add a Timingtests target to check bad behaviour from the streaming engine

* Makefile.am: add a Timingtests target to check bad behaviour
  from the streaming engine
* dbgen.pl dbgenattr.pl: perl script to generate big instances
* xmlreader.c: fix a bad behaviour on large buffer inputs
Daniel

ChangeLog

@@ -1,3 +1,10 @@
+Thu Sep 18 00:31:02 CEST 2003 Daniel Veillard <daniel@veillard.com>
+
+	* Makefile.am: add a Timingtests target to check bad behaviour
+	  from the streaming engine
+	* dbgen.pl dbgenattr.pl: perl script to generate big instances
+	* xmlreader.c: fix a bad behaviour on large buffer inputs
+
 Wed Sep 17 23:25:47 CEST 2003 Daniel Veillard <daniel@veillard.com>
 
 	* dict.c xmlreader.c: two small improvements

Makefile.am

@@ -120,7 +120,7 @@ check-local: tests
 
 testall : tests SVGtests SAXtests
 
-tests: XMLtests XMLenttests NStests Readertests SAXtests HTMLtests Validtests URItests XPathtests XPtrtests XIncludetests C14Ntests Scripttests Catatests @TEST_REGEXPS@ @TEST_SCHEMAS@ @TEST_THREADS@
+tests: XMLtests XMLenttests NStests Readertests SAXtests HTMLtests Validtests URItests XPathtests XPtrtests XIncludetests C14Ntests Scripttests Catatests @TEST_REGEXPS@ @TEST_SCHEMAS@ @TEST_THREADS@ Timingtests
 	@(if [ "@PYTHON_SUBDIR@" != "" ] ; then cd python ; $(MAKE) tests ; fi)
 
 valgrind:
@@ -699,6 +699,24 @@ Automatatests: testAutomata$(EXEEXT)
 	      rm result.$$name ; \
 	  fi ; fi ; done)
 
+$(srcdir)/dba100000.xml: dbgenattr.pl
+	@echo "## generating dba100000.xml"
+	@(dbgenattr.pl 100000 > $(srcdir)/dba100000.xml)
+
+Timingtests: $(srcdir)/dba100000.xml
+	@echo "##"
+	@echo "## Timing tests to try to detect performance"
+	@echo "## as well a memory usage breakage when streaming"
+	@echo "## first when using the file interface"
+	@echo "## then when using the memory interface"
+	@echo "##"
+	-@(xmllint --stream --timing $(srcdir)/dba100000.xml; \
+	   grep "MORY ALLO" .memdump  | grep -v "MEMORY ALLOCATED : 0";\
+	   exit 0)
+	-@(xmllint --stream --timing --memory $(srcdir)/dba100000.xml; \
+	   grep "MORY ALLO" .memdump  | grep -v "MEMORY ALLOCATED : 0";\
+	   exit 0)
+
 C14Ntests : testC14N$(EXEEXT)
 	@echo "##"
 	@echo "## C14N and XPath regression tests"
@@ -869,7 +887,9 @@ EXTRA_DIST = xml2-config.in xml2Conf.sh.in libxml.spec.in libxml2.spec \
 	     $(man_MANS) libxml-2.0.pc.in \
 	     trionan.c trionan.h triostr.c triostr.h trio.c trio.h \
 	     triop.h triodef.h libxml.h \
-	     testThreadsWin32.c genUnicode.py TODO_SCHEMAS
+	     testThreadsWin32.c genUnicode.py TODO_SCHEMAS \
+	     dbgen.pl dbgenattr.pl
+
 
 pkgconfigdir = $(libdir)/pkgconfig
 pkgconfig_DATA = libxml-2.0.pc

dbgen.pl

@@ -0,0 +1,43 @@
+#!/usr/bin/perl
+
+$size = shift;
+
+if ($size eq "") 
+{
+    die "usage:  dbgen.pl [size]\n";
+}
+
+@firstnames = ("Al", "Bob", "Charles", "David", "Egon", "Farbood", 
+               "George", "Hank", "Inki", "James");
+@lastnames = ("Aranow", "Barker", "Corsetti", "Dershowitz", "Engleman", 
+              "Franklin", "Grice", "Haverford", "Ilvedson", "Jones");
+@states = ("AL", "AK", "AZ", "AR", "CA", "CO", "CT", "DE", "FL", "GA", 
+           "HI", "ID", "IL", "IN", "IA", "KS", "KY", "LA", "ME", "MD", 
+           "MA", "MI", "MN", "MS", "MO", "MT", "NE", "NV", "NH", "NJ", 
+           "NM", "NY", "NC", "ND", "OH", "OK", "OR", "PA", "RI", "SC", 
+           "SD", "TN", "TX", "UT", "VT", "VA", "WA", "WV", "WI", "WY");
+
+print "<?xml version=\"1.0\"?>\n";
+print "\n";
+print "<table>\n";
+
+for ($i=0; $i<$size; $i++)
+{
+    $first = $firstnames [$i % 10];
+    $last = $lastnames [($i / 10) % 10];
+    $state = $states [($i / 100) % 50];
+    $zip = 22000 + $i / 5000;
+
+    printf "  <row>\n";
+    printf "    <id>%04d</id>\n", $i;
+    printf "    <firstname>$first</firstname>\n", $i;
+    printf "    <lastname>$last</lastname>\n", $i;
+    printf "    <street>%d Any St.</street>\n", ($i % 100) + 1;
+    printf "    <city>Anytown</city>\n";
+    printf "    <state>$state</state>\n";
+    printf "    <zip>%d</zip>\n", $zip;
+    printf "  </row>\n";
+}
+
+print "</table>\n";
+

dbgenattr.pl

@@ -0,0 +1,42 @@
+#!/usr/bin/perl
+
+$size = shift;
+
+if ($size eq "") 
+{
+    die "usage:  dbgen.pl [size]\n";
+}
+
+@firstnames = ("Al", "Bob", "Charles", "David", "Egon", "Farbood", 
+               "George", "Hank", "Inki", "James");
+@lastnames = ("Aranow", "Barker", "Corsetti", "Dershowitz", "Engleman", 
+              "Franklin", "Grice", "Haverford", "Ilvedson", "Jones");
+@states = ("AL", "AK", "AZ", "AR", "CA", "CO", "CT", "DE", "FL", "GA", 
+           "HI", "ID", "IL", "IN", "IA", "KS", "KY", "LA", "ME", "MD", 
+           "MA", "MI", "MN", "MS", "MO", "MT", "NE", "NV", "NH", "NJ", 
+           "NM", "NY", "NC", "ND", "OH", "OK", "OR", "PA", "RI", "SC", 
+           "SD", "TN", "TX", "UT", "VT", "VA", "WA", "WV", "WI", "WY");
+
+print "<?xml version=\"1.0\"?>\n";
+print "\n";
+print "<table>\n";
+
+for ($i=0; $i<$size; $i++)
+{
+    $first = $firstnames [$i % 10];
+    $last = $lastnames [($i / 10) % 10];
+    $state = $states [($i / 100) % 50];
+    $zip = 22000 + $i / 5000;
+
+    printf "  <row\n";
+    printf "    id='%04d'\n", $i;
+    printf "    firstname='$first'\n", $i;
+    printf "    lastname='$last'\n", $i;
+    printf "    street='%d Any St.'\n", ($i % 100) + 1;
+    printf "    city='Anytown'\n";
+    printf "    state='$state'\n";
+    printf "    zip='%d'/>\n", $zip;
+}
+
+print "</table>\n";
+

xmlreader.c

@@ -803,7 +803,8 @@ xmlTextReaderPushData(xmlTextReaderPtr reader) {
      * Discard the consumed input when needed and possible
      */
     if (reader->mode == XML_TEXTREADER_MODE_INTERACTIVE) {
-	if (reader->cur >= 4096) {
+	if ((reader->cur >= 4096) &&
+	    (inbuf->use - reader->cur <= CHUNK_SIZE)) {
 	    val = xmlBufferShrink(inbuf, reader->cur);
 	    if (val >= 0) {
 		reader->cur -= val;