From 7dae824261bd658c2429371dabee5c7c04f09ab9 Mon Sep 17 00:00:00 2001
From: James Antill <james@fedoraproject.org>
Date: Fri, 2 May 2008 21:36:54 +0000
Subject: [PATCH] - Fix reget's against servers that don't allow Range
 requests, also tweaks - reget == check_timestamp, if anyone/thing uses that.
 - Resolves: bug#435156 - Fix minor typo in progress for single instance.

---
 python-urlgrabber.spec         | 10 +++++-
 urlgrabber-grab-no-range.patch | 64 ++++++++++++++++++++++++++++++++++
 urlgrabber-progress-ui.patch   |  2 +-
 3 files changed, 74 insertions(+), 2 deletions(-)
 create mode 100644 urlgrabber-grab-no-range.patch

diff --git a/python-urlgrabber.spec b/python-urlgrabber.spec
index 2fe2c9c..c3b281f 100644
--- a/python-urlgrabber.spec
+++ b/python-urlgrabber.spec
@@ -3,13 +3,14 @@
 Summary: A high-level cross-protocol url-grabber
 Name: python-urlgrabber
 Version: 3.0.0
-Release: 6%{?dist}
+Release: 7%{?dist}
 Source0: urlgrabber-%{version}.tar.gz
 Patch0: urlgrabber-keepalive.patch
 Patch1: urlgrabber-string-type.patch
 Patch2: urlgrabber-3.0.0-cleanup.patch
 Patch3: urlgrabber-ftp-port.patch
 Patch4: urlgrabber-progress-ui.patch
+Patch5: urlgrabber-grab-no-range.patch
 License: LGPLv2+
 Group: Development/Libraries
 BuildRoot: %{_tmppath}/%{name}-%{version}-%{release}-root
@@ -30,6 +31,7 @@ authentication, proxies and more.
 %patch2 -p1
 %patch3 -p0
 %patch4 -p0
+%patch5 -p1
 
 %build
 python setup.py build
@@ -49,6 +51,12 @@ rm -rf $RPM_BUILD_ROOT
 %{_bindir}/urlgrabber
 
 %changelog
+* Fri May  2 2008 James Antill <james@fedoraproject.org> 3.0.0-7
+- Fix reget's against servers that don't allow Range requests, also tweaks
+- reget == check_timestamp, if anyone/thing uses that.
+- Resolves: bug#435156
+- Fix minor typo in progress for single instance.
+
 * Mon Apr  7 2008 James Antill <james@fedoraproject.org> 3.0.0-6
 - Fix the ftp byterange port problem:
 - Resolves: bug#419241
diff --git a/urlgrabber-grab-no-range.patch b/urlgrabber-grab-no-range.patch
new file mode 100644
index 0000000..32a3d05
--- /dev/null
+++ b/urlgrabber-grab-no-range.patch
@@ -0,0 +1,64 @@
+diff -ru urlgrabber-3.0.0-orig/urlgrabber/grabber.py urlgrabber-3.0.0/urlgrabber/grabber.py
+--- urlgrabber-3.0.0-orig/urlgrabber/grabber.py	2008-05-02 17:18:14.000000000 -0400
++++ urlgrabber-3.0.0/urlgrabber/grabber.py	2008-05-02 17:30:35.000000000 -0400
+@@ -400,6 +400,11 @@
+     HTTPException = None
+ 
+ try:
++    from httplib import HTTPResponse
++except ImportError, msg:
++    HTTPResponse = None
++
++try:
+     # This is a convenient way to make keepalive optional.
+     # Just rename the module so it can't be imported.
+     import keepalive
+@@ -1065,8 +1070,21 @@
+         req = urllib2.Request(self.url, self.opts.data) # build request object
+         self._add_headers(req) # add misc headers that we need
+         self._build_range(req) # take care of reget and byterange stuff
++        
++        def _checkNoReget(fo):
++            # HTTP can disallow Range requests
++            
++            if self.opts.reget is None:
++                return False
++            if 'Range' not in req.headers:
++                return False
++            if not isinstance(fo, HTTPResponse):
++                return False
++
++            return fo.code != 206
+ 
+         fo, hdr = self._make_request(req, opener)
++        fetch_again = 0
+         if self.reget_time and self.opts.reget == 'check_timestamp':
+             # do this if we have a local file with known timestamp AND
+             # we're in check_timestamp reget mode.
+@@ -1078,14 +1096,26 @@
+             except (TypeError,):
+                 fetch_again = 1
+             
++        if _checkNoReget(fo): # doing a "reget" didn't work, so fixup
++            fetch_again = 0
++            self.opts.reget = None
++            self.append = 0
++            self._amount_read = 0
++
++        if True:
+             if fetch_again:
+                 # the server version is newer than the (incomplete) local
+                 # version, so we should abandon the version we're getting
+                 # and fetch the whole thing again.
+                 fo.close()
++                #  Without this we'll have to read all of the previous request
++                # data. For "large" requests, this is very bad.
++                fo.close_connection()
+                 self.opts.reget = None
+                 del req.headers['Range']
+                 self._build_range(req)
++                # This doesn't get reset in _build_range() *sigh*
++                self._amount_read = 0
+                 fo, hdr = self._make_request(req, opener)
+ 
+         (scheme, host, path, parm, query, frag) = urlparse.urlparse(self.url)
diff --git a/urlgrabber-progress-ui.patch b/urlgrabber-progress-ui.patch
index ce911f1..5fe3cda 100644
--- a/urlgrabber-progress-ui.patch
+++ b/urlgrabber-progress-ui.patch
@@ -47,7 +47,7 @@ diff -u -r1.7 urlgrabber/progress.py
 +            bar = '='*int(10 * frac)
 +            ave_dl = format_number(self.re.average_rate())
 +            if sofar_size is None:
-+                out = '\r%-25.25s %3i%% |%-14.14s| %5sB/s | %5sB %9s ETA ' % \
++                out = '\r%-29.29s %3i%% |%-10.10s| %5sB/s | %5sB %9s ETA ' % \
 +                      (text, frac*100, bar, ave_dl, fread, frtime)
 +            else:
 +                fmt ='\r%-22.22s %3i%% |%4i%% |%-10.10s| %5sB/s | %5sB %9s ETA '