index.htm

<html manifest="ac.appcache"><head>
<!-- title must be first, IE 6 (test) and under, put title tag first in -->
<!-- childNodes/children, regardless of HTML src code tree order, if title -->
<!-- not first absolute pos into .children will fail -->
<title>tinymta - Lightweight NYC MTA Subway Train arrival times</title>
<!-- put no-referrer so preload reqs match later fetch reqs instead of double req -->
<meta name=mobileoptimized content=0><meta name="referrer" content="no-referrer">
<!-- preconnect Chrome 46/FF41-70 115-, preconnect Chrome 50/FF 85, inc both -->
<link href="https://api.weather.com/v3/wx/forecast/hourly/1day?geocode=40.74,-73.91&format=json&units=e&language=en-us&apiKey=8de2d8b3a93542c9a2d8b3a935a2c909" as="fetch" rel="preload" crossorigin>
<!-- dont do preconnect fallback for weather, not worth wire bytes -->
<!-- 1062 vs 1059 bytes gz local af rm weather PC tag -->
<!-- 1059 vs 1047 bytes gz local af rm 2 dns-fetch tags -->
<!-- as.js is a CFW, so higher latency vs static files on CF, put these 2 .js -->
<!-- on wire asap, for parallel download and parse/latency by modern UAs with -->
<!-- async, f.js has work arounds for as.js no-async early failed paint/no DOM -->
<!-- body yet problems, IMO as.js is less important than getting "new socket"  -->
<!-- cross origin weather JSON preload on wire, since tinymta socket 25%-99% chance -->
<!-- already started -->
<link href="//tinymta.us.to/as.js" rel=preload as=script>
<script async defer src="//tinymta.us.to/as.js"></script>
<script async defer src="//tinymta.us.to/jsp?url=https%3A%2F%2Fapi.weather.com%2Fv3%2Fwx%2Fforecast%2Fhourly%2F1day%3Fgeocode%3D40.74%2C-73.91%26format%3Djson%26units%3De%26language%3Den-us%26apiKey%3D8de2d8b3a93542c9a2d8b3a935a2c909&callback=j1716324619352"></script>
<script async defer src=dumb.js></script>
<script async defer src=fav.js></script>
<!-- preconnect to status.htm's API -->
<link href="//collector-otp-prod.camsys-apps.com" rel="preconnect" crossorigin>
<!-- not time sensitive, all below are navigates or unimplimented by SF -->
<link href=stations.htm rel=prefetch>
<link href=mn/stations.htm rel=prefetch>
<link href=li/stations.htm rel=prefetch>
<link href=status.htm rel=prefetch>
<!-- draw_fav.js is dev only -->
<!-- <script src=draw_fav.js></script> -->


<style>:root{color-scheme:light dark}span{display:inline-block}img{height:2em}</style>
</head><body>
Do you have a<br>
[1]<a accesskey=1 href=stations.htm>Smart Phone (Subway)</a><br>
[2]<a accesskey=2 href=mn/stations.htm>Smart Phone (Metro North)</a><br>
[3]<a accesskey=3 href=li/stations.htm>Smart Phone (Long Island)</a><br>
[4]<a accesskey=4 href=jsrdt.htm>Dumb Phone (Subway)</a><br>
[5]<a accesskey=5 href=mn/jsrdt.htm>Dumb Phone (Metro North)</a><br>
[6]<a accesskey=6 href=li/jsrdt.htm>Dumb Phone (Long Island)</a><br>
[7]<a accesskey=7 href=status.htm>Service Changes</a><br>
<!-- [8]<a accesskey=8 href=plan.htm>Smart Phone UI (Subway) Trip Planner</a><br> -->
<!-- [9]<a accesskey=9 href=more.htm>More</a><br> -->
[0]<a accesskey=0 href=abt.htm>About</a>
<!-- div used by as.js-->
<div></div>
<script>
//1029->1061 LS "f", 1062 for LS "fav"
//localStorage.setItem("fav", "document.body.appendChild(document.createElement('div')).appendChild(document.createTextNode('FFF'))");
//localStorage.removeItem("fav");
//debugger;
//var CT = performance.now();
try{eval(localStorage.getItem("fav"))} catch (err) {}
//console.log('fav ld '+(performance.now()-CT));

var L;//async race CB fav.js vs inline wea script
var R;
//gz 1021 "document.", gz 1019 "doc." everywhere
var doc = document;
//IE 5.5 and 6.0 don't have document.head
var headElem = doc.documentElement.firstChild;
function y (nosvg, mkWeatherCellContainerPF) {
  fetch(headElem.children[3].href)
  .then(function (r) {
  r.status != 200 ? //put != first for r.status gzip reasons
    alert(r.status)
  :
    r.json().then(function (r) {
/*
was a polyfill func or real doc.CDF, but it was switched to "polyfill always"
DIV tag, to keep this minified/gziped htm file
inside exactly 1 http 1.1 resp packet (1500 bytes)
*/
      var frag = doc.createElement('div'),
      e, //elem
      v, //val
      img_ext = nosvg ? nosvg : '.svg',
      br = doc.createElement('br'),
      drop = this.E || doc.createTextNode("\ud83d\udca7"),
      i = 0;
      /*hardcode 24 hr in a day to save wire*/
      for (; i < 24; i++) {
/*no inline-block for IE 5.0, inline-block 5.5 up.
  A 1 cell table, is a old PF hack inline-block emulate for IE 5.0
  http://mrclay.org/web_design/shrink_to_fit/
  http://mrclay.org/web_design/shrink_to_fit/ieshrink.js
  I fixed a bug in code above "<table><tr><td>" only works for
  .innerHTML/real HTML parser in IE, <tbody> required if appendChild/W3C DOM
  for IE 5 5.5 and 6, in my testing
  see https://blog.tomayac.com/2006/06/23/dynamically-created-tables-in-internet-explorer-214331/
*/
        if (mkWeatherCellContainerPF) {
          e = mkWeatherCellContainerPF(frag);
        } else {
/*
originally a DIV tag, switched to SPAN for IE 5.5/6 compat.
So IE 11 and all FF/WBK/Chrome know style=display=inline-block on a DIV tag
but IE 5.5/6 refuse to honor inline-block on a DIV tag, switch to SPAN
*/
          e = frag.appendChild(doc.createElement('span'));
        }
        v = +r.validTimeLocal[i].slice(11, 13);
        //perf, skip locale/Date objects
        e.appendChild(doc.createTextNode((v % 12 || 12) + (v >= 12 ? 'PM\xa0' : 'AM\xa0')));
/*
Only opera 10 fatal excpt err if missing arg. All IEs don't err and no modern
Safari/WBK/Chrome/FF needs arg. But MDN archives say W3C fuzzyness and spec
flipping between FF 1-3?? and W3C on meaning of no arg, prob caused Opera devs
to mandate the arg 100% of time, false is "clone shallow" true is "clone deep"
meaningless for self closing BR tag
*/
        e.appendChild(br.cloneNode(0));
        e.appendChild(doc.createTextNode(r.temperature[i] + "F"));
        e.appendChild(br.cloneNode(0));
        e.appendChild(doc.createTextNode(r.precipChance[i] + "%"));
//avg of ~50 pg loads, 1.X ms is cumulative (24 calls summed per pg load)
//CE() is plus .src + setAttribute()
//FF115 CE() 1.4670533846736797
//      CN() 1.655915783189822
//C109  CE() 2.156961555116267
//      CN() 1.6185931681633636
//optimize for Chrome since Chrome more popular
        e.appendChild(drop.cloneNode(0));
        e.appendChild(br.cloneNode(0));
        e.appendChild(doc.createElement('img'))
        .src =
/*STARTINSERTPATH*/
          'w/'
/*ENDINSERTPATH*/
          +r.iconCode[i] + img_ext;
      }
      doc.body.appendChild(frag);
    }); //json cb
  }); //end fetch CB
};
if (!this.f) {
  if (this.fetch) {
    y();
  } else {
  //dont double load in dumb.js, async flag or no async flag
    f=1;
    headElem.appendChild(doc.createElement("script")).src = 'f.js';
  }
}
//run DOM load CB from fav.js if needed (chains to as.js CB if needed)
L=L&&L();
</script>
</body></html>