Introduce tolower_ascii_unsafe() and use it for hashing

This one has some obvious collisions for non-alphabetical characters,
like some control characters also hashing to numbers, but we don't
really have those, and these are hash functions which are not
collision free to begin with.

3 files changed, 9 insertions, 4 deletions

diff --git a/apt-pkg/contrib/strutl.h b/apt-pkg/contrib/strutl.h
index f3591d65f..918ac89c7 100644
--- a/apt-pkg/contrib/strutl.h
+++ b/apt-pkg/contrib/strutl.h
@@ -141,6 +141,11 @@ bool CheckDomainList(const std::string &Host, const std::string &List);
 #define isspace_ascii  isspace_ascii_inline
 
 APT_CONST APT_HOT
+static inline int tolower_ascii_unsafe(int const c)
+{
+   return c | 0x20;
+}
+APT_CONST APT_HOT
 static inline int tolower_ascii_inline(int const c)
 {
    return (c >= 'A' && c <= 'Z') ? c + 32 : c;
diff --git a/apt-pkg/deb/deblistparser.cc b/apt-pkg/deb/deblistparser.cc
index 43fc4aa3a..549e75952 100644
--- a/apt-pkg/deb/deblistparser.cc
+++ b/apt-pkg/deb/deblistparser.cc
@@ -369,7 +369,7 @@ unsigned short debListParser::VersionHash()
       {
 	 if (isspace_ascii(*Start) != 0)
 	    continue;
-	 *J++ = tolower_ascii(*Start);
+	 *J++ = tolower_ascii_unsafe(*Start);
 
 	 /* Normalize <= to < and >= to >. This is the wrong way around, but
 	  * more efficient that the right way. And since we're only hashing
diff --git a/apt-pkg/pkgcache.cc b/apt-pkg/pkgcache.cc
index e7e417a5a..b0ba1597f 100644
--- a/apt-pkg/pkgcache.cc
+++ b/apt-pkg/pkgcache.cc
@@ -213,14 +213,14 @@ map_id_t pkgCache::sHash(StringView Str) const
 {
    uint32_t Hash = 5381;
    for (auto I = Str.begin(); I != Str.end(); ++I)
-      Hash = 33 * Hash + tolower_ascii(*I);
+      Hash = 33 * Hash + tolower_ascii_unsafe(*I);
    return Hash % HeaderP->GetHashTableSize();
 }
 map_id_t pkgCache::sHash(const string &Str) const
 {
    uint32_t Hash = 5381;
    for (string::const_iterator I = Str.begin(); I != Str.end(); ++I)
-      Hash = 33 * Hash + tolower_ascii((signed char)*I);
+      Hash = 33 * Hash + tolower_ascii_unsafe((signed char)*I);
    return Hash % HeaderP->GetHashTableSize();
 }
 
@@ -228,7 +228,7 @@ map_id_t pkgCache::sHash(const char *Str) const
 {
    uint32_t Hash = 5381;
    for (const char *I = Str; *I != 0; ++I)
-      Hash = 33 * Hash + tolower_ascii((signed char)*I);
+      Hash = 33 * Hash + tolower_ascii_unsafe((signed char)*I);
    return Hash % HeaderP->GetHashTableSize();
 }