From 41221592b34a5ed6b9c2172c170b9b38dcfca7d8 Mon Sep 17 00:00:00 2001
From: Simon Kornblith <simon@simonster.com>
Date: Sun, 16 Dec 2007 03:59:39 +0000
Subject: [PATCH] Addresses #749, BibTeX issues

- import UTF-8 by default (as far as I can tell, this is no worse than the previous behavior, which would probably have read all BibTeX files as ASCII)
- use only ascii characters in cite keys
- present BibTeX UTF-8 export option on export
- hide disabled export options, rather than graying them out, to avoid the indication that other formats will not be exported as UTF-8
---
 chrome/content/zotero/exportOptions.js       |  9 +++-
 chrome/content/zotero/xpcom/cite.js          | 15 +++++++
 chrome/content/zotero/xpcom/translate.js     |  2 -
 chrome/locale/en-US/zotero/zotero.properties |  1 +
 defaults/preferences/zotero.js               |  1 -
 scrapers.sql                                 | 44 ++++++++++----------
 6 files changed, 46 insertions(+), 26 deletions(-)

diff --git a/chrome/content/zotero/exportOptions.js b/chrome/content/zotero/exportOptions.js
index ba126c017..3ba676735 100644
--- a/chrome/content/zotero/exportOptions.js
+++ b/chrome/content/zotero/exportOptions.js
@@ -109,6 +109,8 @@ var Zotero_File_Interface_Export = new function() {
 		}
 		
 		var optionsBox = document.getElementById("translator-options");
+		optionsBox.hidden = true;
+		var haveOption = false;
 		for(var i=0; i<optionsBox.childNodes.length; i++) {
 			// loop through options to see which should be enabled
 			var node = optionsBox.childNodes[i];
@@ -116,7 +118,8 @@ var Zotero_File_Interface_Export = new function() {
 			
 			if(translatorOptions[optionName] != undefined) {
 				// option should be enabled
-				node.disabled = undefined;
+				optionsBox.hidden = undefined;
+				node.hidden = undefined;
 				
 				var defValue = translatorOptions[optionName];
 				if(typeof(defValue) == "boolean") {
@@ -131,10 +134,12 @@ var Zotero_File_Interface_Export = new function() {
 				}
 			} else {
 				// option should be disabled and unchecked to prevent confusion
-				node.disabled = true;
+				node.hidden = true;
 				node.setAttribute("checked", "false");
 			}
 		}
+		
+		window.sizeToContent();
 	}
 	
 	/*
diff --git a/chrome/content/zotero/xpcom/cite.js b/chrome/content/zotero/xpcom/cite.js
index 8ae864fac..946d5b7ee 100644
--- a/chrome/content/zotero/xpcom/cite.js
+++ b/chrome/content/zotero/xpcom/cite.js
@@ -802,6 +802,21 @@ Zotero.CSL.prototype._processNames = function(item, element, formattedString, co
 							if(formattedString.format == "Sort") {
 								// for sort, we use the plain names
 								var name = creators[i].getNameVariable("lastName");
+								
+								// cut off lowercase parts of otherwise capitalized names (e.g., "de")
+								var lastNameParts = name.split(" ");
+								if(lastNameParts.length > 1 && lastNameParts[0].length <= 4
+										&& lastNameParts[0][0].toLowerCase() == lastNameParts[0][0]
+										&& lastNameParts[lastNameParts.length-1][0].toUpperCase() == lastNameParts[lastNameParts.length-1][0]) {
+									name = "";
+									for(var k=1; k<lastNameParts.length; k++) {
+										if(lastNameParts[k][0].toUpperCase() == lastNameParts[k][0]) {
+											name += " "+lastNameParts[k];
+										}
+									}
+									name = name.substr(1);
+								}
+								
 								var firstName = creators[i].getNameVariable("firstName");
 								if(name && firstName) name += ", ";
 								name += firstName;
diff --git a/chrome/content/zotero/xpcom/translate.js b/chrome/content/zotero/xpcom/translate.js
index ed35829f9..20c641c14 100644
--- a/chrome/content/zotero/xpcom/translate.js
+++ b/chrome/content/zotero/xpcom/translate.js
@@ -661,8 +661,6 @@ Zotero.Translate.prototype._generateSandbox = function() {
 		this._sandbox.Zotero.Utilities = new Zotero.Utilities();
 	}
 	
-	this._sandbox.Zotero.useBibtexUTF8 = Zotero.Prefs.get("export.unicodeBibTeX");
-	
 	if(this.type == "export") {
 		// add routines to retrieve items and collections
 		this._sandbox.Zotero.nextItem = function() { return me._exportGetItem() };
diff --git a/chrome/locale/en-US/zotero/zotero.properties b/chrome/locale/en-US/zotero/zotero.properties
index 67d2fc899..74d16bdc4 100644
--- a/chrome/locale/en-US/zotero/zotero.properties
+++ b/chrome/locale/en-US/zotero/zotero.properties
@@ -451,6 +451,7 @@ fulltext.indexState.partial								= Partial
 
 exportOptions.exportNotes       = Export Notes
 exportOptions.exportFileData    = Export Files
+exportOptions.UTF8				= Export as UTF-8
 
 date.daySuffixes                = st, nd, rd, th
 date.abbreviation.year          = y
diff --git a/defaults/preferences/zotero.js b/defaults/preferences/zotero.js
index 6085729a0..9e04d638b 100644
--- a/defaults/preferences/zotero.js
+++ b/defaults/preferences/zotero.js
@@ -61,7 +61,6 @@ pref("extensions.zotero.export.lastStyle", 'http://www.zotero.org/styles/chicago
 pref("extensions.zotero.export.bibliographySettings", 'save-as-rtf');
 pref("extensions.zotero.export.bibliographyLocale", '');
 pref("extensions.zotero.export.citePaperJournalArticleURL", false);
-pref("extensions.zotero.export.unicodeBibTeX", false);
 
 pref("extensions.zotero.export.quickCopy.setting", 'bibliography=http://www.zotero.org/styles/chicago-note.csl');
 
diff --git a/scrapers.sql b/scrapers.sql
index 8cf0835fb..b3291298b 100644
--- a/scrapers.sql
+++ b/scrapers.sql
@@ -15780,8 +15780,9 @@ function doExport() {
 	}
 }');
 
-REPLACE INTO translators VALUES ('9cb70025-a888-4a29-a210-93ec52da40d4', '1.0.0b4.r1', '', '2007-12-12 12:00:00', '1', '100', '3', 'BibTeX', 'Simon Kornblith', 'bib', 
+REPLACE INTO translators VALUES ('9cb70025-a888-4a29-a210-93ec52da40d4', '1.0.0b4.r1', '', '2007-12-16 03:13:58', '1', '100', '3', 'BibTeX', 'Simon Kornblith', 'bib', 
 'Zotero.configure("dataMode", "block");
+Zotero.addOption("UTF8", false);
 
 function detectImport() {
 	var block = "";
@@ -17482,10 +17483,8 @@ function doImport() {
 	var read = "", text = "", recordCloseElement = false;
 	var type = false;
 	
-	if (Zotero.useBibtexUTF8) {
-	    Zotero.debug("Setting UTF-8 on import");
-	    Zotero.setCharacterSet("UTF-8");
-	}
+	Zotero.setCharacterSet("UTF-8");
+	
 	while(read = Zotero.read(1)) {
 		if(read == "@") {
 			type = "";
@@ -17511,7 +17510,7 @@ function writeMacroField(field, value) {
 	}
 	
 	Zotero.write(",\n\t"+field+" = ");
-	if (!Zotero.useBibtexUTF8) {
+	if (!Zotero.getOption("UTF8")) {
 	    var tvalue = value.toString();
 	    value = "";
 	    for (var j in tvalue) {
@@ -17529,7 +17528,7 @@ function writeMacroField(field, value) {
 function writeField(field, value) {
 	if(!value) return;
 	Zotero.write(",\n\t"+field+" = {");
-	if (!Zotero.useBibtexUTF8) {
+	if (!Zotero.getOption("UTF8")) {
 	    var tvalue = value.toString();
 	    value = "";
 	    for (var j in tvalue) {
@@ -17547,8 +17546,7 @@ function writeField(field, value) {
 
 var numberRe = /^[0-9]+/;
 function doExport() {
-	if (Zotero.useBibtexUTF8) {
-	    Zotero.debug("Setting UTF-8 on import");
+	if(Zotero.getOption("UTF8")) {
 	    Zotero.setCharacterSet("UTF-8");
 	}
 	
@@ -17565,33 +17563,37 @@ function doExport() {
 		// create a unique citation key
 		var basekey = "";
 		if(item.creators && item.creators[0] && item.creators[0].lastName) {
-			basekey = item.creators[0].lastName.toLowerCase().replace(/ /g,"_").replace(/,/g,"");
+			basekey += "_" + item.creators[0].lastName.toLowerCase().replace(/ /g,"_").replace(/,/g,"");
 		}
 		
 		// include the item title as part of the citation key
 		if (item["title"]) {
 			// this is a list of words that should not appear as part of the citation key
-			var bannedTitleKeys = {"a" : 1, "an" : 1, "does": 1, "how": 1, "it''s": 1, "on" : 1, "some": 1, "the" : 1, "this" : 1, "why" : 1 };
-			var titleElements = item["title"].split(" ");
-			var appendKey = "";
-			for (te in titleElements) {
-				if (!bannedTitleKeys[titleElements[te].toLowerCase()]) {
-					appendKey = "_" + titleElements[te].toLowerCase() + "_";
+			var bannedTitleKeys = ["a", "an", "from", "does", "how", "it''s", "its", "on", "some", "the", "this", "why"];
+			var titleElements = item["title"].toLowerCase().split(" ");
+			for(var te in titleElements) {
+				if (bannedTitleKeys.indexOf(titleElements[te]) == -1) {
+					basekey += "_" + titleElements[te];
 					break;
-					}
 				}
-				basekey = basekey + appendKey;
+			}
         }
 
 		if(item.date) {
 			var date = Zotero.Utilities.strToDate(item.date);
 			if(date.year && numberRe.test(date.year)) {
-				basekey += date.year;
+				basekey += "_" + date.year;
 			}
 		}
 		
-		// make sure we do not have any other funny characters
-		basekey = basekey.replace(/[\. ,'':\"!&]/g,"");
+		// for now, remove any characters not explicitly known to be allowed;
+		// we might want to allow UTF-8 citation keys in the future, depending
+		// on implementation support.
+		//
+		// no matter what, we want to make sure we exclude
+		// " # % '' ( ) , = { } ~ and backslash
+		
+		basekey = basekey.substr(1).replace(/[^a-z0-9\!\$\&\*\+\-\.\/\:\;\<\>\?\[\]\^\_\`\|]+/g,"");
 		var citekey = basekey;
 		var i = 0;
 		while(citekeys[citekey]) {