url: find scheme with a "perfect hash"

Instead of a loop to scan over the potentially 30+ scheme names, this
uses a "perfect hash" table. This works fine because the set of schemes
is known and cannot change in a build. The hash algorithm and table size
is made to only make a single scheme index per table entry.

The perfect hash is generated by a separate tool (scripts/schemetable.c)

Closes #12347

diff --git a/lib/url.c b/lib/url.c
index 187dcda9afa27c913a93ee67cc6f37ac7b0b575d..fc35a84b997ce2f14738638f4b896fd1c21ab4ce 100644 (file)
--- a/lib/url.c
+++ b/lib/url.c
@@ -168,130 +168,6 @@ static curl_prot_t get_protocol_family(const struct Curl_handler *h)
   return h->family;
 }
 
-
-/*
- * Protocol table. Schemes (roughly) in 2019 popularity order:
- *
- * HTTPS, HTTP, FTP, FTPS, SFTP, FILE, SCP, SMTP, LDAP, IMAPS, TELNET, IMAP,
- * LDAPS, SMTPS, TFTP, SMB, POP3, GOPHER POP3S, RTSP, RTMP, SMBS, DICT
- */
-static const struct Curl_handler * const protocols[] = {
-
-#if defined(USE_SSL) && !defined(CURL_DISABLE_HTTP)
-  &Curl_handler_https,
-#endif
-
-#ifndef CURL_DISABLE_HTTP
-  &Curl_handler_http,
-#endif
-
-#ifdef USE_WEBSOCKETS
-#if defined(USE_SSL) && !defined(CURL_DISABLE_HTTP)
-  &Curl_handler_wss,
-#endif
-
-#ifndef CURL_DISABLE_HTTP
-  &Curl_handler_ws,
-#endif
-#endif
-
-#ifndef CURL_DISABLE_FTP
-  &Curl_handler_ftp,
-#endif
-
-#if defined(USE_SSL) && !defined(CURL_DISABLE_FTP)
-  &Curl_handler_ftps,
-#endif
-
-#if defined(USE_SSH)
-  &Curl_handler_sftp,
-#endif
-
-#ifndef CURL_DISABLE_FILE
-  &Curl_handler_file,
-#endif
-
-#if defined(USE_SSH) && !defined(USE_WOLFSSH)
-  &Curl_handler_scp,
-#endif
-
-#ifndef CURL_DISABLE_SMTP
-  &Curl_handler_smtp,
-#ifdef USE_SSL
-  &Curl_handler_smtps,
-#endif
-#endif
-
-#ifndef CURL_DISABLE_LDAP
-  &Curl_handler_ldap,
-#if !defined(CURL_DISABLE_LDAPS) && \
-    ((defined(USE_OPENLDAP) && defined(USE_SSL)) || \
-     (!defined(USE_OPENLDAP) && defined(HAVE_LDAP_SSL)))
-  &Curl_handler_ldaps,
-#endif
-#endif
-
-#ifndef CURL_DISABLE_IMAP
-  &Curl_handler_imap,
-#ifdef USE_SSL
-  &Curl_handler_imaps,
-#endif
-#endif
-
-#ifndef CURL_DISABLE_TELNET
-  &Curl_handler_telnet,
-#endif
-
-#ifndef CURL_DISABLE_TFTP
-  &Curl_handler_tftp,
-#endif
-
-#ifndef CURL_DISABLE_POP3
-  &Curl_handler_pop3,
-#ifdef USE_SSL
-  &Curl_handler_pop3s,
-#endif
-#endif
-
-#if !defined(CURL_DISABLE_SMB) && defined(USE_CURL_NTLM_CORE) && \
-   (SIZEOF_CURL_OFF_T > 4)
-  &Curl_handler_smb,
-#ifdef USE_SSL
-  &Curl_handler_smbs,
-#endif
-#endif
-
-#ifndef CURL_DISABLE_RTSP
-  &Curl_handler_rtsp,
-#endif
-
-#ifndef CURL_DISABLE_MQTT
-  &Curl_handler_mqtt,
-#endif
-
-#ifndef CURL_DISABLE_GOPHER
-  &Curl_handler_gopher,
-#ifdef USE_SSL
-  &Curl_handler_gophers,
-#endif
-#endif
-
-#ifdef USE_LIBRTMP
-  &Curl_handler_rtmp,
-  &Curl_handler_rtmpt,
-  &Curl_handler_rtmpe,
-  &Curl_handler_rtmpte,
-  &Curl_handler_rtmps,
-  &Curl_handler_rtmpts,
-#endif
-
-#ifndef CURL_DISABLE_DICT
-  &Curl_handler_dict,
-#endif
-
-  NULL
-};
-
 void Curl_freeset(struct Curl_easy *data)
 {
   /* Free all dynamic strings stored in the data->set substructure. */
@@ -1593,17 +1469,216 @@ const struct Curl_handler *Curl_get_scheme_handler(const char *scheme)
 const struct Curl_handler *Curl_getn_scheme_handler(const char *scheme,
                                                     size_t len)
 {
-  size_t i;
-  /* Scan protocol handler table and match against 'scheme'. The handler may
-     be changed later when the protocol specific setup function is called. */
-  for(i = 0; i < ARRAYSIZE(protocols) - 1; ++i)
-    if(strncasecompare(protocols[i]->scheme, scheme, len) &&
-       !protocols[i]->scheme[len])
-      /* Protocol found in table. */
-      return protocols[i];
-  return NULL; /* not found */
-}
+  /* table generated by schemetable.c:
+     1. gcc schemetable.c && ./a.out
+     2. check how small the table gets
+     3. tweak the hash algorithm, then rerun from 1
+     4. when the table is good enough
+     5. copy the table into this source code
+     6. make sure this function uses the same hash function that worked for
+     schemetable.c
+     7. if needed, adjust the #ifdefs in schemetable.c and rerun
+     */
+  static const struct Curl_handler * const protocols[67] = {
+#ifndef CURL_DISABLE_FILE
+    &Curl_handler_file,
+#else
+    NULL,
+#endif
+    NULL, NULL,
+#if defined(USE_SSL) && !defined(CURL_DISABLE_GOPHER)
+    &Curl_handler_gophers,
+#else
+    NULL,
+#endif
+    NULL,
+#ifdef USE_LIBRTMP
+    &Curl_handler_rtmpe,
+#else
+    NULL,
+#endif
+#ifndef CURL_DISABLE_SMTP
+    &Curl_handler_smtp,
+#else
+    NULL,
+#endif
+#if defined(USE_SSH)
+    &Curl_handler_sftp,
+#else
+    NULL,
+#endif
+#if !defined(CURL_DISABLE_SMB) && defined(USE_CURL_NTLM_CORE) && \
+  (SIZEOF_CURL_OFF_T > 4)
+    &Curl_handler_smb,
+#else
+    NULL,
+#endif
+#if defined(USE_SSL) && !defined(CURL_DISABLE_SMTP)
+    &Curl_handler_smtps,
+#else
+    NULL,
+#endif
+#ifndef CURL_DISABLE_TELNET
+    &Curl_handler_telnet,
+#else
+    NULL,
+#endif
+#ifndef CURL_DISABLE_GOPHER
+    &Curl_handler_gopher,
+#else
+    NULL,
+#endif
+#ifndef CURL_DISABLE_TFTP
+    &Curl_handler_tftp,
+#else
+    NULL,
+#endif
+    NULL, NULL, NULL,
+#if defined(USE_SSL) && !defined(CURL_DISABLE_FTP)
+    &Curl_handler_ftps,
+#else
+    NULL,
+#endif
+#ifndef CURL_DISABLE_HTTP
+    &Curl_handler_http,
+#else
+    NULL,
+#endif
+#ifndef CURL_DISABLE_IMAP
+    &Curl_handler_imap,
+#else
+    NULL,
+#endif
+#ifdef USE_LIBRTMP
+    &Curl_handler_rtmps,
+#else
+    NULL,
+#endif
+#ifdef USE_LIBRTMP
+    &Curl_handler_rtmpt,
+#else
+    NULL,
+#endif
+    NULL, NULL, NULL,
+#if !defined(CURL_DISABLE_LDAP) && \
+  !defined(CURL_DISABLE_LDAPS) && \
+  ((defined(USE_OPENLDAP) && defined(USE_SSL)) || \
+   (!defined(USE_OPENLDAP) && defined(HAVE_LDAP_SSL)))
+    &Curl_handler_ldaps,
+#else
+    NULL,
+#endif
+#if defined(USE_WEBSOCKETS) && \
+  defined(USE_SSL) && !defined(CURL_DISABLE_HTTP)
+    &Curl_handler_wss,
+#else
+    NULL,
+#endif
+#if defined(USE_SSL) && !defined(CURL_DISABLE_HTTP)
+    &Curl_handler_https,
+#else
+    NULL,
+#endif
+    NULL, NULL, NULL, NULL, NULL, NULL, NULL, NULL,
+#ifndef CURL_DISABLE_RTSP
+    &Curl_handler_rtsp,
+#else
+    NULL,
+#endif
+#if defined(USE_SSL) && !defined(CURL_DISABLE_SMB) && \
+  defined(USE_CURL_NTLM_CORE) && (SIZEOF_CURL_OFF_T > 4)
+    &Curl_handler_smbs,
+#else
+    NULL,
+#endif
+#if defined(USE_SSH) && !defined(USE_WOLFSSH)
+    &Curl_handler_scp,
+#else
+    NULL,
+#endif
+    NULL, NULL, NULL,
+#ifndef CURL_DISABLE_POP3
+    &Curl_handler_pop3,
+#else
+    NULL,
+#endif
+    NULL, NULL,
+#ifdef USE_LIBRTMP
+    &Curl_handler_rtmp,
+#else
+    NULL,
+#endif
+    NULL, NULL, NULL,
+#ifdef USE_LIBRTMP
+    &Curl_handler_rtmpte,
+#else
+    NULL,
+#endif
+    NULL, NULL, NULL,
+#ifndef CURL_DISABLE_DICT
+    &Curl_handler_dict,
+#else
+    NULL,
+#endif
+    NULL, NULL, NULL,
+#ifndef CURL_DISABLE_MQTT
+    &Curl_handler_mqtt,
+#else
+    NULL,
+#endif
+#if defined(USE_SSL) && !defined(CURL_DISABLE_POP3)
+    &Curl_handler_pop3s,
+#else
+    NULL,
+#endif
+#if defined(USE_SSL) && !defined(CURL_DISABLE_IMAP)
+    &Curl_handler_imaps,
+#else
+    NULL,
+#endif
+    NULL,
+#if defined(USE_WEBSOCKETS) && !defined(CURL_DISABLE_HTTP)
+    &Curl_handler_ws,
+#else
+    NULL,
+#endif
+    NULL,
+#ifdef USE_LIBRTMP
+    &Curl_handler_rtmpts,
+#else
+    NULL,
+#endif
+#ifndef CURL_DISABLE_LDAP
+    &Curl_handler_ldap,
+#else
+    NULL,
+#endif
+    NULL, NULL,
+#ifndef CURL_DISABLE_FTP
+    &Curl_handler_ftp,
+#else
+    NULL,
+#endif
+  };
 
+  if(len && (len <= 7)) {
+    const char *s = scheme;
+    size_t l = len;
+    const struct Curl_handler *h;
+    unsigned int c = 978;
+    while(l) {
+      c <<= 5;
+      c += Curl_raw_tolower(*s);
+      s++;
+      l--;
+    }
+
+    h = protocols[c % 67];
+    if(h && strncasecompare(scheme, h->scheme, len) && !h->scheme[len])
+      return h;
+  }
+  return NULL;
+}
 
 static CURLcode findprotocol(struct Curl_easy *data,
                              struct connectdata *conn,