Fix parsing of properties containing [topic/parse_property

For example, the name extracted from a requirement of
"python3.10dist(fonttools[unicode])[>= 4.10]" was sometimes
"python3.10dist(fonttools" instead of the expected
"python3.10dist(fonttools[unicode])".

Code parsing such strings existed in many places, it now exists only
in 2 places, a perl version in Resolve.pm and a C version in URPM.xs.

Both codes used to handle both "foo >= 0" and "foo[>= 0]" but at least
the perl code seems to only call it on provides/conflicts/obsoletes
which are always using the second form so the support for it was
dropped from the perl version for the sake of simplicity.

4 files changed, 268 insertions, 79 deletions

diff --git a/URPM.xs b/URPM.xs
index 7d48acf..f1e37dd 100644
--- a/URPM.xs
+++ b/URPM.xs
@@ -309,6 +309,73 @@ ranges_overlap(rpmsenseFlags aflags, char *sa, rpmsenseFlags bflags, char *sb) {
   }
 }
 
+struct property {
+  char *name;
+  rpmsenseFlags flags;
+  char *evr;
+};
+
+static void free_property(struct property *ps) {
+  free(ps->name);
+  ps->name = NULL;
+  ps->flags = 0;
+  free(ps->evr);
+  ps->evr = NULL;
+}
+
+// This parses things like 'ocamlx(bar)[== 42]' or 'ocaml-bar < 42'
+static int parse_property(char *s, struct property *ps) {
+  int l = strlen(s);
+  char *eon = NULL;
+
+  ps->name = NULL;
+  ps->flags = 0;
+  ps->evr = NULL;
+
+  if (l == 0) return -1;
+
+  if (s[l-1] == ']') {
+    eon = strrchr(s, '[');
+
+    if (eon == NULL || eon == s) {
+      // This is a strange one finishing with ] but without a [ before like
+      // "Provides: foo]" or something without a name before like
+      // "Provides: [foo]"
+      ps->name = strdup(s);
+      ps->evr = strdup("");
+      return 0;
+    }
+  }
+
+  /* Drop "[*]" if present at the end of the name */
+  if (eon-s > 3 && !strncmp(eon-3, "[*]", 3)) {
+    eon = eon-3;
+  }
+
+  if (eon == NULL) {
+    // This does not have [ ] at the end but could be in the form "gcc < 42"
+    eon = s;
+    while (*eon && *eon != ' ' && *eon != '<' && *eon != '>') ++eon;
+  }
+
+  ps->name = strndup(s, eon-s);
+  l = strlen(eon);
+  s = eon;
+  while (*s) {
+    if (*s == ' ' || *s == '[' || *s == '*' || *s == ']');
+    else if (*s == '<') ps->flags |= RPMSENSE_LESS;
+    else if (*s == '>') ps->flags |= RPMSENSE_GREATER;
+    else if (*s == '=') ps->flags |= RPMSENSE_EQUAL;
+    else break;
+    ++s;
+    --l;
+  }
+  ps->evr = strndup(s, l-1);
+  /* fprintf(stderr, "name=\"%s\" evr=\"%s\"\n", ps->name, ps->evr); */
+
+  return 0;
+}
+
 typedef int (*callback_list_str)(char *s, int slen, const char *name, const rpmsenseFlags flags, const char *evr, void *param);
 
 static int
@@ -339,9 +406,8 @@ callback_list_str_overlap(char *s, int slen, const char *name, rpmsenseFlags fla
   struct cb_overlap_s *os = (struct cb_overlap_s *)param;
   int result = 0;
   char *eos = NULL;
-  char *eon = NULL;
   char eosc = '\0';
-  char eonc = '\0';
+  struct property ops;
 
   /* we need to extract name, flags and evr from a full sense information, store result in local copy */
   if (s) {
@@ -350,28 +416,16 @@ callback_list_str_overlap(char *s, int slen, const char *name, rpmsenseFlags fla
       eosc = *eos;
       *eos = 0;
     }
-    name = s;
-    while (*s && *s != ' ' && *s != '[' && *s != '<' && *s != '>' && *s != '=') ++s;
-    if (*s) {
-      eon = s;
-      while (*s) {
-	if (*s == ' ' || *s == '[' || *s == '*' || *s == ']');
-	else if (*s == '<') flags |= RPMSENSE_LESS;
-	else if (*s == '>') flags |= RPMSENSE_GREATER;
-	else if (*s == '=') flags |= RPMSENSE_EQUAL;
-	else break;
-	++s;
-      }
-      evr = s;
-    } else
+    if (!parse_property(s, &ops)) {
+      name = ops.name;
+      flags = ops.flags;
+      evr = ops.evr;
+    } else {
+      name = s;
       evr = "";
+    }
   }
 
-  /* mark end of name */
-  if (eon) {
-       eonc = *eon;
-       *eon = 0;
-  }
   /* names should be equal, else it will not overlap */
   if (!strcmp(name, os->name)) {
     /* perform overlap according to direction needed, negative for left */
@@ -384,9 +438,12 @@ callback_list_str_overlap(char *s, int slen, const char *name, rpmsenseFlags fla
   /* fprintf(stderr, "cb_list_str_overlap result=%d, os->direction=%d, os->name=%s, os->evr=%s, name=%s, evr=%s\n",
      result, os->direction, os->name, os->evr, name, evr); */
 
-  /* restore s if needed */
-  if (eon) *eon = eonc;
-  if (eos) *eos = eosc;
+  if (s) {
+    free_property(&ops);
+
+    /* restore s if needed */
+    if (eos) *eos = eosc;
+  }
 
   return result;
 }
@@ -408,21 +465,29 @@ return_list_str(char *s, const Header header, rpmTag tag_name, rpmTag tag_flags,
       if (f(s, 0, NULL, 0, NULL, param))
         return -count;
     } else {
-      char *eos;
+      struct property ops;
       while(ps != NULL) {
-	*ps = 0; eos = strchr(s, '['); if (!eos) eos = strchr(s, ' ');
-	++count;
-	if (f(s, eos ? eos-s : ps-s, NULL, 0, NULL, param)) {
-          *ps = '@';
-          return -count;
+        *ps = 0;
+        if (!parse_property(s, &ops)) {
+          ++count;
+	  if (f(ops.name, strlen(ops.name), NULL, 0, NULL, param)) {
+            *ps = '@';
+            free_property(&ops);
+            return -count;
+          }
+          free_property(&ops);
         }
-	*ps = '@'; /* restore in memory modified char */
-	s = ps + 1; ps = strchr(s, '@');
+        *ps = '@'; /* restore in memory modified char */
+        s = ps + 1; ps = strchr(s, '@');
       }
-      eos = strchr(s, '['); if (!eos) eos = strchr(s, ' ');
       ++count;
-      if (f(s, eos ? eos-s : 0, NULL, 0, NULL, param))
-         return -count;
+      if (!parse_property(s, &ops)) {
+        if (f(ops.name, strlen(ops.name), NULL, 0, NULL, param)) {
+           free_property(&ops);
+           return -count;
+        }
+        free_property(&ops);
+      }
     }
   } else if (header) {
     struct rpmtd_s list, flags, list_evr;
@@ -879,6 +944,7 @@ update_provides(const URPM__Package pkg, HV *provides) {
     }
   } else {
     char *ps, *s, *es;
+    struct property ops;
 
     if ((s = pkg->requires) != NULL && *s != 0) {
       ps = strchr(s, '@');
@@ -900,12 +966,24 @@ update_provides(const URPM__Package pkg, HV *provides) {
     if ((s = pkg->provides) != NULL && *s != 0) {
       ps = strchr(s, '@');
       while(ps != NULL) {
-	*ps = 0; es = strchr(s, '['); if (!es) es = strchr(s, ' '); *ps = '@';
-	update_hash_entry(provides, s, es != NULL ? es-s : ps-s, 1, es != NULL, pkg);
+	*ps = 0;
+	if (parse_property(s, &ops)) {
+	  /* Failed to parse, use the whole string */
+	  update_hash_entry(provides, s, ps-s, 1, 0, pkg);
+	} else {
+	  update_hash_entry(provides, s, strlen(ops.name), 1, ops.flags != 0, pkg);
+	  free_property(&ops);
+	}
+        *ps = '@';
 	s = ps + 1; ps = strchr(s, '@');
       }
-      es = strchr(s, '['); if (!es) es = strchr(s, ' ');
-      update_hash_entry(provides, s, es != NULL ? es-s : 0, 1, es != NULL, pkg);
+      if (parse_property(s, &ops)) {
+	  /* Failed to parse, use the whole string */
+	  update_hash_entry(provides, s, ps-s, 1, 0, pkg);
+	} else {
+	  update_hash_entry(provides, s, strlen(ops.name), 1, ops.flags != 0, pkg);
+	  free_property(&ops);
+	}
     }
   }
 }
@@ -925,16 +1003,28 @@ update_obsoletes(const URPM__Package pkg, HV *obsoletes) {
     char *ps, *s;
 
     if ((s = pkg->obsoletes) != NULL && *s != 0) {
-      char *es;
+      struct property ops;
 
       ps = strchr(s, '@');
       while(ps != NULL) {
-	*ps = 0; es = strchr(s, '['); if (!es) es = strchr(s, ' '); *ps = '@';
-	update_hash_entry(obsoletes, s, es != NULL ? es-s : ps-s, 1, 0, pkg);
+	*ps = 0;
+	if (parse_property(s, &ops)) {
+	  /* Failed to parse, use the whole string */
+	  update_hash_entry(obsoletes, s, ps-s, 1, 0, pkg);
+	} else {
+	  update_hash_entry(obsoletes, s, strlen(ops.name), 1, 0, pkg);
+	  free_property(&ops);
+	}
+        *ps = '@';
 	s = ps + 1; ps = strchr(s, '@');
       }
-      es = strchr(s, '['); if (!es) es = strchr(s, ' ');
-      update_hash_entry(obsoletes, s, es != NULL ? es-s : 0, 1, 0, pkg);
+      if (parse_property(s, &ops)) {
+        /* Failed to parse, use the whole string */
+        update_hash_entry(obsoletes, s, strlen(s), 1, 0, pkg);
+      } else {
+        update_hash_entry(obsoletes, s, ops.flags ? strlen(ops.name) : 0, 1, 0, pkg);
+        free_property(&ops);
+      }
     }
   }
 }
@@ -1507,7 +1597,6 @@ static int get_e_v_r(URPM__Package pkg, int *epoch, char **version, char **relea
     return 0;
 }
 
-
 MODULE = URPM            PACKAGE = URPM::Package       PREFIX = Pkg_
 
 void
@@ -1991,8 +2080,7 @@ Pkg_obsoletes_overlap(pkg, s)
      provides_overlap = 1
   PREINIT:
   struct cb_overlap_s os;
-  char *eon = NULL;
-  char eonc = '\0';
+  struct property ps;
   rpmTag tag_name;
   rpmTag tag_flags, tag_version;
   CODE:
@@ -2008,33 +2096,20 @@ Pkg_obsoletes_overlap(pkg, s)
        tag_version = RPMTAG_OBSOLETEVERSION;
        break;
   }
-  os.name = s;
-  os.flags = 0;
-  while (*s && *s != ' ' && *s != '[' && *s != '<' && *s != '>' && *s != '=') ++s;
-  if (*s) {
-    eon = s;
-    while (*s) {
-      if (*s == ' ' || *s == '[' || *s == '*' || *s == ']');
-      else if (*s == '<') os.flags |= RPMSENSE_LESS;
-      else if (*s == '>') os.flags |= RPMSENSE_GREATER;
-      else if (*s == '=') os.flags |= RPMSENSE_EQUAL;
-      else break;
-      ++s;
-    }
-    os.evr = s;
-  } else
-    os.evr = "";
-  os.direction = ix == 0 ? -1 : 1;
-  /* mark end of name */
-  if (eon) {
-    eonc = *eon;
-    *eon = 0;
-  }
-  /* return_list_str returns a negative value is the callback has returned non-zero */
-  RETVAL = return_list_str(ix == 0 ? pkg->obsoletes : pkg->provides, pkg->h, tag_name, tag_flags, tag_version,
-			   callback_list_str_overlap, &os) < 0;
-  /* restore end of name */
-  if (eon) *eon = eonc;
+  if (parse_property(s, &ps)) {
+    fprintf(stderr, "provides_overlap: failed to parse property %s\n", s);
+    RETVAL = -1;
+  } else {
+    os.name = ps.name;
+    os.flags = ps.flags;
+    os.evr = ps.evr;
+    os.direction = ix == 0 ? -1 : 1;
+    fprintf(stderr, "provides_overlap: property=\"%s\" name=\"%s\" evr=\"%s\"\n", s, os.name, os.evr);
+    /* return_list_str returns a negative value is the callback has returned non-zero */
+    RETVAL = return_list_str(ix == 0 ? pkg->obsoletes : pkg->provides, pkg->h, tag_name, tag_flags, tag_version,
+			     callback_list_str_overlap, &os) < 0;
+    free_property(&ps);
+  }
   OUTPUT:
   RETVAL
 
diff --git a/URPM/Resolve.pm b/URPM/Resolve.pm
index dded564..1816eb1 100644
--- a/URPM/Resolve.pm
+++ b/URPM/Resolve.pm
@@ -58,7 +58,12 @@ Returns the property name (eg: "mageia-release" in above example)
 
 sub property2name {
     my ($property) = @_;
-    $property =~ /^([^\s\[]*)/ && $1;
+    my $name = $property;
+    if ($property =~ /^([^\s]+)\s*\[[^\[]*\]$/) {
+	$name = $1;
+    }
+    $name =~ s/\[\*\]$//;
+    return $name;
 }
 
 =item property2name_range($property)
@@ -69,7 +74,14 @@ Returns the property name & range (eg: "mageia-release" & ">= 1" in above exampl
 
 sub property2name_range {
     my ($property) = @_;
-    $property =~ /^([^\s\[]*)(?:\[\*\])?\[?([^\s\]]*\s*[^\s\]]*)/;
+    my ($name, $range) = ($property, "");
+    $name =~ s/\[\*\]$//;
+    if ($name =~ /^([^\s]+)\s*\[([^\[]*)\]$/) {
+	$name = $1;
+	$range = $2;
+    }
+    $name =~ s/\[\*\]$//;
+    return ($name, $range);
 }
 
 =item property2name_op_version($property)
@@ -81,7 +93,15 @@ Returns the property name, operator & range (eg: "mageia-release", ">=", & "1" i
 
 sub property2name_op_version {
     my ($property) = @_;
-    $property =~ /^([^\s\[]*)(?:\[\*\])?\s*\[?([^\s\]]*)\s*([^\s\]]*)/;
+    my ($name, $op, $version) = ($property, "", "");
+    $name =~ s/\[\*\]$//;
+    if ($name =~ /^([^\s]+)\s*\[([^\s\[]*)\s*([^\s\]]*)\]$/) {
+	$name = $1;
+	$op = $2;
+	$version = $3;
+    }
+    $name =~ s/\[\*\]$//;
+    return ($name, $op, $version);
 }
 
 
diff --git a/t/resolve.t b/t/resolve.t
new file mode 100644
index 0000000..56869b1
--- /dev/null
+++ b/t/resolve.t
@@ -0,0 +1,46 @@
+#!/usr/bin/perl
+
+use strict ;
+use warnings ;
+use Test::More;
+use URPM::Resolve;
+
+my $testcases = [
+    {
+	property => 'mageia-release[>= 1]',
+	name => 'mageia-release',
+	op => '>=',
+	version => '1',
+    }, {
+	property => 'python3.10dist(fonttools[unicode])[>= 4.10]',
+	name => 'python3.10dist(fonttools[unicode])',
+	op => '>=',
+	version => '4.10',
+    }, {
+	property => 'python3.10dist(fonttools[unicode])',
+	name => 'python3.10dist(fonttools[unicode])',
+	op => '',
+	version => '',
+    }, {
+	property => 'openssl[*][>= 0.9.7]',
+	name => 'openssl',
+	op => '>=',
+	version => '0.9.7',
+    }, {
+	property => 'openssl[*]',
+	name => 'openssl',
+	op => '',
+	version => '',
+    }];
+
+foreach my $tc (@$testcases) {
+    is(URPM::property2name($tc->{property}), $tc->{name}, "property2name(\"$tc->{property}\")");
+    my $expected = [$tc->{name}, $tc->{op} ? "$tc->{op} $tc->{version}" : ""];
+    my @got = URPM::property2name_range($tc->{property});
+    is_deeply(\@got, $expected, "property2name_range(\"$tc->{property}\")");
+    $expected = [$tc->{name}, $tc->{op}, $tc->{version}];
+    @got = URPM::property2name_op_version($tc->{property});
+    is_deeply(\@got, $expected, "property2name_op_version(\"$tc->{property}\")");
+}
+
+done_testing();
diff --git a/t/synthesis.t b/t/synthesis.t
index 8246b36..794fd7a 100644
--- a/t/synthesis.t
+++ b/t/synthesis.t
@@ -2,7 +2,7 @@
 
 use strict ;
 use warnings ;
-use Test::More tests => 95;
+use Test::More tests => 113;
 use URPM;
 
 chdir 't' if -d 't';
@@ -27,7 +27,25 @@ $s =~ s/-devel//g;
 print $f $s;
 close $f;
 
-END { unlink $file1, $file2 }
+$s = <<'EOF';
+@provides@python3-fonttools+ufo[== 4.18.2-4.mga9]@python3dist(fonttools[ufo])[== 4.18.2]
+@summary@Metapackage for python3-fonttools: ufo extras
+@info@python3-fonttools+ufo-4.18.2-4.mga9.noarch@0@221683@Development/Other
+@provides@config(dovecot)[== 2.3.19.1-2.mga9]
+@requires@openssl[*]@rpm-helper[*][>= 0.21]
+@summary@Secure IMAP and POP3 server
+@info@dovecot-2.3.19.1-2.mga9.x86_64@0@17815993@System/Servers
+@provides@openimageio[== 2.3.20.0-2.mga9]@font(:lang=aa)
+@summary@Library for reading and writing images
+@info@openimageio-2.3.20.0-2.mga9.x86_64@0@42@Development/Other
+EOF
+
+my $file3 = 'synthesis.squarebrackets.cz';
+open $f, "| gzip -9 >$file3";
+print $f $s;
+close $f;
+
+END { unlink $file1, $file2, $file3 }
 
 my $a = URPM->new;
 ok($a);
@@ -161,3 +179,33 @@ ok($requires[8] eq 'rpmlib(CompressedFileNames)');
 @provides = $pkg->provides_nosense;
 ok(@provides == 1);
 ok($provides[0] eq 'glibc-devel');
+
+my $b = URPM->new;
+$b->parse_synthesis($file3);
+$pkg = $b->{depslist}[0];
+ok($pkg);
+is($pkg->name, 'python3-fonttools+ufo');
+@provides = $pkg->provides;
+is(@provides, 2);
+is($provides[1], 'python3dist(fonttools[ufo])[== 4.18.2]');
+@provides = $pkg->provides_nosense;
+is(@provides, 2);
+is($provides[1], 'python3dist(fonttools[ufo])');
+$pkg = $b->{depslist}[1];
+ok($pkg);
+is($pkg->name, 'dovecot');
+@requires = $pkg->requires;
+is(@requires, 2);
+is($requires[0], 'openssl[*]');
+is($requires[1], 'rpm-helper[*][>= 0.21]');
+@requires = $pkg->requires_nosense;
+is(@requires, 2);
+is($requires[0], 'openssl');
+is($requires[1], 'rpm-helper');
+$pkg = $b->{depslist}[2];
+ok($pkg);
+is($pkg->name, 'openimageio');
+@provides = $pkg->provides;
+is(@provides, 2);
+is($provides[1], 'font(:lang=aa)');
+