fix cum_scanned_length tracking for files larger than initial read buff

app/Makefile

@@ -260,7 +260,7 @@ MORE_SOURCE+= ${YAJL_INCLUDE} ${YAJL_HELPER_INCLUDE} -I${JQ_INCLUDE_DIR}
 MORE_LIBS+=${JQ_LIB} ${LDFLAGS_JQ}
 
 help:
-	@echo "To build: ${MAKE} [DEBUG=1] [clean] [clean-all] [BINDIR=${BINDIR}] [JQ_PREFIX=/usr/local] <install|all|test>"
+	@echo "To build: ${MAKE} [DEBUG=1] [clean] [clean-all] [BINDIR=${BINDIR}] [JQ_PREFIX=/usr/local] <install|all|install-util-lib|test>"
 	@echo
 	@echo "If JQ_PREFIX is not defined, libjq will be built in the build dir"
 	@echo

app/builtin/help.c

@@ -35,7 +35,7 @@ static int main_help(int argc, const char *argv[]) {
     "",
     "Options common to all commands except `prop`, `rm` and `jq`:",
 #ifdef ZSV_EXTRAS
-    "  -L,--limit-rows <n>: limit processing to the given number of rows (including any header row(s))",
+    "  -L,--limit-rows <n>      : limit processing to the given number of rows (including any header row(s))",
 #endif
     "  -c,--max-column-count <n>: set the maximum number of columns parsed per row. defaults to 1024",
     "  -r,--max-row-size <n>    : set the minimum supported maximum row size. defaults to 64k",

app/ext_example/Makefile

@@ -52,7 +52,7 @@ endif
 
 DEBUG=0
 ifeq ($(DEBUG),0)
-  CFLAGS+= -O3 -DNDEBUG -std=gnu11 -Wno-gnu-statement-expression -Wshadow -Wall -Wextra -Wno-missing-braces -pedantic -DSTDC_HEADERS -D_GNU_SOURCE -lm -mavx2 -ftree-vectorize -flto
+  CFLAGS+= -O3 -DNDEBUG -std=gnu11 -Wno-gnu-statement-expression -Wshadow -Wall -Wextra -Wno-missing-braces -pedantic -DSTDC_HEADERS -D_GNU_SOURCE -lm -ftree-vectorize -flto
 else
   CFLAGS += -g
 endif

app/ext_example/test/expected/zsvext-test-3.out

@@ -18,7 +18,7 @@ Usage:
   zsv license [<extension_id>]
 
 Options common to all commands except `prop`, `rm` and `jq`:
-  -L,--limit-rows <n>: limit processing to the given number of rows (including any header row(s))
+  -L,--limit-rows <n>      : limit processing to the given number of rows (including any header row(s))
   -c,--max-column-count <n>: set the maximum number of columns parsed per row. defaults to 1024
   -r,--max-row-size <n>    : set the minimum supported maximum row size. defaults to 64k
   -B,--buff-size <n>       : set internal buffer size. defaults to 256k
@@ -76,7 +76,7 @@ Usage:
   zsv license [<extension_id>]
 
 Options common to all commands except `prop`, `rm` and `jq`:
-  -L,--limit-rows <n>: limit processing to the given number of rows (including any header row(s))
+  -L,--limit-rows <n>      : limit processing to the given number of rows (including any header row(s))
   -c,--max-column-count <n>: set the maximum number of columns parsed per row. defaults to 1024
   -r,--max-row-size <n>    : set the minimum supported maximum row size. defaults to 64k
   -B,--buff-size <n>       : set internal buffer size. defaults to 256k

app/select.c

@@ -962,7 +962,8 @@ int ZSV_MAIN_FUNC(ZSV_COMMAND)(int argc, const char *argv[], struct zsv_opts *op
         while(status == zsv_status_ok
               && !zsv_signal_interrupted && !data.cancelled)
           status = zsv_parse_more(data.parser);
-        zsv_finish(data.parser);
+        if(status == zsv_status_no_more_input)
+          status = zsv_finish(data.parser);
         zsv_delete(data.parser);
       }
     }

configure

@@ -405,8 +405,8 @@ test -n "$MAKE" || fail "$0: cannot find a MAKE tool"
 # Find a C compiler to use
 #
 printf "checking for C compiler... "
-for c in cc gcc gcc-11 gcc-10 gcc-9 gcc-8 gcc-7 clang; do trycc1 "$c"; done
-for c in cc gcc gcc-11 gcc-10 gcc-9 gcc-8 gcc-7 clang; do trycc2 "$c"; done
+for c in cc gcc gcc-14 gcc-13 gcc-12 gcc-11 gcc-10 gcc-9 gcc-8 gcc-7 clang; do trycc1 "$c"; done
+for c in cc gcc gcc-14 gcc-13 gcc-12 gcc-11 gcc-10 gcc-9 gcc-8 gcc-7 clang; do trycc2 "$c"; done
 printf "%s\n" "$CC"
 test -n "$CC" || fail "$0: cannot find a C compiler"
 

include/zsv/api.h

@@ -41,6 +41,9 @@
  * - zsv_delete(): dispose the parser
  ******************************************************************************/
 
+ZSV_EXPORT
+int zsv_peek(zsv_parser);
+
 /**
  * Create a zsv parser. Typically, passed options will at least include a
  * a `row_handler()` callback. Many, but not all, options can be subsequently

include/zsv/utils/dirs.h

@@ -73,7 +73,7 @@ struct zsv_foreach_dirent_handle {
 
   unsigned char verbose:1;
   unsigned char is_dir:1;       /* non-zero if this entry is a directory */
-  unsigned char no_recurse:1;        /* set to 1 when handling a dir to prevent recursing into it */
+  unsigned char no_recurse:1;   /* set to 1 when handling a dir to prevent recursing into it */
   unsigned char _:5;
 };
 

scripts/ci-build.sh

@@ -56,6 +56,7 @@ if [ "$RUN_TESTS" = true ]; then
   echo "[INF] Tests completed successfully!"
 
   echo "[INF] Configuring example extension and running example extension tests"
+  echo "[INF] (cd app/ext_example && $MAKE CONFIGFILE=../../config.mk test)"
   (cd app/ext_example && "$MAKE" CONFIGFILE=../../config.mk test)
   echo "[INF] Tests completed successfully!"
 fi

src/zsv.c

@@ -58,10 +58,11 @@ inline static size_t scanner_pre_parse(struct zsv_scanner *scanner) {
     scanner->old_bytes_read = 0;
   }
 
-  scanner->cum_scanned_length += scanner->scanned_length;
+  scanner->cum_scanned_length += scanner->scanned_length - scanner->partial_row_length;
 
   size_t capacity = scanner->buff.size - scanner->partial_row_length;
-  if(VERY_UNLIKELY(capacity == 0)) { // our row size was too small to fit a single row of data
+  if(VERY_UNLIKELY(capacity == 0)) {
+    // our row size was too small to fit a single row of data
     fprintf(stderr, "Warning: row %zu truncated\n", scanner->data_row_count);
     if(scanner->mode == ZSV_MODE_FIXED) {
       if(VERY_UNLIKELY(row_fx(scanner, scanner->buff.buff, 0, scanner->buff.size)))
@@ -310,6 +311,13 @@ ZSV_EXPORT enum zsv_status zsv_set_fixed_offsets(zsv_parser parser, size_t count
   return zsv_status_ok;
 }
 
+ZSV_EXPORT
+int zsv_peek(zsv_parser z) {
+  if(z->scanned_length + 1 < z->buff.size)
+    return z->buff.buff[z->scanned_length+1];
+  return -1;
+}
+
 /**
  * Create a zsv parser
  * @param opts