[Bug-wget] [PATCH 14/27] New: Metalink file size mismatch returns error code METALINK_SIZE_ERROR

[Matthew White]

* src/wget.h (uerr_t): Add error code METALINK_SIZE_ERROR to enum
* src/metalink.c (retrieve_from_metalink): Use boolean variable
  size_ok, when false set retr_err to METALINK_SIZE_ERROR
* testenv/Makefile.am: Add new file
* testenv/Test-metalink-xml-size.py: New file. Metalink/XML file size
  tests (<size></size>)

Before this patch, no appropriate error code was returned to inform a
file size mismatch.

This patch introduces the error code METALINK_SIZE_ERROR to inform a
file size mismatch.
---
 src/metalink.c                    |  11 +++
 src/wget.h                        |   3 +-
 testenv/Makefile.am               |   3 +-
 testenv/Test-metalink-xml-size.py | 202 ++++++++++++++++++++++++++++++++++++++
 4 files changed, 217 insertions(+), 2 deletions(-)
 create mode 100755 testenv/Test-metalink-xml-size.py

diff --git a/src/metalink.c b/src/metalink.c
index b729450..5108a5e 100644
--- a/src/metalink.c
+++ b/src/metalink.c
@@ -102,6 +102,7 @@ retrieve_from_metalink (const metalink_t* metalink)
       char *basename = NULL;
       char *safename = NULL;
       char *destname = NULL;
+      bool size_ok = false;
       bool hash_ok = false;
 
       uerr_t retr_err = METALINK_MISSING_RESOURCE;
@@ -325,10 +326,12 @@ retrieve_from_metalink (const metalink_t* metalink)
                   continue;
                 }
 
+              size_ok = false;
               logprintf (LOG_VERBOSE, _("Computing size for %s\n"), quote 
(destname));
 
               if (!mfile->size)
                 {
+                  size_ok = true;
                   logprintf (LOG_VERBOSE, _("File size not declared. Skipping 
check.\n"));
                 }
               else
@@ -356,6 +359,7 @@ retrieve_from_metalink (const metalink_t* metalink)
                     }
                   else
                     {
+                      size_ok = true;
                       logputs (LOG_VERBOSE, _("Size matches.\n"));
                     }
                 }
@@ -679,6 +683,13 @@ gpg_skip_verification:
           logprintf (LOG_VERBOSE, _("Failed to download %s. Skipping 
resource.\n"),
                      quote (destname ? destname : safename));
         }
+      else if (!size_ok)
+        {
+          retr_err = METALINK_SIZE_ERROR;
+          logprintf (LOG_NOTQUIET,
+                     _("File %s retrieved but size does not match. "
+                       "\n"), quote (destname));
+        }
       else if (!hash_ok)
         {
           retr_err = METALINK_CHKSUM_ERROR;
diff --git a/src/wget.h b/src/wget.h
index eacf725..ed57245 100644
--- a/src/wget.h
+++ b/src/wget.h
@@ -361,7 +361,8 @@ typedef enum
   TIMECONV_ERR,
   METALINK_PARSE_ERROR, METALINK_RETR_ERROR,
   METALINK_CHKSUM_ERROR, METALINK_SIG_ERROR, METALINK_MISSING_RESOURCE,
-  RETR_WITH_METALINK
+  RETR_WITH_METALINK,
+  METALINK_SIZE_ERROR
 } uerr_t;
 
 /* 2005-02-19 SMS.
diff --git a/testenv/Makefile.am b/testenv/Makefile.am
index f8c3da9..b671438 100644
--- a/testenv/Makefile.am
+++ b/testenv/Makefile.am
@@ -44,7 +44,8 @@ if METALINK_IS_ENABLED
     Test-metalink-xml-prefix-trust.py               \
     Test-metalink-xml-relprefix-trust.py            \
     Test-metalink-xml-absprefix-trust.py            \
-    Test-metalink-xml-homeprefix-trust.py
+    Test-metalink-xml-homeprefix-trust.py           \
+    Test-metalink-xml-size.py
 else
   METALINK_TESTS =
 endif
diff --git a/testenv/Test-metalink-xml-size.py 
b/testenv/Test-metalink-xml-size.py
new file mode 100755
index 0000000..4c7ccb7
--- /dev/null
+++ b/testenv/Test-metalink-xml-size.py
@@ -0,0 +1,202 @@
+#!/usr/bin/env python3
+from sys import exit
+from test.http_test import HTTPTest
+from misc.wget_file import WgetFile
+import hashlib
+
+"""
+    This is to test Metalink/XML file size check in Wget.
+
+    With --trust-server-names, trust the metalink:file names.
+
+    Without --trust-server-names, don't trust the metalink:file names:
+    use the basename of --input-metalink, and add a sequential number
+    (e.g. .#1, .#2, etc.).
+
+    Strip the directory from unsafe paths.
+"""
+############# File Definitions ###############################################
+bad = "Ouch!"
+
+File1 = "Would you like some Tea?"
+File1_lowPref = "Do not take this"
+File1_sha256 = hashlib.sha256 (File1.encode ('UTF-8')).hexdigest ()
+
+File2 = "This is gonna be good"
+File2_lowPref = "Not this one too"
+File2_sha256 = hashlib.sha256 (File2.encode ('UTF-8')).hexdigest ()
+
+File3 = "A little more, please"
+File3_lowPref = "That's just too much"
+File3_sha256 = hashlib.sha256 (File3.encode ('UTF-8')).hexdigest ()
+
+File4 = "Maybe a biscuit?"
+File4_lowPref = "No, thanks"
+File4_sha256 = hashlib.sha256 (File4.encode ('UTF-8')).hexdigest ()
+
+File5 = "More Tea...?"
+File5_lowPref = "I have to go..."
+File5_sha256 = hashlib.sha256 (File5.encode ('UTF-8')).hexdigest ()
+File5_size = str (len (File5))
+
+MetaXml = \
+"""<?xml version="1.0" encoding="utf-8"?>
+<metalink version="3.0" xmlns="http://www.metalinker.org/";>
+  <publisher>
+    <name>GNU Wget</name>
+  </publisher>
+  <license>
+    <name>GNU GPL</name>
+    <url>http://www.gnu.org/licenses/gpl.html</url>
+  </license>
+  <identity>Wget Test Files</identity>
+  <version>1.2.3</version>
+  <description>Wget Test Files description</description>
+  <files>
+    <file name="File1">
+      <size></size> <!-- left empty on purpose, same as 0, ignored -->
+      <verification>
+        <hash type="sha256">{{FILE1_HASH}}</hash>
+      </verification>
+      <resources>
+        <url type="http" 
preference="35">http://{{SRV_HOST}}:{{SRV_PORT}}/wrong_file</url>
+        <url type="http" 
preference="40">http://{{SRV_HOST}}:{{SRV_PORT}}/404</url>
+        <url type="http" 
preference="25">http://{{SRV_HOST}}:{{SRV_PORT}}/File1_lowPref</url>
+        <url type="http" 
preference="30">http://{{SRV_HOST}}:{{SRV_PORT}}/File1</url>
+      </resources>
+    </file>
+    <file name="File2">
+      <size>2</size> <!-- wrong size -->
+      <verification>
+        <hash type="sha256">{{FILE2_HASH}}</hash>
+      </verification>
+      <resources>
+        <url type="http" 
preference="35">http://{{SRV_HOST}}:{{SRV_PORT}}/wrong_file</url>
+        <url type="http" 
preference="40">http://{{SRV_HOST}}:{{SRV_PORT}}/404</url>
+        <url type="http" 
preference="25">http://{{SRV_HOST}}:{{SRV_PORT}}/File2_lowPref</url>
+        <url type="http" 
preference="30">http://{{SRV_HOST}}:{{SRV_PORT}}/File2</url>
+      </resources>
+    </file>
+    <file name="File3">
+      <size>0</size> <!-- 0 is ignored -->
+      <verification>
+        <hash type="sha256">{{FILE3_HASH}}</hash>
+      </verification>
+      <resources>
+        <url type="http" 
preference="35">http://{{SRV_HOST}}:{{SRV_PORT}}/wrong_file</url>
+        <url type="http" 
preference="40">http://{{SRV_HOST}}:{{SRV_PORT}}/404</url>
+        <url type="http" 
preference="25">http://{{SRV_HOST}}:{{SRV_PORT}}/File3_lowPref</url>
+        <url type="http" 
preference="30">http://{{SRV_HOST}}:{{SRV_PORT}}/File3</url>
+      </resources>
+    </file>
+    <file name="File4">
+      <size>-240</size> <!-- negative values are parsed as 0 by libmetalink, 
then ignored -->
+      <verification>
+        <hash type="sha256">{{FILE4_HASH}}</hash>
+      </verification>
+      <resources>
+        <url type="http" 
preference="35">http://{{SRV_HOST}}:{{SRV_PORT}}/wrong_file</url>
+        <url type="http" 
preference="40">http://{{SRV_HOST}}:{{SRV_PORT}}/404</url>
+        <url type="http" 
preference="25">http://{{SRV_HOST}}:{{SRV_PORT}}/File4_lowPref</url>
+        <url type="http" 
preference="30">http://{{SRV_HOST}}:{{SRV_PORT}}/File4</url>
+      </resources>
+    </file>
+    <file name="File5">
+      <size>{{FILE5_SIZE}}</size> <!-- right size -->
+      <verification>
+        <hash type="sha256">{{FILE5_HASH}}</hash>
+      </verification>
+      <resources>
+        <url type="http" 
preference="35">http://{{SRV_HOST}}:{{SRV_PORT}}/wrong_file</url>
+        <url type="http" 
preference="40">http://{{SRV_HOST}}:{{SRV_PORT}}/404</url>
+        <url type="http" 
preference="25">http://{{SRV_HOST}}:{{SRV_PORT}}/File5_lowPref</url>
+        <url type="http" 
preference="30">http://{{SRV_HOST}}:{{SRV_PORT}}/File5</url>
+      </resources>
+    </file>
+  </files>
+</metalink>
+"""
+
+wrong_file = WgetFile ("wrong_file", bad)
+
+File1_orig = WgetFile ("File1", File1)
+File1_down = WgetFile ("test.metalink.#1", File1)
+File1_nono = WgetFile ("File1_lowPref", File1_lowPref)
+
+# wrong size
+File2_orig = WgetFile ("File2", File2)
+File2_nono = WgetFile ("File2_lowPref", File2_lowPref)
+
+File3_orig = WgetFile ("File3", File3)
+File3_down = WgetFile ("test.metalink.#3", File3)
+File3_nono = WgetFile ("File3_lowPref", File3_lowPref)
+
+File4_orig = WgetFile ("File4", File4)
+File4_down = WgetFile ("test.metalink.#4", File4)
+File4_nono = WgetFile ("File4_lowPref", File4_lowPref)
+
+File5_orig = WgetFile ("File5", File5)
+File5_down = WgetFile ("test.metalink.#5", File5)
+File5_nono = WgetFile ("File5_lowPref", File5_lowPref)
+
+MetaFile = WgetFile ("test.metalink", MetaXml)
+
+WGET_OPTIONS = "--input-metalink test.metalink"
+WGET_URLS = [[]]
+
+Files = [[
+    wrong_file,
+    File1_orig, File1_nono,
+    File2_orig, File2_nono,
+    File3_orig, File3_nono,
+    File4_orig, File4_nono,
+    File5_orig, File5_nono
+]]
+Existing_Files = [MetaFile]
+
+ExpectedReturnCode = 1
+ExpectedDownloadedFiles = [
+    File1_down,
+    File3_down,
+    File4_down,
+    File5_down,
+    MetaFile
+]
+
+################ Pre and Post Test Hooks #####################################
+pre_test = {
+    "ServerFiles"       : Files,
+    "LocalFiles"        : Existing_Files
+}
+test_options = {
+    "WgetCommands"      : WGET_OPTIONS,
+    "Urls"              : WGET_URLS
+}
+post_test = {
+    "ExpectedFiles"     : ExpectedDownloadedFiles,
+    "ExpectedRetcode"   : ExpectedReturnCode
+}
+
+http_test = HTTPTest (
+                pre_hook=pre_test,
+                test_params=test_options,
+                post_hook=post_test,
+)
+
+http_test.server_setup()
+### Get and use dynamic server sockname
+srv_host, srv_port = http_test.servers[0].server_inst.socket.getsockname ()
+
+MetaXml = MetaXml.replace('{{FILE1_HASH}}', File1_sha256)
+MetaXml = MetaXml.replace('{{FILE2_HASH}}', File2_sha256)
+MetaXml = MetaXml.replace('{{FILE3_HASH}}', File3_sha256)
+MetaXml = MetaXml.replace('{{FILE4_HASH}}', File4_sha256)
+MetaXml = MetaXml.replace('{{FILE5_HASH}}', File5_sha256)
+MetaXml = MetaXml.replace('{{FILE5_SIZE}}', File5_size)
+MetaXml = MetaXml.replace('{{SRV_HOST}}', srv_host)
+MetaXml = MetaXml.replace('{{SRV_PORT}}', str (srv_port))
+MetaFile.content = MetaXml
+
+err = http_test.begin ()
+
+exit (err)
-- 
2.7.3