From 3962b1bdfb2a8a2e3a5ff4f4e51a61b0c44f2e6b Mon Sep 17 00:00:00 2001
From: =?UTF-8?q?John=20Ankarstr=C3=B6m?= <john@ankarstrom.se>
Date: Tue, 23 Aug 2022 01:43:09 +0200
Subject: Add Managed (generic RAII type).

---
 c/data.cpp | 98 ++++++++++++++++++++++++--------------------------------------
 1 file changed, 38 insertions(+), 60 deletions(-)

(limited to 'c/data.cpp')

diff --git a/c/data.cpp b/c/data.cpp
index c1e7177..8d064a8 100644
--- a/c/data.cpp
+++ b/c/data.cpp
@@ -7,63 +7,25 @@
 #include "data.h"
 #include "win.h"
 
-struct InternetFile
+struct XmlError : public std::exception
 {
-	InternetFile(const wchar_t* url)
+	const char* msg;
+	XmlError()
 	{
-		hi = InternetOpen(L"Episode Browser", INTERNET_OPEN_TYPE_DIRECT, nullptr, nullptr, 0);
-		if (!hi)
-			throw Win32Error();
-
-		hiUrl = InternetOpenUrl(hi, url,
-		    nullptr, 0, INTERNET_FLAG_NO_UI, 0);
-		if (!hiUrl) {
-			DWORD e = GetLastError();
-			InternetCloseHandle(hi);
-			throw InternetError(e);
-		}
+		msg = xmlGetLastError()->message;
 	}
-
-	~InternetFile()
+	virtual const char* what() const noexcept
 	{
-		InternetCloseHandle(hiUrl);
-		InternetCloseHandle(hi);
-	}
-
-	DWORD Read(void* buf, DWORD cb)
-	{
-		DWORD cbRead;
-		if (InternetReadFile(hiUrl, buf, cb, &cbRead))
-			return cbRead;
-		else
-			throw InternetError();
+		return msg;
 	}
-
-	HINTERNET hi;
-	HINTERNET hiUrl;
 };
 
-template <auto F, typename T>
-struct XmlPtr
-{
-	XmlPtr(T v) : v(v)
-	{
-		if (!v)
-			throw std::runtime_error(xmlGetLastError()->message);
-	}
-	~XmlPtr() { F(v); }
-	operator T() { return v; }
-	T operator ->() { return v; }
-private:
-	T v;
-};
-
-using HtmlParserCtxtPtr = XmlPtr<xmlFreeParserCtxt, htmlParserCtxtPtr>;
-using HtmlDocPtr = XmlPtr<xmlFreeDoc, htmlDocPtr>;
-using XmlXPathContextPtr = XmlPtr<xmlXPathFreeContext, xmlXPathContextPtr>;
-using XmlXPathObjectPtr = XmlPtr<xmlXPathFreeObject, xmlXPathObjectPtr>;
+using HtmlParserCtxtPtr = Managed<htmlParserCtxtPtr, xmlFreeParserCtxt, XmlError>;
+using HtmlDocPtr = Managed<htmlDocPtr, xmlFreeDoc, XmlError>;
+using XmlXPathContextPtr = Managed<xmlXPathContextPtr, xmlXPathFreeContext, XmlError>;
+using XmlXPathObjectPtr = Managed<xmlXPathObjectPtr, xmlXPathFreeObject, XmlError>;
 static inline void XmlFree(void* p) { xmlFree(p); }
-using XmlCharPtr = XmlPtr<XmlFree, xmlChar*>;
+using XmlCharPtr = Managed<xmlChar*, XmlFree, XmlError>;
 
 template <size_t N>
 bool WcharsFromXmlchars(wchar_t (&dst)[N], XmlCharPtr utf8) noexcept
@@ -93,24 +55,36 @@ bool WcharsFromXmlchars(wchar_t (&dst)[N], XmlCharPtr utf8) noexcept
 void FetchData()
 {
 	LIBXML_TEST_VERSION;
+	using InternetHandle = Managed<HINTERNET, InternetCloseHandle, InternetError>;
 
-	InternetFile inf(L"https://www.detectiveconanworld.com/wiki/Anime");
-	//InternetFile inf(L"file://C:/Users/John/Desktop/dcw.html");
-	char buf[1024];
+	/* The remote data is retrieved using WinINet from the
+	 * Detective Conan World wiki. Using libxml2's "push parser",
+	 * the HTML is parsed piece by piece as it is retrieved. */
 
+	InternetHandle hi = InternetOpen(L"Episode Browser", INTERNET_OPEN_TYPE_DIRECT, nullptr, nullptr, 0);
+	InternetHandle hiUrl = InternetOpenUrl(hi, L"https://www.detectiveconanworld.com/wiki/Anime",
+		    nullptr, 0, INTERNET_FLAG_NO_UI, 0); //L"file://C:/Users/John/Desktop/dcw.html"
+
+	char buf[1024];
 	HtmlParserCtxtPtr ctxt = htmlCreatePushParserCtxt(nullptr, nullptr,
 	    buf, sizeof(buf), "https://www.detectiveconanworld.com/wiki/Anime",
 	    XML_CHAR_ENCODING_UTF8);
-
 	htmlCtxtUseOptions(ctxt, HTML_PARSE_RECOVER|HTML_PARSE_NOERROR|HTML_PARSE_NOWARNING);
 
-	while (const DWORD cbRead = inf.Read(&buf, sizeof(buf))) {
+	BOOL r;
+	DWORD cbRead;
+	while (r = InternetReadFile(hiUrl, buf, sizeof(buf), &cbRead), cbRead) {
+		if (!r)
+			throw InternetError();
 		if (!htmlParseChunk(ctxt, buf, cbRead, 0))
 			throw std::runtime_error(xmlGetLastError()->message);
 	}
 	htmlParseChunk(ctxt, buf, 0, 1); /* Stop parsing. */
 
-	/* Find table rows containing episode data. */
+	/* The episode data are contained in table rows matching a
+	 * (very!) specific XPath query. This is fragile
+	 * theoretically, but unlikely to break practically. */
+
 	HtmlDocPtr doc = ctxt->myDoc;
 	XmlXPathContextPtr xpathCtx = xmlXPathNewContext(doc);
 	XmlXPathObjectPtr xpathObj = xmlXPathEvalExpression(
@@ -133,7 +107,10 @@ void FetchData()
 		ElvDataA& e = g_fvElv.At(i);
 		DlvDataA& d = g_fvDlv.At(i);
 
-		/* Get cells. */
+		/* Each datum is contained within a specific cell in
+		 * the row. The child element count above ensures that
+		 * none of the following nodes are null. */
+
 		const xmlNodePtr nodeEp = xmlFirstElementChild(node);
 		const xmlNodePtr nodeTitle = xmlNextElementSibling(xmlNextElementSibling(nodeEp));
 		const xmlNodePtr nodeDate = xmlNextElementSibling(nodeTitle);
@@ -142,17 +119,18 @@ void FetchData()
 		const xmlNodePtr nodeHint = xmlNextElementSibling(nodeSource);
 
 		WcharsFromXmlchars(e.siEp, xmlNodeGetContent(nodeEp));
-		e.siEp[wcscspn(e.siEp, L"W")] = 0; /* Remove WPS suffix. */
+		e.siEp[wcscspn(e.siEp, L"W")] = 0; /* Remove potential "WPS" suffix. */
 		WcharsFromXmlchars(e.title, xmlNodeGetContent(nodeTitle));
 		WcharsFromXmlchars(d.date, xmlNodeGetContent(nodeDate));
 		WcharsFromXmlchars(d.source, xmlNodeGetContent(nodeSource));
 		WcharsFromXmlchars(d.hint, xmlNodeGetContent(nodeHint));
 
-		/* Get wiki URL. */
+		/* Retrieve the link to the episode's wiki entry,
+		 * which should be the first (and only) child element
+		 * of the title node. */
 		const xmlNodePtr nodeLink = xmlFirstElementChild(nodeTitle);
 		if (nodeLink)
-			WcharsFromXmlchars(d.wiki,
-			    xmlGetProp(nodeLink, reinterpret_cast<const xmlChar*>("href")));
+			WcharsFromXmlchars(d.wiki, xmlGetProp(nodeLink, reinterpret_cast<const xmlChar*>("href")));
 	}
 }
 
-- 
cgit v1.2.3