
                        var dat = new Date( strMonth + " " + strDay + " " + strYear + " " + strTime );
                        // months number is zero based ?!?  wtf
                        strMonth = (dat.getUTCMonth()+1).toString().replace( /^(\d)$/, "0$1" );
                        return strYear + "." + strMonth + "." + strDay + " " + strTime; 
                });
        } catch(e) {
            strDate = e.message;
        }
        return strDate ;
    }

    //]]>
    </msxsl:script>

</xsl:stylesheet>

i think i can live with that for a while. the results will be evident later.

tech notes: whistler's son

fractalnavel — Mon, 16 Mar 2009 22:23:14 GMT

scratch that last cvrap, replace the GetAndPostJob class with the following, making the whole thing a console app to run from the task scheduler:

    class CMain
    {
        [STAThread]
        static void Main(string[] args)
        {
            //do work
            XmlDocument xmldocSettings = new XmlDocument();
            xmldocSettings.Load( args[0]);
            GetAndPost gap = new GetAndPost(xmldocSettings.DocumentElement);
            if ( gap.Get() ) {
                gap.Transform();
                gap.Post();
            } else if ( xmldocSettings.Attributes["bPostNone"].Value == "true" ) {
                gap.Post( "nothin'" );
            }
        }
    }

    [XmlRpcUrl("http://blogs.no-ip.org/fractalnavel/metablog.ashx")] 
    public interface IMetaWebLogNewPost
    {
        [XmlRpcMethod("metaWeblog.newPost")]
        string newPost(
            string blogid,
            string username,
            string password,
            MetaWeblog.Post post,
            bool publish);
    }

and replace the last couple of lines in the Post method:

    
            IMetaWebLogNewPost mwlProxy = XmlRpcProxyGen.Create<IMetaWebLogNewPost>();
            string ret = mwlProxy.newPost( blogid, userid, pwd, p, bPublish);

there's really no good way to get an "at" functionality from cs tasks (unless there's something undocumented; can't tell from the obfuscated code). i mean shoot, i was going to end up recreating the whole scheduler business, may as well just use the thing.

duh.

yeah, all the rest of the stuff needs cleaning, of course, but this will at least do what i want, when i want.

oh yeah - the argument to the app is the name of a file that contains the task node that used to be in communityserver.config, minus the community server specific stuff. i suppose i could use an actual config file; this was a straight forward adaptation of the previous approach.

whatever. just getting things done. i'm not in a very ocd mood. that's what happens with (b) fueled code, instead of (c).

tech notes: whistling in the dark

fractalnavel — Mon, 16 Mar 2009 17:50:44 GMT

been playing with auto-capturing daily twittering & posting here. not that i use twitter much; pehaps this auto-posting lack is why. at first i tried loudtwitter, but community server doesn't have a way of accepting what it provides out of the box - email (needs cs enterprise license), atom, xmlrpc (what api?). so i had it email to tumblr, and then mirrored the feed from there (needs cs pro license). you may have seen the recent example of that. either i did something grossly wrong somewhere, or either loudtwitter or tumblr is eating the markup. and the post naming would cause problems with my current url rewriting scheme - i just noticed that. huh - no wonder it didn't post last night.

none of this was very acceptable (or even workable). i suppose i could have figured out what xmlrpc was being sent from loudtwitter and then used that (a more portable solution - still wopuld have the bad post naming). funny that there's nothing about anyone doing this anyplace that i can see. just me, i guess. and no, still no a.p.p. in cs2008.5sp1 (i upgraded last week).

apparently, i chose the "roll your own" route (see below), but it's not finished. well, it works, but i need to tweak scheduling and formatting, and generalizing the configuration would be nice, as would fixing the credential security issue. but hey, this is just for me. all that detail stuff is a lot of work.

not a sterling example of code or design, but, well, any landing you can walk away from ...

fractalnavel.CS.Components.GetAndPostJob

/*
get xml, transform to post, post to blog.

***twitter specific at the moment.***

config is as follows for now, until controlpanel stuff is done:
    hrefget - get: string (must return results as xml; basically assuming a rest interface)
    useridGet, pwdget - get: authentication = userid, password (yeah, not very secure, need to do something different)
    hrefXsl - transform reference (url; this xslt is applied to the "get" xml, and will result in the post body)
    idBlog - post: destination blog
    useridBlog, pwdBlog - post: authentication (userid, password) (again, security issue)
    titleBlog - post title (appended with date)
    hours - how far back to look
    bPostNone - post when nothing was retrieved ?

also, should log processing status to events...
*/

using System;
using System.IO;
using System.Net;
using System.Text;
using System.Xml;
using System.Xml.Xsl;

using CommunityServer.Blogs.Components;

using Telligent.Tasks;


namespace fractalnavel.CS.Components
{

    // the job
    public class GetAndPostJob : ITask
    {

        public GetAndPostJob()
        {
        }

        public void Execute(XmlNode node)
        {
            //do work
            GetAndPost gap = new GetAndPost(node);
            if ( gap.Get() ) {
                gap.Transform();
                gap.Post();
            } else if ( node.Attributes["bPostNone"].Value == "true" ) {
                gap.Post( "nothin'" );
            }
        }
    }

    // the work
    public class GetAndPost
    {
        private XmlNode _nodSettings;
        private string _strPostBody;
        private XmlDocument _xmldocAll = new XmlDocument();
        
        public GetAndPost( XmlNode node )
        {
            _nodSettings = node;
        }
        
        public bool Get() 
        {
            string href = _nodSettings.Attributes["hrefGet"].Value;
            int intHours = Int32.Parse(_nodSettings.Attributes["hours"].Value);
            // Tue%2C+27+Mar+2007+22%3A55%3A48+GMT
            href += "?since=" + ((DateTime.Now).AddHours(-intHours)).ToString("r");
            string userid = _nodSettings.Attributes["useridGet"].Value;
            string pwd = _nodSettings.Attributes["pwdGet"].Value;
            _strPostBody = GetTextFromHref( href, userid, pwd );
            _xmldocAll.LoadXml( _strPostBody );
            return (_xmldocAll.SelectNodes("//status")).Count > 0;
        }
        
        public void Transform()
        {
            string href = _nodSettings.Attributes["hrefXsl"].Value;
            XslTransform xsl = new XslTransform();
            xsl.Load( href, new XmlUrlResolver() );
            StringBuilder sbResult = new StringBuilder();
            StringWriter swResult = new StringWriter( sbResult );
            xsl.Transform( _xmldocAll, null, swResult, null );
            _strPostBody = sbResult.ToString();
        }
        
        public void Post()
        {
            Post( _strPostBody );
        }
        
        public void Post( string strPostBody )
        {
            string blogid = _nodSettings.Attributes["idBlog"].Value;
            string userid = _nodSettings.Attributes["useridBlog"].Value;
            string pwd = _nodSettings.Attributes["pwdBlog"].Value;
            string title = _nodSettings.Attributes["titleBlog"].Value;
            bool bPublish = true;
            DateTime dtNow = DateTime.Now;
            MetaWeblog.Post p = new MetaWeblog.Post();
            p.title = title + " " + (dtNow).ToString("yyyy.MM.dd");
            p.dateCreated = dtNow;
            p.description = strPostBody;
            IMetaWeblog metablog = new MetaWeblog();
            metablog.newPost(blogid, userid, pwd, p, bPublish);
        }

        private string GetTextFromHref(string url, string user, string password)
        {
            HttpWebRequest request = (HttpWebRequest)HttpWebRequest.Create(url);
            request.Credentials = new NetworkCredential(user, password);
            WebResponse response = request.GetResponse();
            StreamReader reader = new StreamReader(response.GetResponseStream());
            string responseString = reader.ReadToEnd();
            reader.Close();
            return responseString;
        }
    }

}

excerpt from communityserver.config

...
<!-- 2009.03.15, css: added for daily twitter summary post; daily update relies on daily app pool cycling !! -->
<Thread minutes="1320">
    <task 
        name = "fnGetAndPostJob" 
        type = "fractalnavel.CS.Components.GetAndPostJob, fractalnavel.CS" 
        enabled = "true" 
        enableShutDown = "false" 
        hrefGet="http://twitter.com/statuses/user_timeline.xml" 
        useridGet="***" 
        pwdGet="***" 
        hrefXsl="http://blogs.no-ip.org/GetAndPost.xsl" 
        idBlog="fractalnavel" 
        useridBlog="***" 
        pwdBlog="***" 
        titleBlog="a day in a life" 
        hours="24" 
        bPostNone="false" 
        />
</Thread>
...

GetAndPost.xsl

<?xml version="1.0" ?>
<xsl:stylesheet version="1.0" xmlns:xsl="http://www.w3.org/1999/XSL/Transform" >
    
<xsl:output method='html' 
            omit-xml-declaration="yes"
            version="1.0"
            encoding="UTF-8"
            indent="yes"
            cdata-section-elements=""
            /> 
    
    <xsl:template match="/statuses">

        <div class="divLTAll">

            <div class="divLTHead">
                via <a href="http://www.twitter.com/fractalnavel" target="_blank" 
                title="twitter!">@fractalnavel</a>: in the last twenty-four hours:
            </div>
        
            <div class="divLTBody">
                <ul>
                    <xsl:apply-templates select="status" >
                        <xsl:sort select="position()" order="descending" data-type="number"/>
                    </xsl:apply-templates>
                </ul>
            </div>

            <div class="divLTFoot">
                (pulled direct from twitter via custom cs job)
            </div>
            
        </div>

    </xsl:template>

    <xsl:template match="status">
        <li><xsl:value-of select="created_at" />: <xsl:value-of select="text" /></li>
    </xsl:template>

</xsl:stylesheet>

so the intent should be clear anyway.

there were a lot of links that assisted in one way or another, but the power went phlooey this morning for no apparent reason (interestingly, my modem/router was presciently bombing just minutes before that), so that combined with my laziness means you won't see them here. mostly just stuff on generic community server task creation, and the twitter api.

other twitter related stuff to-do:

do an updated CSModule for updating twitter when a post is created. there is an old one out there. somewhere. but my data stream is heading the other direction.
do a sidebar widget with ajaxy & configurable timeline display. have it update & scroll, that sort of thing. eh, for the addicts out there, sure, but for me ? nahhh...

i did finally do something twitter-ish on the xo: installed TTYtter. too bad there's not a sugar-ized gui client. and it seems any other linux friendly gui twitter clients will bloat too much once all the supporting stuff gets put in place. best to do that on an xo running linux on a stick. seriously, what's the issue with expanding the storage there ? for just a few bucks more, could increase it by an order of magnitude. but it's easy being an ~~armchair~~ couch quarterback.

technical difficulties

fractalnavel — Wed, 03 Dec 2008 19:57:32 GMT

apparently craig's s.o. decided december, holidays, a recession and major global economic difficulties were a prime indicator to put an addition onto the house. this is where our server also resides. and wouldn't you know it, murphy also put in an appearance, directing the builders to cut the internet / tv cable.

this was monday, 12/1. apparently the cable company told them they were on a "10 day schedule" to get the cable fixed. so craig went out and spliced in a temporary fix this morning. yay for personal initiative ! boo for corporate america !

doesn't take long to lose what little subscribers we have, and to lose various other connections.

so hang on, there may be some more down time coming soon.

impending hermit mode

fractalnavel — Tue, 11 Nov 2008 19:11:39 GMT

i swear i wasn't going to go here this time, but it looks like it's one of those "best laid plans..." things.

been messing with this site again. added a first shot at video previews to the media gallery. that's not quite working the way i want, and it's missing flexibility as well, but hey, a start is a start. and i can always change the code myself.

what else... oh yeah, community server 2008.5 sp1 is out now, and i'm thinking of applying that. will have to be careful not to bump all these other customizations. it has a big list of fixes and enhancements.

i keep trying to tell myself "do the last yard work of the season" - but it never happens. excuses, excuses. rain to come, wet leaves, long grass. here i sit.

haven't felt like working out or shaving the last couple of days. what's that all about ?

demotivated. as always. tech is to avoid facing other "necessities". i deny their essential necessariness. (huh?)

need new tires. not really a good idea to drive up into snow / ice areas without better treads than i have. i will get to that - when ? can't order / install tires online. *sigh*

some interesting browsing lately, of which only some is found on any of my bookmark lists. best to track my browsing feed. no history to that, unless you've been keeping up, or if you're me and have access to the source. yeah, it's boring twaddle (twiddle?) 95% of the time, but then there are days... unfortunately, i do a lot of browsing that never reaches my google history, being from machines that don't have the google toolbar installed. did a good bit of rabbit-holing over at wikipedia the other day. something along the lines of historiography would be a start. there, it's on the list now too.

* * * * *

part of the trouble seems to be that brief two-day return to work last week. i need to put that shit out of my brain, but it always takes a week or two. was getting a late dinner at the local dive, and i was barely sitting down before someone was in my ear about a local event downtown over the weekend, which was run by the customer i need a break from. they have been exceeding themselves lately in ways i never imagined - and not positively.

well, it was nice outside today, but ...

fractalnavel — Sun, 02 Nov 2008 21:46:58 GMT

i'm, like, a world class slacker. absolutely gorgeous fall weekend, colors at their peak, and really, i should have been hiking & doing yard work & ...

but no, not me. i sit inside and tweak on the blog more. just some niggling stuff:

i want any brief post excerpts i create to appear as post subtitles. that's because the post titles tend to be playful, while i actually try and be more or less plainly descriptive in the excerpts.
post lists have had a date-time heading for each post, whereas there should only be a date heading for all the posts on that day.

so i monkeyed around with those until i got all the details straight. yeah, big deal, i know...

so the same customer that i've been ~~playing~~, er, working at the last few years contacted me last week, hopefully just some minor short-term stuff. unless they really wanted to start doing all the good shit they should be doing. not likely. there's a new manager there, i think this is more of his way of getting to know what resources he has available to him. eh, whatever. i've had october off, and this feeds right into my avoidance syndrome, so what the hey. i find out more tomorrow.

larosa's delivery is on the way.

at least i did one good thing for self last week - zero beer intake. yeah, that'll throw some of you, but i was doing the neighborhood pub thing monday through friday, and then stocking up on six-packs for the weekends. not good. especially since i've been trying to get in regular workouts for the last month, shed some of this pudge, add some shape, whatever. those beer etc. calories are the real problem. poor self. he's, uh - beside himself. that makes two of us ;-)

how many is that now ?

tech notes: what, again ?

fractalnavel — Wed, 29 Oct 2008 01:50:20 GMT

so i’m yakking with craig yesterday, and he mentions that stuff on this blog is broken in firefox – version 3. i’m like, cvrap, it was working in v2…

well, took a tweak here & there over the last couple of days, and some leftover config of community server that i hadn’t done, and things are working again. even works on my xo’s browser now, so i’m guessing it’ll work on other stuff where it wasn’t before too. not sure if it still works in ff2 now, but hey, i don’t care – upgrade :-p

mostly had to do with adapting to more recent changes in the mozilla / gecko dom, javascript, xslt & other components. nah, you don’t really want the details…^[1]

and apparently feedburner was not (any longer?) accepting server-side xmlhttp requests without a user agent set, so i took care of that too. kind of strange, though – they were returning a 404, where maybe a 403 of some sort would have been more appropriate.

^[1] silly stuff:

javascript “eval” changed – can’t eval a statement fragment
xml doc “evaluate” changed – must use the xml doc being evaluated against
error object changes; error message changes (better, but different)
cs config: needed to prevent web.config inheritance for my cross-domain feed proxy
xslt change – doesn’t just blow off msxsl:script anymore
xslt even wraps <xsl:output method=”html”/> results in a “transformiix” node if there isn’t a single root node in the result – needed some changes for this to allow proper “refresh” with my transform behavior

i don’t even think that’s a complete list. good thing i had already built-in a lot of tracing / error handling when i went through this cross-browser stuff last winter.

tech notes: …dink dink dink…

fractalnavel — Sat, 25 Oct 2008 18:48:08 GMT

still getting long-outstanding things done here:

finally fixed the dates on the images in the gallery. they have been screwed up since the original migration to community server two years ago
fixed search operation from blog pages (was only working from aggregate & other pages). gotta be careful with the order of things in siteurls.config when placing blogs at the site root
fixed a strange bug where if a blog post name or number is referenced and doesn’t exist, a page with all the comments from that blog is returned – !

that last one is really a problem with the <CSBlog:WeblogFeedbackList /> control, but never mind that, i don’t even want the post page showing up at all – why isn’t a 404 generated ?

so i created post404.ascx in the blog theme folder:

<%@ Control Language="C#" %>
<%@ Import Namespace="CommunityServer.Controls" %>
<%@ Import Namespace="CommunityServer.Components" %>

<script runat="server">
protected override void OnInit(EventArgs e)
{
    base.OnInit(e);
    Post post = CSContext.Current.GetCurrent<Post>();
    if ( post == null || ( post.PostID == 0 && String.IsNullOrEmpty(post.Name) ) )
        throw new System.Web.HttpException( 404, "" );
}
</script>

and added the following

<%@ Register TagPrefix="CSUserControl" TagName="Post404Check" Src="Post404.ascx" %>
...
<CSUserControl:Post404Check runat="server" />

in the blog theme’s post.aspx page. that last part is at the beginning of the first asp:Content tag’s contents.

what i don’t like is that the 404 page comes up with aspxerrorpath set to the rewritten path of the page, rather than the external form. couldn’t find anything on changing that while still preserving the 404 error code.

tech notes: anything to avoid having to do “real” work

fractalnavel — Fri, 24 Oct 2008 18:05:20 GMT

i lied.

sort of.

i dinked around with a bunch of different community server stuff this past week:

separated the home page & “dashboard” views
added identicon handling for the media gallery, and fixed a couple glitches
added detailed meta tag description handling
fixed an issue with del.icio.us auto-posting
made a number of changes to siteurls.config so that old style post links would still work
monkeyed around with comment spam filter settings

oh, there’s probably a few other things.

although we do accept anonymous comments, please log in first, it keeps your stuff associated with you, and let’s you play with other fun features.

tech notes: wrapping up the blog update thing

fractalnavel — Fri, 17 Oct 2008 20:11:22 GMT

ok:

upgraded to community server 2008.5, including keeping all the add-ons working;
added identicon handling for anonymous & non-avatar weblog comments – see CarKnee.CS.Bundle v4.5;
fixed the del.icio.us auto-posting – see del.icio.us.ashx for community server;

disclaimer on those code mods: they were just quick & dirty “fixes”, but i posted them anyway just in case someone may find them useful. i couldn’t immediately find similar work out there, so why not.

that should be it for this site for the moment. the fun part is now just getting to know & start using all these shiny new toys.

tech notes: more to fix

fractalnavel — Wed, 08 Oct 2008 20:06:21 GMT

job_name out_username out_password out_url out_blog_id out_cat_id

posttofractalnavel (encrypted) (encrypted) http://blogs.no-ip.org/blogs/del.icio.us.ashx fractalnavel 1040

Next run at:
[10/09/08 11:00:00 AM -0700]
Last run notes:
[10/08/08 11:00:02 AM -0700] Setting category '1040' for blog post '4597' ...ERROR: Failed Category Post due to XML RPC2 Exception: 0 : request contains string value where integer expected [request : parameter 1]

job_name	out_username	out_password	out_url	out_blog_id	out_cat_id
posttofractalnavel	(encrypted)	(encrypted)	http://blogs.no-ip.org/blogs/del.icio.us.ashx	fractalnavel	1040
Next run at: [10/09/08 11:00:00 AM -0700] Last run notes: [10/08/08 11:00:02 AM -0700] Setting category '1040' for blog post '4597' ...ERROR: Failed Category Post due to XML RPC2 Exception: 0 : request contains string value where integer expected [request : parameter 1]

i seem to remember that i had to change this before too. wtf ? oh well, add it to the list *sigh*

metablog: it's nice to share

fractalnavel — Thu, 13 Mar 2008 02:08:08 GMT

added ScottWater.CS.Modules ShareIt module to this site. interestingly, this two year old piece of code written for an older community server version (2.1) works just fine as is in v2007.1.

csmodule config as follows:

  <!--
  2008.03.12, css: added for v2.1 of shareit csmodule - will it work ?
  note that there's a bug in the link element's title attribute handling.
  -->
  <add
  name="ShareIt" 
  type ="ScottWater.CS.Modules.ShareItModule, ScottWater.CS.Modules" 
  web="true" syndicate="false" text="">

    <link 
    text='&lt;img src="/images/shareit/delicious.gif"
    	height="16" border="0"
    	style="vertical-align: -20%;" /&gt;' 
    url="http://del.icio.us/post?url=[url][ ^ ];title=[title]" 
    title="Submit [title] to del.icio.us" 
    />

    <link 
    text='&lt;img src="/images/shareit/digg.gif"
    	height="16" border="0"
    	style="vertical-align: -20%;" /&gt;' 
    url="http://www.digg.com/submit?url=[url][ ^ ];phase=2" 
    title="Submit [title] to digg.com" 
    />

    <!-- AddThis Button BEGIN -->
    <!-- 2008.03.12, css: note hack of url attribute value
    	 in order to get event handlers inline. -->
    <link 
    text='&lt;img src="http://s9.addthis.com/button1-share.gif"
    	width="125" height="16" border="0" alt=""
    	style="vertical-align: -20%;" /&gt;' 
    url='http://www.addthis.com/bookmark.php"
    	onMouseOver="return addthis_open(this, &#39;&#39;, &#39;[url]&#39;, &#39;[title]&#39;)"
    	onMouseOut="addthis_close()"
    	onClick="return addthis_sendto()' 
    title="" 
    />
    <!-- AddThis Button END -->

  </add>

notes:

yeah, there is a bug in that code that prevents a link element's title attribute value from being used:
```
...
  att = addNode.Attributes["title"];
  string addTitle = null;
  if (att == null)
  {
      addTitle = att.Value.Replace("[title]","{1}");
  }
  else
  {
      addTitle = "Post {2}";
  }
...
```
that condition should be "!=".
for the "addthis" link, managed to construct the url value so that the module spits out an anchor with event handlers.

there were a couple of other minor changes i made to this module for a client: removed the hard-coded use of <strong/> tags around the text; and added link item separator configuration. and i also use images and not text for the links as well.

and oh yeah: specific settings like the account etc. go in script tags in the "site header"; while the addthis script reference goes into the "copyright" setting (like the google analytics call), since there's no "site footer".

yeah, ok, for some of you this is just soo exciting. hey, but i like the results. just ignore this post, and use the things.

metablog: sunrise, sunset...

fractalnavel — Wed, 12 Mar 2008 15:27:33 GMT

fiddler on the roof ? nah. i just put the world sunrise map image in the sidebar. i always like following the equinox there (about a week from today), seeing the straight up and down terminator.

world sunlight map at die.net

of course, that's only relative to the projection being parallel to the axis of rotation, but what the hey, anthropic biases be damned, it's still fun to watch.

ok, i admit, i'm oddly - if not necessarily easily - entertained.