Add functions psl_latest() and psl_dist_filename()

Also add a new ./configure function to set a distribution wide
PSL file used by psl_latest(): --with-psl-distfile
If possible that filename should point to a DAFSA PSL file that
becomes updated regularly.

configure.ac

@@ -243,6 +243,11 @@ else
   TESTS_INFO="Valgrind testing not enabled"
 fi
 
+# Check for distribution-wide PSL file
+AC_ARG_WITH(psl-distfile,
+  AC_HELP_STRING([--with-psl-distfile=[PATH]], [path to distribution-wide PSL file]),
+  PSL_DISTFILE=$withval AC_SUBST(PSL_DISTFILE))
+
 # Check for custom PSL file
 AC_ARG_WITH(psl-file,
   AC_HELP_STRING([--with-psl-file=[PATH]], [path to PSL file]),
@@ -283,6 +288,7 @@ AC_MSG_NOTICE([Summary of build options:
   Libs:              ${LIBS}
   Runtime:           ${enable_runtime}
   Builtin:           ${enable_builtin}
+  PSL Dist File:     ${PSL_DISTFILE}
   PSL File:          ${PSL_FILE}
   PSL Test File:     ${PSL_TESTFILE}
   Tests:             ${TESTS_INFO}

docs/libpsl/libpsl-sections.txt

@@ -13,6 +13,7 @@ psl_error_t
 psl_ctx_t
 psl_load_file
 psl_load_fp
+psl_latest
 psl_builtin
 psl_free
 psl_is_public_suffix
@@ -27,6 +28,7 @@ psl_builtin_sha1sum
 psl_builtin_filename
 psl_builtin_outdated
 psl_is_cookie_domain_acceptable
+psl_dist_filename
 psl_get_version
 psl_check_version_number
 psl_str_to_utf8lower

include/libpsl.h.in

@@ -91,6 +91,10 @@ psl_ctx_t *
 const psl_ctx_t *
 	psl_builtin(void);
 
+/* retrieves most recent PSL data */
+psl_ctx_t *
+	psl_latest(const char *fname);
+
 /* checks whether domain is a public suffix or not */
 int
 	psl_is_public_suffix(const psl_ctx_t *psl, const char *domain);
@@ -139,6 +143,10 @@ const char *
 const char *
 	psl_builtin_filename(void);
 
+/* returns name of distribution PSL data file */
+const char *
+	psl_dist_filename(void);
+
 /* returns library version string */
 const char *
 	psl_get_version(void);

src/Makefile.am

@@ -7,7 +7,7 @@ CLEANFILES = suffixes_dafsa.c
 lib_LTLIBRARIES = libpsl.la
 
 libpsl_la_SOURCES = psl.c lookup_string_in_fixed_set.c
-libpsl_la_CPPFLAGS = -I$(top_srcdir)/include
+libpsl_la_CPPFLAGS = -I$(top_srcdir)/include -DPSL_DISTFILE=\"$(PSL_DISTFILE)\"
 # include ABI version information
 libpsl_la_LDFLAGS = -version-info $(LIBPSL_SO_VERSION)
 if WITH_LIBICU

src/psl.c

@@ -176,6 +176,12 @@ struct _psl_ctx_st {
 static const psl_ctx_t
 	_builtin_psl;
 
+#ifdef PSL_DISTFILE
+static const char _psl_dist_filename[] = PSL_DISTFILE;
+#else
+static const char _psl_dist_filename[] = "";
+#endif
+
 static _psl_vector_t *_vector_alloc(int max, int (*cmp)(const _psl_entry_t **, const _psl_entry_t **))
 {
 	_psl_vector_t *v;
@@ -1416,6 +1422,23 @@ int psl_builtin_outdated(void)
 	return 0;
 }
 
+/**
+ * psl_dist_filename:
+ *
+ * This function returns the file name of the distribution/system PSL data file.
+ * This file will be considered by psl_latest().
+ *
+ * Return the filename that is set by ./configure --with-psl-distfile, or an empty string.
+ *
+ * Returns: String containing a PSL file name or an empty string.
+ *
+ * Since: 0.16
+ */
+const char *psl_dist_filename(void)
+{
+	return _psl_dist_filename;
+}
+
 /**
  * psl_get_version:
  *
@@ -1741,3 +1764,70 @@ out:
 
 	return ret;
 }
+
+/* if file is newer than the builtin data, insert it reverse sorted by mtime */
+static int _insert_file(const char *fname, const char **psl_fname, time_t *psl_mtime, int n)
+{
+	struct stat st;
+	int it;
+
+	if (fname && *fname && stat(fname, &st) == 0 && st.st_mtime > _psl_file_time) {
+		/* add file name and mtime to end of array */
+		psl_fname[n] = fname;
+		psl_mtime[n++] = st.st_mtime;
+
+		/* move the new entry to it's correct position */
+		for (it = n - 2; it >= 0 && st.st_mtime > psl_mtime[it]; it--) {
+			psl_fname[it + 1] = psl_fname[it];
+			psl_mtime[it + 1] = psl_mtime[it];
+			psl_fname[it] = fname;
+			psl_mtime[it] = st.st_mtime;
+		}
+	}
+
+	return n;
+}
+
+/**
+ * psl_latest:
+ * @fname: Name of PSL file or %NULL
+ *
+ * This function loads the the latest available PSL data from either
+ * - @fname (application specific filename, may be %NULL)
+ * - location specified during built-time (filename from ./configure --with-psl-distfile)
+ * - built-in PSL data (generated from ./configure --with-psl-file)
+ * - location of built-in data (filename from ./configure --with-psl-file)
+ * 
+ * If none of the above is available, the function returns %NULL.
+ *
+ * To free the allocated resources, call psl_free().
+ *
+ * Returns: Pointer to a PSL context or %NULL on failure.
+ *
+ * Since: 0.16
+ */
+psl_ctx_t *psl_latest(const char *fname)
+{
+	psl_ctx_t *psl;
+	const char *psl_fname[3];
+	time_t psl_mtime[3];
+	int it, ntimes;
+
+	psl_fname[0] = NULL; /* silence gcc 6.2 false warning */
+
+	/* create array of PSL files reverse sorted by mtime (latest first) */
+	ntimes = _insert_file(fname, psl_fname, psl_mtime, 0);
+	ntimes = _insert_file(_psl_filename, psl_fname, psl_mtime, ntimes);
+	ntimes = _insert_file(_psl_dist_filename, psl_fname, psl_mtime, ntimes);
+
+	/* load PSL data from the latest file, falling back to the second recent, ... */
+	for (psl = NULL, it = 0; it < ntimes; it++) {
+		if (psl_mtime[it] > _psl_file_time)
+			if ((psl = psl_load_file(psl_fname[it])))
+				break;
+	}
+
+	/* if file loading failed or there is no file newer than the builtin data,
+	 * then return the builtin data. */
+	return psl ? psl : (psl_ctx_t *) psl_builtin();
+}

tools/psl.c

@@ -72,7 +72,7 @@ int main(int argc, const char *const *argv)
 {
 	int mode = 1;
 	const char *const *arg, *psl_file = NULL, *cookie_domain = NULL;
-	psl_ctx_t *psl = (psl_ctx_t *) psl_builtin();
+	psl_ctx_t *psl = (psl_ctx_t *) psl_latest(NULL);
 
 	/* set current locale according to the environment variables */
 	#include <locale.h>
@@ -197,6 +197,8 @@ int main(int argc, const char *const *argv)
 			printf("%s: %d\n", *arg, psl_is_cookie_domain_acceptable(psl, *arg, cookie_domain));
 	}
 	else if (mode == 99) {
+		printf("dist filename: %s\n", psl_dist_filename());
+
 		if (psl && psl != psl_builtin()) {
 			printf("suffixes: %d\n", psl_suffix_count(psl));
 			printf("exceptions: %d\n", psl_suffix_exception_count(psl));