Newer
Older
1001
1002
1003
1004
1005
1006
1007
1008
1009
1010
1011
1012
1013
1014
1015
1016
1017
1018
1019
1020
1021
1022
1023
1024
1025
1026
1027
int u8bytes=0; while(u16[u8bytes++]); u8bytes*=3;
p=(POSTYPE)malloc(++u8bytes);
pOrig=p;
do {
if(!(*u16&~0x7f)) *p++=*u16;
else {
if(!(*u16&~0x7ff)) {
*p++=0xC0|((*u16)>>6);
} else {
*p++=0xE0|(((*u16)>>12)&15);
*p++=0x80|(((*u16)>>6)&0x3F);
}
*p++=0x80|((*u16)&0x3F);
}
} while(*u16++);
GlobalUnlock(hClipMemory);
CloseClipboard();
char fname[MAX_PATH];
#ifndef _WINCE
GetTempPathA(sizeof(fname) - 7, fname);
strcat(fname,"c.html"); /* c for clipboard */
outFile = fopen(fname,"w");
#endif
if (!outFile) {
strcpy(fname,"\\c.html");
outFile=fopen(fname,"w");
}
OutWriteStr("<html><head><meta http-equiv=\"Content-Type\" content=\"text/html; charset=utf-8\"><meta name=\"mobileoptimized\" content=\"0\"><meta name=\"viewport\" content=\"width=device-width\"></head><body><style id=\"ruby\">ruby { display: inline-table; vertical-align: bottom; -webkit-border-vertical-spacing: 1px; padding-top: 0.5ex; } ruby * { display: inline; vertical-align: top; line-height:1.0; text-indent:0; text-align:center; white-space: nowrap; } rb { display: table-row-group; font-size: 100%; } rt { display: table-header-group; font-size: 100%; line-height: 1.1; }</style>\n<!--[if !IE]>-->\n<style>rt { font-family: Gandhari, DejaVu Sans, Lucida Sans Unicode, Times New Roman, serif !important; }</style>\n<!--<![endif]-->\n");
p=pOrig; copyP=p;
matchAll();
free(pOrig);
OutWriteStr("<script><!--\nfunction treewalk(n) { var c=n.firstChild; while(c) { if (c.nodeType==1 && c.nodeName!=\"SCRIPT\" && c.nodeName!=\"TEXTAREA\" && !(c.nodeName==\"A\" && c.href)) { treewalk(c); if(c.nodeName==\"RUBY\" && c.title && !c.onclick) c.onclick=Function(\"alert(this.title)\") } c=c.nextSibling; } } function tw() { treewalk(document.body); window.setTimeout(tw,5000); } treewalk(document.body); window.setTimeout(tw,1500);\n//--></script></body></html>");
fclose(outFile);
TCHAR fn2[sizeof(fname)]; int i;
for(i=0; fname[i]; i++) fn2[i]=fname[i]; fn2[i]=(TCHAR)0;
SHELLEXECUTEINFO sei;
memset(&sei, 0, sizeof(sei));
sei.cbSize = sizeof(sei);
sei.lpVerb = TEXT("open");
sei.lpFile = fn2;
sei.nShow = SW_SHOWNORMAL;
if (!ShellExecuteEx(&sei)) errorExit("ShellExecuteEx");
// TODO: sleep(); remove{fname); ?
// (although it will probably be the same on each run)
DestroyWindow(win); // TODO: needed?
}
"""
else: c_end += r"""
Silas S. Brown
committed
#ifndef Omit_main
Silas S. Brown
committed
int main(int argc,char*argv[]) {
Silas S. Brown
committed
int i; for(i=1; i<argc; i++) {
if(!strcmp(argv[i],"--help")) {
puts("Use --ruby to output ruby markup (default)");
puts("Use --raw to output just the annotations without the base text");
puts("Use --braces to output as {base-text|annotation}");
return 0;
} else if(!strcmp(argv[i],"--ruby")) {
annotation_mode = ruby_markup;
} else if(!strcmp(argv[i],"--raw")) {
annotation_mode = annotations_only;
} else if(!strcmp(argv[i],"--braces")) {
annotation_mode = brace_notation;
} else {
fprintf(stderr,"Unknown argument '%s'\n(Text should be on standard input)\n",argv[i]); return 1;
Silas S. Brown
committed
}
}
matchAll();
}
#endif
"""
# ANDROID: setDefaultTextEncodingName("utf-8") is included as it might be needed if you include file:///android_asset/ URLs in your app (files put into assets/) as well as remote URLs. (If including ONLY file URLs then you don't need to set the INTERNET permission in Manifest, but then you might as well pre-annotate the files and use a straightforward static HTML app like http://people.ds.cam.ac.uk/ssb22/gradint/html2apk.html )
# Also we get shouldOverrideUrlLoading to return true for URLs that end with .apk .pdf .epub .mp3 etc so the phone's normal browser can handle those (search code below for ".apk" for the list)
additional_intents=''.join(('\n<intent-filter><action android:name="android.intent.action.VIEW" /><category android:name="android.intent.category.DEFAULT" /><category android:name="android.intent.category.BROWSABLE" /><data android:scheme="%s" android:host="%s" android:pathPrefix="%s" /></intent-filter>'%(urlparse.urlparse(x).scheme,urlparse.urlparse(x).netloc,urlparse.urlparse(x).path)) for x in os.environ.get("ANNOGEN_ANDROID_URLS","").split())
if additional_intents and "?" in os.environ["ANNOGEN_ANDROID_URLS"]: errExit("Can't include '?' queries in ANNOGEN_ANDROID_URLS (it and anything after it would be ignored)")
Silas S. Brown
committed
/* COMPILING
---------
As I've been unable to make "Android Studio" work
on my equipment (I'm not entirely sure why), all I
can offer are these instructions for compiling on the
older "Android Developer Tools" (ADT), which were
deprecated in June 2015 and the download was removed
in June 2017. But in case you still have them:
1. You might need to increase the amount of RAM it's
allowed to use, e.g. put -Xmx2g into eclipse.ini
(be sure to remove any existing -Xmx settings
otherwise they might override your new setting)
Silas S. Brown
committed
2. Go to File / New / Android application project
3. Application name = anything you want (for the phone's app menu)
Project name = anything you want (unique on your development machine)
Silas S. Brown
committed
Minimum Required SDK = API 1: Android 1.0
Leave everything else as default
but make a note of the project directory
(usually on the second setup screen as "location")
(If you DON'T want the app to run in full screen,
see "Delete the following line if you don't want full screen" below)
5. Edit project.properties and add the line
dex.force.jumbo=true
6. Edit AndroidManifest.xml and make it look as below
(you might need to change targetSdkVersion="19" if
your SDK has a different targetSdkVersion setting,
and if you're creating a new version of a
previously-released app then you might want to
increase the values of android:versionCode and
android:versionName for your new app version)
Silas S. Brown
committed
---------------------- cut here ----------------------
<?xml version="1.0" encoding="utf-8"?>
<manifest xmlns:android="http://schemas.android.com/apk/res/android" package="%%JPACKAGE%%" android:versionCode="1" android:versionName="1.0" >
Silas S. Brown
committed
<uses-permission android:name="android.permission.INTERNET" />
<uses-sdk android:minSdkVersion="1" android:targetSdkVersion="19" />
Silas S. Brown
committed
<application android:icon="@drawable/ic_launcher" android:label="@string/app_name" android:theme="@style/AppTheme" >
<activity android:configChanges="orientation|screenSize|keyboardHidden" android:name="%%JPACKAGE%%.MainActivity" android:label="@string/app_name" android:launchMode="singleInstance" >
Silas S. Brown
committed
<intent-filter><action android:name="android.intent.action.MAIN" /><category android:name="android.intent.category.LAUNCHER" /></intent-filter>
<intent-filter><action android:name="android.intent.action.SEND" /><category android:name="android.intent.category.DEFAULT" /><data android:mimeType="text/plain" /></intent-filter>"""+additional_intents+r"""
Silas S. Brown
committed
</activity></application></manifest>
---------------------- cut here ----------------------
7. Copy new AndroidManifest.xml to the bin/ directory
Silas S. Brown
committed
(so there will be 2 copies, one in the top level
and the other in bin/ )
8. Edit res/layout/activity_main.xml and make it like:
Silas S. Brown
committed
---------------------- cut here ----------------------
<?xml version="1.0" encoding="utf-8"?>
<LinearLayout xmlns:android="http://schemas.android.com/apk/res/android" android:layout_height="fill_parent" android:layout_width="fill_parent" android:orientation="vertical">
<WebView android:id="@+id/browser" android:layout_height="match_parent" android:layout_width="match_parent" />
Silas S. Brown
committed
</LinearLayout>
---------------------- cut here ----------------------
9. Restart ADT, do Run / Run As / Android application
10. Watch ADT's Console window until it says the app
Silas S. Brown
committed
has started, then interact with the Android virtual
device to test. (If install fails, try again.)
11. .apk file should now be in the bin subdirectory.
On a real phone go to "Application settings" or
"Security" and enable "Unknown sources". Or if
you're ready to ship your .apk, select it in
Eclipse's Package Explorer (left-hand pane) and
do File / Export / Export Android Application (it
lets you create a keystore and private signing key)
12. If you ship your app on Play Store, you are advised
to use the "beta test" facility before going live.
Play Store has been known to somehow 'corrupt' APKs
generated by Annogen, for an unknown reason. (The
APK works just fine when run standalone, but fails
to annotate when downloaded from Play Store.) When
this happens, simply incrementing the
version numbers in the AndroidManifest.xml files
and re-uploading to Play Store somehow 'fixes' it.
(Similarly, you might find one version works fine
but the next does not, even if you've only fixed a
'typo' between the versions. Use beta test, and if
it goes wrong then re-upload.)
To copy/paste from the annotated text, make sure to
start the long-press ON a word (not in a space). This
appears to be an Android/Chrome limitation (especially
in version 4, but I haven't been able to test all versions).
You can annotate local HTML files as well as Web pages.
Local HTML is placed in the 'assets' folder and referred
to via --android=file:///android_asset/FILENAME
where FILENAME is the name of your HTML file.
A clipboard viewer is placed in clipboard.html.
You can also set ANNOGEN_ANDROID_URLS to a
whitespace-separated list of URL prefixes to
offer to be a browser for. For example,
ANNOGEN_ANDROID_URLS="http://example.com http://example.org/documents"
Silas S. Brown
committed
*/
Silas S. Brown
committed
import android.webkit.WebView;
import android.webkit.WebChromeClient;
import android.webkit.WebViewClient;
import android.annotation.SuppressLint;
Silas S. Brown
committed
import android.app.Activity;
import android.os.Bundle;
import android.view.KeyEvent;
public class MainActivity extends Activity {
@SuppressLint("SetJavaScriptEnabled")
@android.annotation.TargetApi(19) // 19 for setWebContentsDebuggingEnabled; 7 for setAppCachePath; 3 for setBuiltInZoomControls (but only API 1 is required)
@SuppressWarnings("deprecation") // for conditional SDK below
Silas S. Brown
committed
@Override
public void onCreate(Bundle savedInstanceState) {
super.onCreate(savedInstanceState);
// ---------------------------------------------
// Delete the following line if you DON'T want full screen:
requestWindowFeature(android.view.Window.FEATURE_NO_TITLE); getWindow().addFlags(android.view.WindowManager.LayoutParams.FLAG_FULLSCREEN);
// ---------------------------------------------
Silas S. Brown
committed
setContentView(R.layout.activity_main);
browser = (WebView)findViewById(R.id.browser);
// ---------------------------------------------
// Delete the following long line if you DON'T want caching (Android 2.1+); caching is useful for persistence if app is removed from memory and then switched back to while user is offline
if(Integer.valueOf(android.os.Build.VERSION.SDK) >= 7) { browser.getSettings().setAppCachePath(getApplicationContext().getCacheDir().getAbsolutePath()); browser.getSettings().setAppCacheMaxSize(10*1048576) /* if API==7 i.e. exactly Android 2.1 (deprecated in API 8) */ ; browser.getSettings().setAppCacheEnabled(true); if(Integer.valueOf(android.os.Build.VERSION.SDK)<=19 && savedInstanceState==null) browser.clearCache(true); } // (Android 4.4 has Chrome 33 which has Issue 333804 XMLHttpRequest not revalidating, which breaks some sites, so clear cache when we 'cold start' on 4.4 or below)
// ---------------------------------------------
if(Integer.valueOf(android.os.Build.VERSION.SDK) >= 19) WebView.setWebContentsDebuggingEnabled(true); // so you can use chrome://inspect in desktop Chromium when connected via USB to Android 4.4+
Silas S. Brown
committed
browser.getSettings().setJavaScriptEnabled(true);
browser.setWebChromeClient(new WebChromeClient());
public A(MainActivity act) { this.act = act; }
MainActivity act;
@android.webkit.JavascriptInterface public String annotate(String t,boolean inLink) { String r=new %%JPACKAGE%%.Annotator(t).result(); if(!inLink) r=r.replaceAll("<ruby","<ruby onclick=\"annotPopAll(this)\""); return r; } // now we have a Copy button, it's convenient to put this on ALL ruby elements, not just ones with title
@android.webkit.JavascriptInterface public void alert(String t,String a) {
class DialogTask implements Runnable {
String tt,aa;
DialogTask(String t,String a) { tt=t; aa=a; }
public void run() {
android.app.AlertDialog.Builder d = new android.app.AlertDialog.Builder(act);
d.setTitle(tt); d.setMessage(aa);
d.setNegativeButton("Copy",new android.content.DialogInterface.OnClickListener() {
@android.annotation.TargetApi(11)
public void onClick(android.content.DialogInterface dialog,int id) {
String text=tt+" "+aa;
if(Integer.valueOf(android.os.Build.VERSION.SDK) < android.os.Build.VERSION_CODES.HONEYCOMB) // SDK_INT requires API 4 but this works on API 1
((android.text.ClipboardManager)getSystemService(android.content.Context.CLIPBOARD_SERVICE)).setText(text);
else ((android.content.ClipboardManager)getSystemService(android.content.Context.CLIPBOARD_SERVICE)).setPrimaryClip(android.content.ClipData.newPlainText(text,text));
}
});
d.setPositiveButton("OK", null); // or can just click outside the dialog to clear. (TODO: would be nice if it could pop up somewhere near the word that was touched)
d.create().show();
}
}
act.runOnUiThread(new DialogTask(t,a));
@android.webkit.JavascriptInterface public String getClip() { return readClipboard(); }
@android.webkit.JavascriptInterface public String getSentText() { return sentText; }
Silas S. Brown
committed
}
browser.addJavascriptInterface(new A(this),"ssb_local_annotator"); // hope no conflict with web JS
Silas S. Brown
committed
browser.setWebViewClient(new WebViewClient() {
public boolean shouldOverrideUrlLoading(WebView view,String url) { if(url.endsWith(".apk") || url.endsWith(".pdf") || url.endsWith(".epub") || url.endsWith(".mp3") || url.endsWith(".zip")) { startActivity(new Intent(Intent.ACTION_VIEW,android.net.Uri.parse(url))); return true; } else return false; }
Silas S. Brown
committed
public void onPageFinished(WebView view,String url) {
if(Integer.valueOf(android.os.Build.VERSION.SDK) < 19) // Pre-Android 4.4, so below runTimer() alternative won't work. This version has to wait for the page to load entirely (including all images) before annotating.
browser.loadUrl("javascript:"+js_common+"function AnnotMonitor() { AnnotIfLenChanged();window.setTimeout(AnnotMonitor,1000)} AnnotMonitor()");
Silas S. Brown
committed
else browser.loadUrl("javascript:"+js_common+"AnnotIfLenChanged();var m=window.MutationObserver;if(m)new m(function(mut,obs){if(mut[0].type=='childList'){AnnotIfLenChanged()}}).observe(document.body,{childList:true,subtree:true})"); // (no point waiting the rest of the second for runTimer() to run, especially if this is the initial assets page; also Android 4.4+ has MutationObserver for even faster response to changes, so set that up as well) (and yes we do need to include js_common on this line because we don't know if runTimer has yet happened on this new page)
Silas S. Brown
committed
} });
if(Integer.valueOf(android.os.Build.VERSION.SDK) >= 3) {
browser.getSettings().setBuiltInZoomControls(true);
}
int size=Math.round(16*getResources().getConfiguration().fontScale); // from device accessibility settings
browser.getSettings().setDefaultFontSize(size);
browser.getSettings().setDefaultFixedFontSize(size);
Silas S. Brown
committed
browser.getSettings().setDefaultTextEncodingName("utf-8");
if (savedInstanceState!=null) browser.restoreState(savedInstanceState); else
if (!handleIntent(getIntent())) browser.loadUrl("%%ANDROID-URL%%");
}
@Override
public void onNewIntent(Intent intent) {
super.onNewIntent(intent); handleIntent(intent);
}
if (Intent.ACTION_SEND.equals(intent.getAction()) && "text/plain".equals(intent.getType())) {
sentText = intent.getStringExtra(Intent.EXTRA_TEXT);
if (sentText == null) return false;
browser.loadUrl("javascript:document.close();document.rubyScriptAdded=0;document.write('<html><head><meta name=\"mobileoptimized\" content=\"0\"><meta name=\"viewport\" content=\"width=device-width\"></head><body>'+ssb_local_annotator.getSentText().replace(/&/g,'&').replace(/</g,'<').replace('\\n','<br>')+'</body>')");
else if (Intent.ACTION_VIEW.equals(intent.getAction())) {
browser.loadUrl("javascript:document.close();document.write('<html><head><meta name=\"mobileoptimized\" content=\"0\"><meta name=\"viewport\" content=\"width=device-width\"></head><body>Loading, please wait...</body>')"); // to avoid misunderstanding if there's already a page there and user thinks it's not working
browser.loadUrl(intent.getData().toString());
}
Silas S. Brown
committed
}
static final String js_common="""+'"'+jsAnnot("ssb_local_annotator.alert(f(e.firstChild)+' '+f(e.firstChild.nextSibling),e.title||'')","function AnnotIfLenChanged() { var getLen=function(w) { var r=0; if(w.frames && w.frames.length) { var i; for(i=0; i<w.frames.length; i++) r+=getLen(w.frames[i]) } if(w.document && w.document.body && w.document.body.innerHTML) r+=w.document.body.innerHTML.length; return r },curLen=getLen(window); if(curLen!=window.curLen) { annotScan(); window.curLen=getLen(window) } }","","tw0(); "+jsAddRubyCss,"var nv=ssb_local_annotator.annotate(cnv,inLink); if(nv!=cnv) { var newNode=document.createElement('span'); newNode.className='_adjust0'; n.replaceChild(newNode, c); newNode.innerHTML=nv }")+r"""";
android.os.Handler theTimer;
@SuppressWarnings("deprecation")
@android.annotation.TargetApi(19)
void runTimerLoop() {
if(Integer.valueOf(android.os.Build.VERSION.SDK) >= 19) { // on Android 4.4+ we can do evaluateJavascript while page is still loading (useful for slow-network days) - but setTimeout won't usually work so we need an Android OS timer
theTimer = new android.os.Handler();
theTimer.postDelayed(new Runnable() {
@Override
public void run() {
browser.evaluateJavascript(js_common+"AnnotIfLenChanged()",null);
theTimer.postDelayed(this,1000);
}
},0);
Silas S. Brown
committed
@Override public boolean onKeyDown(int keyCode, KeyEvent event) {
if ((keyCode == KeyEvent.KEYCODE_BACK) &&
browser.canGoBack()) {
browser.goBack(); return true;
} else return super.onKeyDown(keyCode, event);
}
@SuppressWarnings("deprecation") // using getText so works on API 1 (TODO consider adding a version check and the more-modern alternative android.content.ClipData c=((android.content.ClipboardManager)getSystemService(android.content.Context.CLIPBOARD_SERVICE)).getPrimaryClip(); if (c != null && c.getItemCount()>0) return c.getItemAt(0).coerceToText(this).toString(); return ""; )
@android.annotation.TargetApi(11)
if(Integer.valueOf(android.os.Build.VERSION.SDK) < android.os.Build.VERSION_CODES.HONEYCOMB) // SDK_INT requires API 4 but this works on API 1
return ((android.text.ClipboardManager)getSystemService(android.content.Context.CLIPBOARD_SERVICE)).getText().toString();
android.content.ClipData c=((android.content.ClipboardManager)getSystemService(android.content.Context.CLIPBOARD_SERVICE)).getPrimaryClip();
if (c != null && c.getItemCount()>0) {
return c.getItemAt(0).coerceToText(this).toString();
}
return "";
}
@Override protected void onSaveInstanceState(Bundle outState) { browser.saveState(outState); }
Silas S. Brown
committed
}
"""
if ndk: c_start = c_start.replace("%%android_src%%",android_src.replace("Put *.java into src/%%JPACK2%%","Optionally edit this file, but beware it will be overwritten if the script to generate it is re-run").replace('new %%JPACKAGE%%.Annotator(t).result()','jniAnnotate(t)').replace('%%JPACKAGE%%',ndk).replace('public class MainActivity extends Activity {','public class MainActivity extends Activity {\n static { System.loadLibrary("Annotator"); }\n static synchronized native String jniAnnotate(String in);').replace('%%ANDROID-URL%%',android))
android_clipboard = r"""<html><head><meta name="mobileoptimized" content="0"><meta name="viewport" content="width=device-width"><meta http-equiv="Content-Type" content="text/html; charset=utf-8"></head><body>
<script>window.onerror=function(msg,url,line){ssb_local_annotator.alert('Error!',''+msg); return true}</script>
<h3>Clipboard</h3>
<div id="clip">waiting for clipboard contents</div>
<script>
var curClip="";
function update() {
var newClip = ssb_local_annotator.getClip();
if (newClip != curClip) {
document.getElementById('clip').innerHTML = newClip.replace(/&/g,'&').replace(/</g,'<').replace(/\u200b/g,'');
curClip = newClip;
} window.setTimeout(update,1000) } update(); </script>
</body></html>"""
if ndk: c_start = c_start.replace("%%android_clipboard%%",android_clipboard)
java_src = r"""package %%JPACKAGE%%;
public class Annotator {
// use: new Annotator(txt).result()
public Annotator(String txt) { nearbytes=%%YBYTES%%; inBytes=s2b(txt); inPtr=0; writePtr=0; needSpace=false; outBuf=new java.util.ArrayList<Byte>(); }
Silas S. Brown
committed
int nearbytes;
byte[] inBytes;
public int inPtr,writePtr; boolean needSpace;
Silas S. Brown
committed
java.util.List<Byte> outBuf; // TODO improve efficiency (although hopefully this annotator is called for only small strings at a time)
Silas S. Brown
committed
static final byte EOF = (byte)0; // TODO: a bit hacky
Silas S. Brown
committed
if (inPtr==inBytes.length) return EOF;
return inBytes[inPtr++];
}
Silas S. Brown
committed
// for Yarowsky-like matching (use Strings rather than byte arrays or Java compiler can get overloaded)
Silas S. Brown
committed
int offset=inPtr, maxPos=inPtr+nearbytes;
if (maxPos > inBytes.length) maxPos = inBytes.length;
maxPos -= bytes.length;
if(offset>nearbytes) offset-=nearbytes; else offset = 0;
while(offset <= maxPos) {
boolean ok=true;
for(int i=0; i<bytes.length; i++) {
if(bytes[i]!=inBytes[offset+i]) { ok=false; break; }
}
if(ok) return true;
offset++;
}
return false;
}
public void o(String s) { byte[] b=s2b(s); for(int i=0; i<b.length; i++) outBuf.add(b[i]); } // TODO: is there a more efficient way to do it than this?
Silas S. Brown
committed
if (needSpace) o((byte)' ');
else needSpace=true;
}
Silas S. Brown
committed
s();
o("<ruby><rb>");
for(;numBytes>0;numBytes--)
o(inBytes[writePtr++]);
o("</rb><rt>"); o(annot);
o("</rt></ruby>");
}
public void o2(int numBytes,String annot,String title) {
Silas S. Brown
committed
s();
o("<ruby title=\""); o(title);
o("\"><rb>");
for(;numBytes>0;numBytes--)
o(inBytes[writePtr++]);
o("</rb><rt>"); o(annot);
o("</rt></ruby>");
}
// Convert string to bytes - version that works before Android API level 9 i.e. in Java 5 not 6. (Some versions of Android Lint sometimes miss the fact that s.getBytes(UTF8) where UTF8==java.nio.charset.Charset.forName("UTF-8")) won't always work. We could do an API9+ version and use @android.annotation.TargetApi(9) around the class (android.os.Build.VERSION.SDK_INT won't work on API less than 4 but Integer.valueOf(android.os.Build.VERSION.SDK) works), but anyway we'd rather not have to generate a special Android-specific version of Annotator as well as putting Android stuff in a separate class.)
try { return s.getBytes("UTF-8"); }
catch(java.io.UnsupportedEncodingException e) {
// should never happen for UTF-8
return null;
}
}
Silas S. Brown
committed
public String result() {
while(inPtr < inBytes.length) {
int oldPos=inPtr;
Silas S. Brown
committed
if (oldPos==inPtr) { needSpace=false; o(nB()); writePtr++; }
}
byte[] b=new byte[outBuf.size()];
for(int i=0; i<b.length; i++) b[i]=outBuf.get(i); // TODO: is this as efficient as we can get??
try { return new String(b, "UTF-8"); } catch(java.io.UnsupportedEncodingException e) { return null; }
Silas S. Brown
committed
}
}
"""
if os.environ.get("ANNOGEN_CSHARP_NO_MAIN",""):
cSharp_mainNote = ""
else: cSharp_mainNote = r"""
// or just use the Main() at end (compile with csc, and
// see --help for usage)
// (to omit this Main() from the generated file, set
// the environment variable ANNOGEN_CSHARP_NO_MAIN before
// running Annotator Generator)"""
cSharp_start = "// C# code "+version_stamp+r"""
// use: new Annotator(txt).result()
// (can also set annotation_mode on the Annotator)"""+cSharp_mainNote+r"""
enum Annotation_Mode { ruby_markup, annotations_only, brace_notation };
class Annotator {
public const string version="""+'"'+version_stamp+r"""";
1414
1415
1416
1417
1418
1419
1420
1421
1422
1423
1424
1425
1426
1427
1428
1429
1430
1431
1432
1433
1434
1435
1436
1437
1438
1439
1440
1441
1442
1443
1444
1445
1446
1447
1448
1449
1450
public Annotator(string txt) { nearbytes=%%YBYTES%%; inBytes=System.Text.Encoding.UTF8.GetBytes(txt); inPtr=0; writePtr=0; needSpace=false; outBuf=new System.IO.MemoryStream(); annotation_mode = Annotation_Mode.ruby_markup; }
int nearbytes;
public Annotation_Mode annotation_mode;
byte[] inBytes;
int inPtr,writePtr; bool needSpace;
System.IO.MemoryStream outBuf;
const byte EOF = (byte)0; // TODO: a bit hacky
byte nB() {
if (inPtr==inBytes.Length) return EOF;
return inBytes[inPtr++];
}
bool near(string s) {
byte[] bytes=System.Text.Encoding.UTF8.GetBytes(s);
int offset=inPtr, maxPos=inPtr+nearbytes;
if (maxPos > inBytes.Length) maxPos = inBytes.Length;
maxPos -= bytes.Length;
if(offset>nearbytes) offset-=nearbytes; else offset = 0;
while(offset <= maxPos) {
bool ok=true;
for(int i=0; i<bytes.Length; i++) {
if(bytes[i]!=inBytes[offset+i]) { ok=false; break; }
}
if(ok) return true;
offset++;
}
return false;
}
void o(byte c) { outBuf.WriteByte(c); }
void o(string s) { byte[] b=System.Text.Encoding.UTF8.GetBytes(s); outBuf.Write(b,0,b.Length); }
void s() {
if (needSpace) o((byte)' ');
else needSpace=true;
}
void o(int numBytes,string annot) {
s();
switch (annotation_mode) {
case Annotation_Mode.annotations_only:
1452
1453
1454
1455
1456
1457
1458
1459
1460
1461
1462
1463
1464
1465
1466
1467
1468
1469
1470
1471
1472
1473
1474
1475
1476
1477
1478
1479
1480
1481
1482
1483
1484
case Annotation_Mode.ruby_markup:
o("<ruby><rb>");
outBuf.Write(inBytes,writePtr,numBytes);
o("</rb><rt>"); o(annot);
o("</rt></ruby>"); break;
case Annotation_Mode.brace_notation:
o("{");
outBuf.Write(inBytes,writePtr,numBytes);
o("|"); o(annot);
o("}"); break;
}
writePtr += numBytes;
}
void o2(int numBytes,string annot,string title) {
if (annotation_mode == Annotation_Mode.ruby_markup) {
s();
o("<ruby title=\""); o(title);
o("\"><rb>");
outBuf.Write(inBytes,writePtr,numBytes);
writePtr += numBytes;
o("</rb><rt>"); o(annot);
o("</rt></ruby>");
} else o(numBytes,annot);
}
public string result() {
while(inPtr < inBytes.Length) {
int oldPos=inPtr;
topLevelMatch();
if (oldPos==inPtr) { needSpace=false; o(nB()); writePtr++; }
}
return System.Text.Encoding.UTF8.GetString(outBuf.ToArray());
}
"""
cSharp_end = "}\n"
if cSharp_mainNote: cSharp_end += r"""
1487
1488
1489
1490
1491
1492
1493
1494
1495
1496
1497
1498
1499
1500
1501
1502
1503
1504
1505
1506
1507
1508
1509
1510
1511
1512
class Test {
static void Main(string[] args) {
Annotation_Mode annotation_mode = Annotation_Mode.ruby_markup;
for(int i=0; i<args.Length; i++) {
if (args[i]=="--help") {
System.Console.WriteLine("Use --ruby to output ruby markup (default)");
System.Console.WriteLine("Use --raw to output just the annotations without the base text");
System.Console.WriteLine("Use --braces to output as {base-text|annotation}");
return;
} else if(args[i]=="--ruby") {
annotation_mode = Annotation_Mode.ruby_markup;
} else if(args[i]=="--raw") {
annotation_mode = Annotation_Mode.annotations_only;
} else if(args[i]=="--braces") {
annotation_mode = Annotation_Mode.brace_notation;
}
}
System.Console.InputEncoding=System.Text.Encoding.UTF8;
System.Console.OutputEncoding=System.Text.Encoding.UTF8;
Annotator a=new Annotator(System.Console.In.ReadToEnd());
a.annotation_mode = annotation_mode;
System.Console.Write(a.result());
}
}
"""
golang_start = '/* "Go" code '+version_stamp+r"""
1514
1515
1516
1517
1518
1519
1520
1521
1522
1523
1524
1525
1526
1527
1528
1529
1530
1531
1532
1533
1534
1535
1536
1537
1538
1539
1540
1541
To set up a Web service on GAE, put this file in a
subdirectory of your project, and create a top-level .go
file with something like:
package server
import (
"net/http"
"%%PKG%%"
)
func init() {
http.HandleFunc("/", %%PKG%%_handler)
// add other handlers as appropriate
}
func %%PKG%%_handler(w http.ResponseWriter, r *http.Request) {
%%PKG%%.Annotate(r.Body,w)
}
Then in app.yaml:
application: whatever
version: 1
runtime: go
api_version: go1
handlers:
- url: /.*
script: _go_app
Then test with: goapp serve
(and POST to localhost:8080, e.g. via Web Adjuster --htmlFilter="http://localhost:8080")
(To deploy with Web Adjuster also on GAE, you'll need 2 different GAE instances, because
although you could add Web Adjuster on the SAME one - put adjuster's app.yaml into a
python-api.yaml with "module: pythonapi" - there will be the issue of how to set the
URL handlers while making sure that Golang's has priority if it's an exception to .*
- unless you want to port the whole of Web Adjuster to Golang and integrate it into your
annotator that way.)
1550
1551
1552
1553
1554
1555
1556
1557
1558
1559
1560
1561
1562
1563
1564
1565
1566
1567
1568
1569
1570
1571
1572
1573
1574
1575
1576
1577
1578
1579
1580
1581
1582
1583
1584
1585
1586
1587
1588
1589
1590
1591
1592
1593
1594
1595
1596
1597
1598
1599
1600
1601
1602
1603
1604
1605
1606
1607
1608
1609
1610
1611
1612
1613
1614
1615
1616
1617
1618
1619
1620
1621
1622
1623
1624
1625
1626
1627
1628
1629
1630
1631
1632
1633
1634
1635
1636
1637
1638
1639
1640
1641
1642
1643
1644
1645
1646
1647
1648
1649
1650
1651
1652
1653
1654
1655
1656
1657
1658
1659
1660
1661
1662
1663
1664
1665
1666
1667
1668
1669
1670
1671
1672
1673
1674
1675
1676
*/
package %%PKG%%
import (
"sync"
"bytes"
"io"
)
// We have a Mutex for thread safety. TODO: option to put
// the global variables into a per-instance struct instead
var mutex sync.Mutex
var inBytes []byte = nil
var outBuf bytes.Buffer
var inPtr int
var writePtr int
var needSpace bool
var nearbytes int = 15
func nB() byte {
if (inPtr == len(inBytes)) {
return 0
}
tmp := inBytes[inPtr]
inPtr++
return tmp
}
func near(s0 string) bool {
s := make([]byte, len(s0))
copy (s,s0)
offset := inPtr
maxPos := inPtr + nearbytes
if maxPos > len(inBytes) {
maxPos = len(inBytes)
}
maxPos -= len(s)
if (offset > nearbytes) {
offset -= nearbytes
} else {
offset = 0
}
for(offset <= maxPos) {
ok := true ; i := 0
for (i < len(s)) {
if s[i] != inBytes[offset+i] {
ok = false ; break
}
i++
}
if (ok) {
return true
}
offset++
}
return false
}
func oB(c byte) {
outBuf.WriteByte(c)
}
func oS(s string) {
outBuf.WriteString(s)
}
func s() {
if(needSpace) {
oB(' ')
} else {
needSpace = true
}
}
func o(numBytes int,annot string) {
s()
oS("<ruby><rb>")
for (numBytes > 0) {
// TODO: does Go have a way to do this in 1 operation?
oB(inBytes[writePtr])
numBytes--
writePtr++
}
oS("</rb><rt>")
oS(annot)
oS("</rt></ruby>")
}
func o2(numBytes int,annot string,title string) {
s()
oS("<ruby title=\"")
oS(title)
oS("\"><rb>")
for (numBytes > 0) {
// TODO: as above
oB(inBytes[writePtr])
numBytes--
writePtr++
}
oS("</rb><rt>")
oS(annot)
oS("</rt></ruby>")
}
""".replace("%%PKG%%",golang)
golang_end=r"""
func Annotate(src io.Reader, dest io.Writer) {
inBuf := new(bytes.Buffer)
io.Copy(inBuf, src)
mutex.Lock()
inBytes = inBuf.Bytes()
inBuf.Reset() ; outBuf.Reset()
needSpace = false
inPtr = 0 ; writePtr = 0
for(inPtr < len(inBytes)) {
oldPos := inPtr
topLevelMatch()
if oldPos==inPtr {
needSpace = false
oB(nB())
writePtr++
}
}
// outBuf.WriteTo(dest) // may hold up if still locked, try this instead:
outBytes := outBuf.Bytes()
mutex.Unlock()
dest.Write(outBytes)
}
"""
Silas S. Brown
committed
class BytecodeAssembler:
Silas S. Brown
committed
# Bytecode for a virtual machine run by the Javascript version etc
opcodes = {
'jump': 50, # params: address
'call': 51, # params: function address
'return': 52, # (or 'end program' if top level)
'switchbyte': 60, # switch(NEXTBYTE) (params: numBytes-1, bytes (sorted, TODO take advantage of this), addresses, default address)
'copyBytes':71,'o':72,'o2':73, # (don't change these numbers, they're hard-coded below)
'savepos':80, # local to the function
'restorepos':81,
'neartest':90, # params: true-label, false-label, byte nbytes, addresses of conds strings until first of the 2 labels is reached (normally true-label, unless the whole neartest is negated)
}
Silas S. Brown
committed
def __init__(self):
self.l = []
self.d2l = {}
self.lastLabelNo = 0
self.addingPosStack = []
def addOpcode(self,opcode): self.l.append((opcode,))
Silas S. Brown
committed
1695
1696
1697
1698
1699
1700
1701
1702
1703
1704
1705
1706
1707
1708
1709
1710
1711
1712
1713
1714
1715
1716
1717
1718
1719
def addBytes(self,bStr):
if type(bStr)==int: self.l.append(chr(bStr))
elif type(bStr)==str: self.l.append(bStr)
else: raise Exception("unspported bytes type")
def startAddingFunction(self):
self.addingPosStack.append((len(self.l),self.lastLabelNo))
self.lastLabelNo = 0
def finishFunctionAndAddCall(self):
# make sure to add a return instruction before this!
fPtr, self.lastLabelNo = self.addingPosStack[-1]
del self.addingPosStack[-1]
fBody = tuple(self.l[fPtr:]) ; self.l=self.l[:fPtr]
if not fBody in self.d2l: # not a duplicate
self.d2l[fBody] = (-len(self.d2l)-1,)
self.addOpcode('call')
self.l.append(self.d2l[fBody])
def addByteswitch(self,byteArray,labelArray):
assert len(byteArray) + 1 == len(labelArray)
# labelArray has the default case added also (TODO: could re-organize code so the bytes immediately after the switch are either the default or one of the items, saving 1 address)
if not len(byteArray): return # empty switch = no-op
self.addOpcode('switchbyte')
self.addBytes(len(byteArray)-1) # num of bytes in list - 1 (so all 256 values can be accounted for if needed)
self.addBytes("".join(byteArray))
for i in labelArray: self.addRef(i)
def addActions(self,actionList):
# assert type(actionList) in [list,tuple], repr(actionList)
Silas S. Brown
committed
for a in actionList:
assert 1 <= len(a) <= 3 and type(a[0])==int, repr(a)
assert 1 <= a[0] <= 255, "bytecode currently supports markup or copy between 1 and 255 bytes only, not %d (but 0 is reserved for expansion)" % a[0]
Silas S. Brown
committed
1724
1725
1726
1727
1728
1729
1730
1731
1732
1733
1734
1735
1736
1737
1738
1739
1740
1741
1742
1743
1744
1745
1746
1747
1748
1749
1750
1751
1752
1753
1754
1755
1756
1757
1758
1759
1760
1761
1762
1763
1764
1765
1766
self.addBytes(70+len(a)) # 71=copyBytes 72=o() 73=o2
self.addBytes(a[0])
for i in a[1:]: self.addRefToString(i)
def addActionDictSwitch(self,byteSeq_to_action_dict,isFunc=True,labelToJump=None):
# a modified stringSwitch for the bytecode
# Actions aren't strings: they list tuples of either
# 1, 2 or 3 items for copyBytes, o(), o2()
# labelToJump is a jump to insert afterwards if not isFunc and if we don't emit an unconditional 'return'. Otherwise, will ALWAYS end up with a 'return' (even if not isFunc i.e. the main program)
allBytes = set(b[0] for b in byteSeq_to_action_dict.iterkeys() if b)
if isFunc:
self.startAddingFunction()
savePos = len(self.l)
self.addOpcode('savepos')
elif ("" in byteSeq_to_action_dict and len(byteSeq_to_action_dict) > 1) or not labelToJump: # ('not labelToJump' and 'not isFunc' == main program)
savePos = len(self.l)
self.addOpcode('savepos')
else: savePos = None
if "" in byteSeq_to_action_dict and len(byteSeq_to_action_dict) > 1 and len(byteSeq_to_action_dict[""])==1 and not byteSeq_to_action_dict[""][0][1] and all((len(a)==1 and a[0][0][:len(byteSeq_to_action_dict[""][0][0])]==byteSeq_to_action_dict[""][0][0] and not a[0][1]) for a in byteSeq_to_action_dict.itervalues()):
self.addActions(byteSeq_to_action_dict[""][0][0])
l = len(byteSeq_to_action_dict[""][0][0])
byteSeq_to_action_dict = dict((x,[(y[l:],z)]) for x,[(y,z)] in byteSeq_to_action_dict.iteritems())
del self.l[savePos] ; savePos = None
del byteSeq_to_action_dict[""]
self.addActionDictSwitch(byteSeq_to_action_dict) # as a subfunction (ends up adding the call to it)
byteSeq_to_action_dict[""] = [("",[])] # for the end of this func
self.addOpcode('return')
elif allBytes:
allBytes = list(allBytes)
labels = [self.makeLabel() for b in allBytes+[0]]
self.addByteswitch(allBytes,labels)
for case in allBytes:
self.addLabelHere(labels[0]) ; del labels[0]
self.addActionDictSwitch(dict([(k[1:],v) for k,v in byteSeq_to_action_dict.iteritems() if k and k[0]==case]),False,labels[-1])
self.addLabelHere(labels[0])
if not savePos==None: self.addOpcode('restorepos')
if isFunc:
self.addOpcode('return')
if self.l[-1]==self.l[-2]: del self.l[-1] # double return
return self.finishFunctionAndAddCall()
elif "" in byteSeq_to_action_dict:
default_action = ""
for action,conds in byteSeq_to_action_dict[""]:
if conds:
if type(conds)==tuple: negate,conds,nbytes = conds
else: negate,nbytes = False,ybytes_max
Silas S. Brown
committed
assert 1 <= nbytes <= 255, "bytecode supports only single-byte nbytes (but nbytes=0 is reserved for expansion)"
trueLabel,falseLabel = self.makeLabel(),self.makeLabel()
self.addOpcode('neartest')
self.addRef(trueLabel)
self.addRef(falseLabel)
assert type(nbytes)==int
self.addBytes(nbytes)
for c in conds: self.addRefToString(c.encode(outcode))
if negate: trueLabel,falseLabel = falseLabel,trueLabel
Silas S. Brown
committed
1778
1779
1780
1781
1782
1783
1784
1785
1786
1787
1788
1789
1790
1791
1792
1793
1794
1795
1796
1797
1798
1799
1800
1801
self.addLabelHere(trueLabel)
self.addActions(action)
self.addOpcode('return')
self.addLabelHere(falseLabel)
else: default_action = action
if default_action or not byteSeq_to_action_dict[""]:
self.addActions(default_action)
self.addOpcode('return') ; return
if labelToJump:
self.addOpcode('jump')
self.addRef(labelToJump)
else: self.addOpcode('return')
def makeLabel(self):
self.lastLabelNo += 1
return self.lastLabelNo
def addLabelHere(self,labelNo):
assert type(labelNo)==int
assert labelNo, "label 0 not allowed"
self.l.append(labelNo)
def addRef(self,labelNo):
assert type(labelNo)==int
self.l.append(-labelNo)
def addRefToString(self,string):
assert type(string)==str
if python or javascript:
# prepends with a length hint if possible (or if not
# prepends with 0 and null-terminates it)
if 1 <= len(string) < 256:
Silas S. Brown
committed
string = chr(len(string))+string
else: string = chr(0)+string+chr(0)
else: string += chr(0) # just null-termination for C
Silas S. Brown
committed
1809
1810
1811
1812
1813
1814
1815
1816
1817
1818
1819
1820
1821
1822
1823
1824
1825
1826
1827
1828
1829
1830
1831
1832
1833
1834
1835
1836
1837
if not string in self.d2l:
self.d2l[string] = (-len(self.d2l)-1,)
self.l.append(self.d2l[string])
def link(self): # returns resulting bytes
# (add an 'end program' instruction before calling)
def f(*args): raise Exception("Must call link() only once")
self.link = f
sys.stderr.write("Linking... ")
for dat,ref in self.d2l.iteritems():
assert type(ref)==tuple and type(ref[0])==int
self.l.append((-ref[0],)) # the label
if type(dat)==str:
self.l.append(dat) ; continue
# otherwise it's a function, and non-reserved labels are local, so we need to rename them
l2l = {}
for i in dat:
if type(i)==int:
if i>0: j=i
else: j=-i
if not j in l2l:
l2l[j] = self.makeLabel()
if i>0: self.addLabelHere(l2l[j])
else: self.addRef(l2l[j])
else: self.l.append(i) # str or tuple just cp
del self.d2l
# elements of self.l are now:
# - byte strings (just copied in)
# - positive integers (labels)
# - negative integers (references to labels)
# - +ve or -ve integers in tuples (reserved labels: a different counter, used for functions etc)
# strings in tuples: opcodes
Silas S. Brown
committed
# 1st byte of o/p is num bytes needed per address
class TooNarrow(Exception): pass
for numBytes in xrange(1,256):
sys.stderr.write("(%d-bit) " % (numBytes*8))
try:
lDic = {} # the label dictionary
for P in [1,2,3]:
labelMove = 0 # amount future labels have to move by, due to instructions taking longer than we thought on pass 2. NB: this labelMove logic relies on the assumption that, if a short-forward-jump is confirmed in pass 2, then the instructions it jumps over will not have to expand in that pass (otherwise it's possible that the label it jumps to will be moved out of range and the instruction will have to expand on pass 3, causing labels to move on pass 3 which would necessitate another pass; assert should catch this). Assumption should hold in the code we generate ('nested switch' stuff: a 'break' from an inner switch can't possibly refer to a label that occurs after the one referred to by 'break's in the outer switch before that inner switch started, hence if the outer switch is confirmed to be within range of its end label then the inner switch must necessarily be in range of ITS end label) but this might not hold if the generator were to start to emit spaghetti state jumps
1848
1849
1850
1851
1852
1853
1854
1855
1856
1857
1858
1859
1860
1861
1862
1863
1864
1865
1866
1867
1868
1869
1870
1871
1872
1873
1874
1875
1876
1877
1878
1879
1880
1881
1882
1883
1884
1885
compacted = 0
labels_seen_this_pass = set() # to avoid backward jumps (as we can't just apply labelMove to them and see if they're behind the program counter, since need to know if they're backward before knowing if labelMove applies)
r = [chr(numBytes)] ; ll = 1
count = 0
while count < len(self.l):
i = self.l[count] ; count += 1
if type(i)==tuple and type(i[0])==str:
# an opcode: consider rewriting with additional_compact_opcodes if present
opcode = i[0]
i = chr(BytecodeAssembler.opcodes[opcode])
if additional_compact_opcodes:
if opcode=='jump' and type(self.l[count])==int:
# Maybe we can use a 1-byte relative forward jump (up to 128 bytes), useful for 'break;' in a small switch
bytesSaved = numBytes # as we're having a single byte instead of byte + numBytes-addr
if P==1: i = ' ' # optimistic placeholder on pass 1 (might have to replace with a normal jump if the label turns out to be too far away)
elif -self.l[count] in lDic and not -self.l[count] in labels_seen_this_pass and lDic[-self.l[count]]+labelMove-(ll+1) < 0x80: # it fits
compacted += bytesSaved
i = chr(0x80 | (lDic[-self.l[count]]+labelMove-(ll+1)))
else:
if P==2: labelMove += bytesSaved # because we need a normal jump (if P==3 then the labels should already have been moved into place on pass 2)
count -= 1 # counteract the below
count += 1
elif opcode=='switchbyte' and self.l[count] < 20: # might be able to do the short version of switchbyte as well
numItems = self.l[count]+1 # it's len-1
# self.l[count+1] is the bytes; labels start at self.l[count+2]
numLabels = numItems+1 # there's an extra default label at the end
instrLen = 1+numItems+numLabels # 1-byte len, bytes, 1-byte address offsets
bytesSaved = 1+1+numItems+numBytes*numLabels-instrLen
if P==1: i=' '*instrLen # optimistic
elif all(type(self.l[count+N])==int and -self.l[count+N] in lDic and not -self.l[count+N] in labels_seen_this_pass and lDic[-self.l[count+N]]+labelMove-(ll+instrLen) <= 0xFF for N in xrange(2,2+numLabels)): # it fits
compacted += bytesSaved
i = chr(self.l[count])+self.l[count+1]+''.join(chr(self.l[count+N]) for N in xrange(2,2+numLabels))
else:
if P==2: labelMove += bytesSaved
count -= 2+numLabels
count += 2+numLabels
# end of opcode handling/rewriting
if type(i) in [int,tuple]: # labels
Silas S. Brown
committed
if type(i)==int: i2,iKey = i,-i
else: i2,iKey = i[0],(-i[0],)
assert type(i2)==int
# iKey is the lDic key *IF* i is a reference (i.e. i2 is -ve). But i might also be the label itself, in which case lKey is irrelevant.
Silas S. Brown
committed
if i2 > 0: # label going in here
labels_seen_this_pass.add(i)
assert not (i in lDic and not lDic[i] == ll-labelMove), "Changing %s from %d to %d (labelMove=%d P=%d)\n" % (repr(i),lDic[i],ll,labelMove,P)
lDic[i] = ll ; i = ""
Silas S. Brown
committed
elif iKey in lDic: # known label
i = lDic[iKey]
shift = 8*numBytes
if (i >> shift): raise TooNarrow()
j = []
for b in xrange(numBytes):
# MSB-LSB (easier to do in JS)
shift -= 8
j.append(chr((i>>shift)&0xFF))
i = "".join(j)
assert len(i)==numBytes
else: # as-yet unknown label
assert P==1, "undefined label %d" % -i
ll += numBytes
i = ""
if len(i):
r.append(i) ; ll += len(i)
if P==2:
if not additional_compact_opcodes: break # need only 2 passes if have fixed-length addressing
else: assert not labelMove, "Labels move only on pass 2"
sys.stderr.write('.')
r = "".join(r)
if zlib:
self.origLen = ll # needed for efficient malloc in the C code later
r = zlib.compress(r,9)
if additional_compact_opcodes: sys.stderr.write("%d bytes (zlib compressed from %d after opcode compaction saved %d)\n" % (len(r),ll,compacted))
else: sys.stderr.write("%d bytes (zlib compressed from %d)\n" % (len(r),ll))
elif additional_compact_opcodes: sys.stderr.write("%d bytes (opcode compaction saved %d)\n" % (ll,compacted))
else: sys.stderr.write("%d bytes\n" % ll)
Silas S. Brown
committed
except TooNarrow: pass
assert 0, "can't even assemble it with 255-byte addressing !?!"
js_start = '/* Javascript '+version_stamp+r"""
Silas S. Brown
committed
Usage:
- You could just include this code and then call the
annotate() function i.e. var result = annotate(input)
- Or you could use (and perhaps extend) the Annotator
object, and call its annotate() method. If you have
Backbone.JS, Annotator will instead be a generator
(extending Backbone.Model) which you will have to
instantiate yourself (possibly after extending it).
The Annotator object/class is also what will be
exported by this module if you're using Common.JS.
- On Unix systems with Node.JS, you can run this file in
"node" to annotate standard input as a simple test.
*/
var Annotator={
version: '"""+version_stamp+"',\n"
Silas S. Brown
committed
1949
1950
1951
1952
1953
1954
1955
1956
1957
1958
1959
1960
1961
1962
1963
1964
1965
1966
1967
1968
1969
1970
1971
1972
1973
1974
1975
1976
1977
1978
1979
1980
1981
1982
1983
1984
1985
1986
1987
1988
1989
1990
1991
1992
js_end = r"""
annotate: function(input) {
/* TODO: if input is a whole html doc, insert css in head
(e.g. from annoclip and/or adjuster), and hope there's
no stuff that's not to be annotated (form fields...) */
input = unescape(encodeURIComponent(input)); // to UTF-8
var data = this.data;
var addrLen = data.charCodeAt(0);
var dPtr;
var inputLength = input.length;
var p = 0; // read-ahead pointer
var copyP = 0; // copy pointer
var output = new Array();
var needSpace = 0;
function readAddr() {
var i,addr=0;
for (i=addrLen; i; i--) addr=(addr << 8) | data.charCodeAt(dPtr++);
return addr;
}
function readRefStr() {
var a = readAddr(); var l=data.charCodeAt(a);
if (l != 0) return data.slice(a+1,a+l+1);
else return data.slice(a+1,data.indexOf('\x00',a+1));
}
function s() {
if (needSpace) output.push(" ");
else needSpace=1; // for after the word we're about to write (if no intervening bytes cause needSpace=0)
}
function readData() {
var sPos = new Array();
while(1) {
switch(data.charCodeAt(dPtr++)) {
case 50: dPtr = readAddr(); break;
case 51: {
var f = readAddr(); var dO=dPtr;
dPtr = f; readData() ; dPtr = dO;
break; }
case 52: return;
case 60: {
var nBytes = data.charCodeAt(dPtr++)+1;
var i = ((p>=input.length)?-1:data.slice(dPtr,dPtr+nBytes).indexOf(input.charAt(p++)));
Silas S. Brown
committed
if (i==-1) i = nBytes;
dPtr += (nBytes + i * addrLen);
dPtr = readAddr(); break; }
case 71: {
var numBytes = data.charCodeAt(dPtr++);
output.push(input.slice(copyP,copyP+numBytes));
copyP += numBytes; break; }