vhost2ident + parsetm

2 files changed, 40 insertions, 4 deletions

diff --git a/src/lib.rs b/src/lib.rs
index e129a5f..aacde83 100644
--- a/src/lib.rs
+++ b/src/lib.rs
@@ -1,5 +1,6 @@
 #![feature(test)]
 extern crate test;
+extern crate time;
 
 mod parser;
 
@@ -7,7 +8,7 @@ pub struct Line<'a> {
     host: &'a str,
     ident: &'a str,
     user: &'a str,
-    date: &'a str,
+    tm: &'a str,
     method: &'a str,
     path: &'a str,
     proto: &'a str,
@@ -22,4 +23,39 @@ impl<'a> Line<'a> {
     pub fn new(line: &str) -> Result<Line, ()> {
         parser::Parser::new(line).parse()
     }
+
+    // Lighttpd by default stores the vhost in the ident field. Some other servers (notably
+    // varnishncsa) instead store a full URL in the path field. This function normalizes those two
+    // variations into the Lighttpd version: If the path is a full URL, the ident field will be
+    // updated to point to the hostname of the URL, and the path field is updated to only list the
+    // path component. This function does nothing if the path field isn't a full URL.
+    pub fn vhost2ident(&mut self) {
+        let start = if self.path.starts_with("http://") { 7 }
+            else if self.path.starts_with("https://") { 8 } else { return };
+        for (i, c) in (&self.path[start..]).char_indices() {
+            if c == '/' {
+                self.ident = &self.path[start..i+start];
+                self.path = &self.path[i+start..];
+                return;
+            }
+        }
+    }
+
+    pub fn parsetm(&self) -> Result<time::Tm, time::ParseError> {
+        time::strptime(self.tm, "%d/%b/%Y:%H:%M:%S %z")
+    }
+}
+
+
+#[test]
+fn test_vhost2ident() {
+    let mut l = Line::new("host - user [tm] \"method http://blicky.net/ proto\" 200 1").unwrap();
+    l.vhost2ident();
+    assert_eq!(l.ident, "blicky.net");
+    assert_eq!(l.path, "/");
+
+    let mut l = Line::new("host - user [tm] \"method https://whatever.blicky.net/Longer_path?with_stuff proto\" 200 1").unwrap();
+    l.vhost2ident();
+    assert_eq!(l.ident, "whatever.blicky.net");
+    assert_eq!(l.path, "/Longer_path?with_stuff");
 }
diff --git a/src/parser.rs b/src/parser.rs
index 4230498..faa15da 100644
--- a/src/parser.rs
+++ b/src/parser.rs
@@ -98,7 +98,7 @@ impl<'a> Parser<'a> {
             host:    try!(self.uptochar(32, false)),
             ident:   try!(self.uptochar(32, false)),
             user:    try!(self.uptochar(32, false)),
-            date:    { try!(self.consume(91)); try!(self.uptochar(93, false)) },
+            tm:      { try!(self.consume(91)); try!(self.uptochar(93, false)) },
             method:  { try!(self.consume(32)); try!(self.consume(34)); try!(self.uptochar(32, false)) },
             path:    try!(self.uptochar(32, false)),
             proto:   try!(self.uptochar(34, false)),
@@ -120,7 +120,7 @@ fn test_extended() {
     assert_eq!(line.host, "host");
     assert_eq!(line.ident, "ident");
     assert_eq!(line.user, "user");
-    assert_eq!(line.date, "da te");
+    assert_eq!(line.tm, "da te");
     assert_eq!(line.method, "method");
     assert_eq!(line.path, "path");
     assert_eq!(line.proto, "proto");
@@ -140,7 +140,7 @@ fn test_common() {
     assert_eq!(line.host, "host");
     assert_eq!(line.ident, "ident");
     assert_eq!(line.user, "user");
-    assert_eq!(line.date, "date");
+    assert_eq!(line.tm, "date");
     assert_eq!(line.method, "method");
     assert_eq!(line.path, "path");
     assert_eq!(line.proto, "proto");