From: Michael Schroeder <mls@suse.de>
Date: Mon, 20 Sep 2021 13:10:38 +0000 (+0200)
Subject: Fix misparsing of '&' in attributes with libxml2
X-Git-Tag: 0.7.20~1
X-Git-Url: http://git.ipfire.org/cgi-bin/gitweb.cgi?a=commitdiff_plain;h=9a13593b43bd883be6bf4260edbfc8eaf766ffa8;p=thirdparty%2Flibsolv.git

Fix misparsing of '&' in attributes with libxml2

For some reason libxml2 insists on re-escaping '&' characters
by substituting them with "&#38;". The rpm dependencies are
stored in attributes, so this is not an academic matter.

Undo the damage done by libxml2 by replacing all &#38; occurences
by a single '&'.
---

diff --git a/ext/solv_xmlparser.c b/ext/solv_xmlparser.c
index 6292663c..87bd0969 100644
--- a/ext/solv_xmlparser.c
+++ b/ext/solv_xmlparser.c
@@ -53,6 +53,47 @@ character_data(void *userData, const XML_Char *s, int len)
   xmlp->lcontent += len; 
 }
 
+#ifdef WITH_LIBXML2
+static void fixup_att_inplace(char *at)
+{
+  while ((at = strchr(at, '&')) != 0)
+    {
+      at++;
+      if (!memcmp(at, "#38;", 4))
+	memmove(at, at + 4, strlen(at + 4) + 1);
+    }
+}
+
+static const xmlChar **fixup_atts(struct solv_xmlparser *xmlp, const xmlChar **atts)
+{
+  size_t needsize = 0;
+  size_t natts;
+  char **at;
+
+  for (natts = 0; atts[natts]; natts++)
+    if (strchr((char *)atts[natts], '&'))
+      needsize += strlen((const char *)atts[natts]) + 1;
+  if (!needsize)
+    return atts;
+  at = xmlp->attsdata = solv_realloc(xmlp->attsdata, (natts + 1) * sizeof(xmlChar *) + needsize);
+  needsize = (natts + 1) * sizeof(xmlChar *);
+  for (natts = 0; atts[natts]; natts++)
+    {
+      at[natts] = (char *)atts[natts];
+      if (strchr(at[natts], '&'))
+        {
+	  size_t l = strlen(at[natts]) + 1;
+	  memcpy((char *)at + needsize, at[natts], l);
+	  at[natts] = (char *)at + needsize;
+	  needsize += l;
+	  fixup_att_inplace(at[natts]);
+        }
+    }
+  at[natts] = 0;
+  return (const xmlChar **)at;
+}
+#endif
+
 #ifdef WITH_LIBXML2
 static void
 start_element(void *userData, const xmlChar *name, const xmlChar **atts)
@@ -97,6 +138,8 @@ start_element(void *userData, const char *name, const char **atts)
       static const char *nullattr;
       atts = (const xmlChar **)&nullattr;
     }
+  else if (xmlp->state != oldstate)
+    atts = fixup_atts(xmlp, atts);
 #endif
   if (xmlp->state != oldstate)
     xmlp->startelement(xmlp, xmlp->state, el->element, (const char **)atts);
@@ -177,6 +220,7 @@ solv_xmlparser_free(struct solv_xmlparser *xmlp)
   queue_free(&xmlp->elementq);
   xmlp->content = solv_free(xmlp->content);
   xmlp->errstr = solv_free(xmlp->errstr);
+  xmlp->attsdata = solv_free(xmlp->attsdata);
 }
 
 static void
diff --git a/ext/solv_xmlparser.h b/ext/solv_xmlparser.h
index ced0571f..717983f1 100644
--- a/ext/solv_xmlparser.h
+++ b/ext/solv_xmlparser.h
@@ -30,6 +30,7 @@ struct solv_xmlparser {
 
   Id *elementhelper;
   void *parser;
+  void *attsdata;
 };
 
 #define SOLV_XMLPARSER_OK	0