Use a fixed-size reader stack

This improves performance, and makes the reader more suitable for embedded or
network-facing applications, at the cost of requiring the user to specify a
maximum stack size.

1 files changed, 16 insertions, 1 deletions

diff --git a/src/serdi.c b/src/serdi.c
index 6049c31b..d2828ea2 100644
--- a/src/serdi.c
+++ b/src/serdi.c
@@ -28,8 +28,10 @@
 #  include <io.h>
 #endif
 
+#include <limits.h>
 #include <stdbool.h>
 #include <stdio.h>
+#include <stdlib.h>
 #include <string.h>
 
 #define SERDI_ERROR(msg) fprintf(stderr, "serdi: " msg)
@@ -102,6 +104,7 @@ print_usage(const char* name, bool error)
   fprintf(os, "  -f           Keep full URIs in input (don't qualify).\n");
   fprintf(os, "  -h           Display this help and exit.\n");
   fprintf(os, "  -i SYNTAX    Input syntax: turtle/ntriples/trig/nquads.\n");
+  fprintf(os, "  -k BYTES     Parser stack size.\n");
   fprintf(os, "  -l           Lax (non-strict) parsing.\n");
   fprintf(os, "  -o SYNTAX    Output syntax: turtle/ntriples/nquads.\n");
   fprintf(os, "  -p PREFIX    Add PREFIX to blank node IDs.\n");
@@ -174,6 +177,7 @@ main(int argc, char** argv)
   bool        full_uris     = false;
   bool        lax           = false;
   bool        quiet         = false;
+  size_t      stack_size    = 4194304;
   const char* add_prefix    = NULL;
   const char* chop_prefix   = NULL;
   const char* root_uri      = NULL;
@@ -212,6 +216,17 @@ main(int argc, char** argv)
       if (!(input_syntax = get_syntax(argv[a]))) {
         return print_usage(argv[0], true);
       }
+    } else if (argv[a][1] == 'k') {
+      if (++a == argc) {
+        return missing_arg(argv[0], 'k');
+      }
+      char*      endptr = NULL;
+      const long size   = strtol(argv[a], &endptr, 10);
+      if (size <= 0 || size == LONG_MAX || *endptr != '\0') {
+        SERDI_ERRORF("invalid stack size `%s'\n", argv[a]);
+        return 1;
+      }
+      stack_size = (size_t)size;
     } else if (argv[a][1] == 'o') {
       if (++a == argc) {
         return missing_arg(argv[0], 'o');
@@ -284,7 +299,7 @@ main(int argc, char** argv)
     output_syntax, writer_flags, env, (SerdWriteFunc)fwrite, out_fd);
 
   SerdReader* const reader =
-    serd_reader_new(input_syntax, serd_writer_sink(writer));
+    serd_reader_new(input_syntax, serd_writer_sink(writer), stack_size);
 
   serd_reader_set_strict(reader, !lax);
   if (quiet) {