add checking sort -V

Makefile

@@ -1,5 +1,14 @@
 testinput: run.py
-	./run.py > testinput
+	./run.py /var/lib/apt/lists/ftp.debian.org_debian_dists_sid_main_binary-amd64_Packages > testinput
+
+testinput.1000: run.py
+	./run.py /var/lib/apt/lists/ftp.debian.org_debian_dists_sid_main_binary-amd64_Packages | awk 'NR % 1000 == 1' > testinput.1000
+
+testinput.unequal: testinput inequalityfilter
+	./inequalityfilter < testinput > testinput.unequal
+
+testinput.unequal.1000: testinput inequalityfilter
+	./inequalityfilter < testinput | awk 'NR % 1000 == 1' > testinput.unequal.1000
 
 testdpkg: testdpkg.c
 	gcc testdpkg.c -ldpkg -o testdpkg
@@ -10,12 +19,31 @@ testapt: testapt.cc
 testdose: testdose.ml
 	ocamlfind ocamlc -package dose3 -linkpkg testdose.ml -o testdose
 
+inequalityfilter: inequalityfilter.cc
+	g++ inequalityfilter.cc -lapt-pkg -o inequalityfilter
+
 .PHONY: test
-test: testinput testdpkg testapt
+test: testinput testdpkg testapt testdose rundpkg.sh
 	./rundpkg.sh < testinput | md5sum
 	./testapt < testinput | md5sum
 	./testdose < testinput | md5sum
 
+.PHONY: test.1000
+test.1000: testinput.1000 testdpkg testapt testdose rundpkg.sh
+	./rundpkg.sh < testinput.1000 | md5sum
+	./testapt < testinput.1000 | md5sum
+	./testdose < testinput.1000 | md5sum
+
+.PHONY: testsort
+testsort: testinput.unequal testapt testsort.sh
+	./testapt < testinput.unequal | md5sum
+	./testsort.sh < testinput.unequal | md5sum
+
+.PHONY: testsort.1000
+testsort.1000: testinput.unequal.1000 testapt testsort.sh
+	./testapt < testinput.unequal.1000 | md5sum
+	./testsort.sh < testinput.unequal.1000 | md5sum
+
 .PHONY: clean
 clean:
-	rm -f testapt testdose testdose.cmi testdose.cmo testdpkg testinput
+	rm -f testapt testdose testdose.cmi testdose.cmo testdpkg testinput inequalityfilter

inequalityfilter.cc

@@ -0,0 +1,41 @@
+#include <apt-pkg/debversion.h>
+#include <stdio.h>
+#include <stdlib.h>
+#include <string.h>
+
+int main()
+{
+    char *line = NULL;
+    char *token;
+    char *orig1;
+    char *orig2;
+    char *ver1;
+    char *ver2;
+    size_t len = 0;
+    size_t read;
+    int ret;
+    while ((read = getline(&line, &len, stdin)) != -1) {
+        //fprintf(stderr, "%s", line);
+        orig1 = strdup(line);
+        orig2 = strdup(line);
+        token = orig1;
+        ver1 = strsep(&token, "\t");
+        if (ver1 == NULL) {
+            fprintf(stderr, "cannot read token1");
+            exit(EXIT_FAILURE);
+        }
+        ver2 = strsep(&token, "\n");
+        if (ver2 == NULL) {
+            fprintf(stderr, "cannot read token2");
+            exit(EXIT_FAILURE);
+        }
+        ret = debVS.CmpVersion(ver1, ver2);
+        if (ret != 0) {
+            fputs(orig2, stdout);
+        }
+        //fputs(orig2, stdout);
+        free(orig1);
+        free(orig2);
+    }
+    exit(EXIT_SUCCESS);
+}

run.py

@@ -1,10 +1,12 @@
 #!/usr/bin/env python
 
+from __future__ import print_function
 import deb822
+import sys
 
 versions = set()
-with open("/home/josch/gsoc2012/bootstrap/Packages") as f:
-    for pkg in deb822.Deb822.iter_paragraphs(f, use_apt_pkg=False):
+with open(sys.argv[1]) as f:
+    for pkg in deb822.Deb822.iter_paragraphs(f):
         ver = pkg.get('Version')
         if ver:
             versions.add(ver)
@@ -13,6 +15,8 @@ with open("/home/josch/gsoc2012/bootstrap/Packages") as f:
 # through the list of versions and comparing every version with all that come
 # after it
 versions = sorted(versions)
+l = len(versions)
 for i,v1 in enumerate(versions):
+    print("%f %%\r"%((i*100.0)/l), file=sys.stderr, end="")
     for v2 in versions[i+1:]:
         print("%s\t%s"%(v1,v2))

testapt.cc

@@ -37,6 +37,10 @@ int main()
         } else {
             fputc('<', stdout);
         }
+        /*
+        fputc(' ', stdout);
+        fputs(orig2, stdout);
+        */
         //fputs(orig2, stdout);
         free(orig1);
         free(orig2);

testsort.sh

@@ -0,0 +1,29 @@
+#!/bin/sh
+
+# some examples where `sort -V` does the wrong thing:
+# 0.04-1-5             0.4-3
+# 0.1+20080921-2       0.1-1+b8
+# 0.1+dfsg-4           000.001-4
+# 0.1.0+20071012-1.2   0.1.0-1.1
+# 0.3.0+20091229-1     0.3.0-7.2
+# 0.6.0+git20130305-5  0.6.0-5+b1
+# 0:2009.10.04-1       1.0pre11-1
+# 0:2009.10.04-1       1.240-1
+# 0:2009.10.04-1       1.7.2.4-4.1
+# 0:2009.10.04-1       2.3-12
+# 0:2009.10.04-1       8.0.184.15484+dfsg-2
+# 1.0+dfsg-1           1.00-6
+
+i=0
+while read line; do
+	printf "$i\r" >&2
+	i=$((i+1))
+	set -- $line
+	newest=$( ( echo "$1"; echo "$2" ) | sort -V | tail -n1)
+	if [ "$1" != "$newest" ]; then
+		printf "<"
+	else
+		printf ">"
+	fi
+	printf " %s\t%s\n" $1 $2
+done