obj: move parsing utilities out of io_common, since they are fairly obj specific

As pointed out in https://developer.blender.org/rB213cd39b6db387bd88f12589fd50ff0e6563cf56#341113,
the utilities are quite OBJ specific due to treating backslash as a line
continuation character. It's unlikely that other formats need that.

No functionality changes, just pure code move (and renamed tests so that
their names reflect obj).

Reviewed By: Campbell Barton
Differential Revision: https://developer.blender.org/D14871

6 files changed, 295 insertions, 6 deletions

diff --git a/source/blender/io/wavefront_obj/CMakeLists.txt b/source/blender/io/wavefront_obj/CMakeLists.txt
index e0fe7ed4992..a7c4253f4d3 100644
--- a/source/blender/io/wavefront_obj/CMakeLists.txt
+++ b/source/blender/io/wavefront_obj/CMakeLists.txt
@@ -4,7 +4,6 @@ set(INC
   .
   ./exporter
   ./importer
-  ../common
   ../../blenkernel
   ../../blenlib
   ../../bmesh
@@ -15,6 +14,7 @@ set(INC
   ../../makesrna
   ../../nodes
   ../../windowmanager
+  ../../../../extern/fast_float
   ../../../../extern/fmtlib/include
   ../../../../intern/guardedalloc
 )
@@ -35,6 +35,7 @@ set(SRC
   importer/obj_import_mesh.cc
   importer/obj_import_mtl.cc
   importer/obj_import_nurbs.cc
+  importer/obj_import_string_utils.cc
   importer/obj_importer.cc
 
   IO_wavefront_obj.h
@@ -50,12 +51,12 @@ set(SRC
   importer/obj_import_mtl.hh
   importer/obj_import_nurbs.hh
   importer/obj_import_objects.hh
+  importer/obj_import_string_utils.hh
   importer/obj_importer.hh
 )
 
 set(LIB
   bf_blenkernel
-  bf_io_common
 )
 
 if(WITH_TBB)
@@ -69,6 +70,7 @@ blender_add_lib(bf_wavefront_obj "${SRC}" "${INC}" "${INC_SYS}" "${LIB}")
 if(WITH_GTESTS)
   set(TEST_SRC
     tests/obj_exporter_tests.cc
+    tests/obj_import_string_utils_tests.cc
     tests/obj_importer_tests.cc
     tests/obj_mtl_parser_tests.cc
 
diff --git a/source/blender/io/wavefront_obj/importer/obj_import_file_reader.cc b/source/blender/io/wavefront_obj/importer/obj_import_file_reader.cc
index be322f49840..fa89b49b605 100644
--- a/source/blender/io/wavefront_obj/importer/obj_import_file_reader.cc
+++ b/source/blender/io/wavefront_obj/importer/obj_import_file_reader.cc
@@ -8,9 +8,8 @@
 #include "BLI_string_ref.hh"
 #include "BLI_vector.hh"
 
-#include "IO_string_utils.hh"
-
 #include "obj_import_file_reader.hh"
+#include "obj_import_string_utils.hh"
 
 namespace blender::io::obj {
 
diff --git a/source/blender/io/wavefront_obj/importer/obj_import_mtl.cc b/source/blender/io/wavefront_obj/importer/obj_import_mtl.cc
index c2ecd8a37de..f39def0a4af 100644
--- a/source/blender/io/wavefront_obj/importer/obj_import_mtl.cc
+++ b/source/blender/io/wavefront_obj/importer/obj_import_mtl.cc
@@ -13,13 +13,12 @@
 #include "DNA_material_types.h"
 #include "DNA_node_types.h"
 
-#include "IO_string_utils.hh"
-
 #include "NOD_shader.h"
 
 /* TODO: move eMTLSyntaxElement out of following file into a more neutral place */
 #include "obj_export_io.hh"
 #include "obj_import_mtl.hh"
+#include "obj_import_string_utils.hh"
 
 namespace blender::io::obj {
 
diff --git a/source/blender/io/wavefront_obj/importer/obj_import_string_utils.cc b/source/blender/io/wavefront_obj/importer/obj_import_string_utils.cc
new file mode 100644
index 00000000000..c60306c8375
--- /dev/null
+++ b/source/blender/io/wavefront_obj/importer/obj_import_string_utils.cc
@@ -0,0 +1,99 @@
+/* SPDX-License-Identifier: GPL-2.0-or-later */
+
+#include "obj_import_string_utils.hh"
+
+/* Note: we could use C++17 <charconv> from_chars to parse
+ * floats, but even if some compilers claim full support,
+ * their standard libraries are not quite there yet.
+ * LLVM/libc++ only has a float parser since LLVM 14,
+ * and gcc/libstdc++ since 11.1. So until at least these are
+ * the minimum spec, use an external library. */
+#include "fast_float.h"
+#include <charconv>
+
+namespace blender::io::obj {
+
+StringRef read_next_line(StringRef &buffer)
+{
+  const char *start = buffer.begin();
+  const char *end = buffer.end();
+  size_t len = 0;
+  char prev = 0;
+  const char *ptr = start;
+  while (ptr < end) {
+    char c = *ptr++;
+    if (c == '\n' && prev != '\\') {
+      break;
+    }
+    prev = c;
+    ++len;
+  }
+
+  buffer = StringRef(ptr, end);
+  return StringRef(start, len);
+}
+
+static bool is_whitespace(char c)
+{
+  return c <= ' ' || c == '\\';
+}
+
+StringRef drop_whitespace(StringRef str)
+{
+  while (!str.is_empty() && is_whitespace(str[0])) {
+    str = str.drop_prefix(1);
+  }
+  return str;
+}
+
+StringRef drop_non_whitespace(StringRef str)
+{
+  while (!str.is_empty() && !is_whitespace(str[0])) {
+    str = str.drop_prefix(1);
+  }
+  return str;
+}
+
+static StringRef drop_plus(StringRef str)
+{
+  if (!str.is_empty() && str[0] == '+') {
+    str = str.drop_prefix(1);
+  }
+  return str;
+}
+
+StringRef parse_float(StringRef str, float fallback, float &dst, bool skip_space)
+{
+  if (skip_space) {
+    str = drop_whitespace(str);
+  }
+  str = drop_plus(str);
+  fast_float::from_chars_result res = fast_float::from_chars(str.begin(), str.end(), dst);
+  if (res.ec == std::errc::invalid_argument || res.ec == std::errc::result_out_of_range) {
+    dst = fallback;
+  }
+  return StringRef(res.ptr, str.end());
+}
+
+StringRef parse_floats(StringRef str, float fallback, float *dst, int count)
+{
+  for (int i = 0; i < count; ++i) {
+    str = parse_float(str, fallback, dst[i]);
+  }
+  return str;
+}
+
+StringRef parse_int(StringRef str, int fallback, int &dst, bool skip_space)
+{
+  if (skip_space) {
+    str = drop_whitespace(str);
+  }
+  str = drop_plus(str);
+  std::from_chars_result res = std::from_chars(str.begin(), str.end(), dst);
+  if (res.ec == std::errc::invalid_argument || res.ec == std::errc::result_out_of_range) {
+    dst = fallback;
+  }
+  return StringRef(res.ptr, str.end());
+}
+
+}  // namespace blender::io::obj
diff --git a/source/blender/io/wavefront_obj/importer/obj_import_string_utils.hh b/source/blender/io/wavefront_obj/importer/obj_import_string_utils.hh
new file mode 100644
index 00000000000..532224569ac
--- /dev/null
+++ b/source/blender/io/wavefront_obj/importer/obj_import_string_utils.hh
@@ -0,0 +1,72 @@
+/* SPDX-License-Identifier: GPL-2.0-or-later */
+
+#pragma once
+
+#include "BLI_string_ref.hh"
+
+/*
+ * Various text parsing utilities used by OBJ importer.
+ * The utilities are not directly usable by other formats, since
+ * they treat backslash (\) as a whitespace character (OBJ format
+ * allows backslashes to function as a line-continuation character).
+ */
+
+namespace blender::io::obj {
+
+/**
+ * Fetches next line from an input string buffer.
+ *
+ * The returned line will not have '\n' characters at the end;
+ * the `buffer` is modified to contain remaining text without
+ * the input line.
+ *
+ * Note that backslash (\) character is treated as a line
+ * continuation.
+ */
+StringRef read_next_line(StringRef &buffer);
+
+/**
+ * Drop leading white-space from a StringRef.
+ * Note that backslash character is considered white-space.
+ */
+StringRef drop_whitespace(StringRef str);
+
+/**
+ * Drop leading non-white-space from a StringRef.
+ * Note that backslash character is considered white-space.
+ */
+StringRef drop_non_whitespace(StringRef str);
+
+/**
+ * Parse an integer from an input string.
+ * The parsed result is stored in `dst`. The function skips
+ * leading white-space unless `skip_space=false`. If the
+ * number can't be parsed (invalid syntax, out of range),
+ * `fallback` value is stored instead.
+ *
+ * Returns the remainder of the input string after parsing.
+ */
+StringRef parse_int(StringRef str, int fallback, int &dst, bool skip_space = true);
+
+/**
+ * Parse a float from an input string.
+ * The parsed result is stored in `dst`. The function skips
+ * leading white-space unless `skip_space=false`. If the
+ * number can't be parsed (invalid syntax, out of range),
+ * `fallback` value is stored instead.
+ *
+ * Returns the remainder of the input string after parsing.
+ */
+StringRef parse_float(StringRef str, float fallback, float &dst, bool skip_space = true);
+
+/**
+ * Parse a number of white-space separated floats from an input string.
+ * The parsed `count` numbers are stored in `dst`. If a
+ * number can't be parsed (invalid syntax, out of range),
+ * `fallback` value is stored instead.
+ *
+ * Returns the remainder of the input string after parsing.
+ */
+StringRef parse_floats(StringRef str, float fallback, float *dst, int count);
+
+}  // namespace blender::io::obj
diff --git a/source/blender/io/wavefront_obj/tests/obj_import_string_utils_tests.cc b/source/blender/io/wavefront_obj/tests/obj_import_string_utils_tests.cc
new file mode 100644
index 00000000000..addb1fa473e
--- /dev/null
+++ b/source/blender/io/wavefront_obj/tests/obj_import_string_utils_tests.cc
@@ -0,0 +1,118 @@
+/* SPDX-License-Identifier: Apache-2.0 */
+
+#include "obj_import_string_utils.hh"
+
+#include "testing/testing.h"
+
+namespace blender::io::obj {
+
+#define EXPECT_STRREF_EQ(str1, str2) EXPECT_STREQ(str1, std::string(str2).c_str())
+
+TEST(obj_import_string_utils, read_next_line)
+{
+  std::string str = "abc\n  \n\nline with \\\ncontinuation\nCRLF ending:\r\na";
+  StringRef s = str;
+  EXPECT_STRREF_EQ("abc", read_next_line(s));
+  EXPECT_STRREF_EQ("  ", read_next_line(s));
+  EXPECT_STRREF_EQ("", read_next_line(s));
+  EXPECT_STRREF_EQ("line with \\\ncontinuation", read_next_line(s));
+  EXPECT_STRREF_EQ("CRLF ending:\r", read_next_line(s));
+  EXPECT_STRREF_EQ("a", read_next_line(s));
+  EXPECT_TRUE(s.is_empty());
+}
+
+TEST(obj_import_string_utils, drop_whitespace)
+{
+  /* Empty */
+  EXPECT_STRREF_EQ("", drop_whitespace(""));
+  /* Only whitespace */
+  EXPECT_STRREF_EQ("", drop_whitespace(" "));
+  EXPECT_STRREF_EQ("", drop_whitespace("   "));
+  EXPECT_STRREF_EQ("", drop_whitespace(" \t\n\r "));
+  /* Drops leading whitespace */
+  EXPECT_STRREF_EQ("a", drop_whitespace(" a"));
+  EXPECT_STRREF_EQ("a b", drop_whitespace("   a b"));
+  EXPECT_STRREF_EQ("a b   ", drop_whitespace(" a b   "));
+  /* No leading whitespace */
+  EXPECT_STRREF_EQ("c", drop_whitespace("c"));
+  /* Case with backslash, should be treated as whitespace */
+  EXPECT_STRREF_EQ("d", drop_whitespace(" \\ d"));
+}
+
+TEST(obj_import_string_utils, parse_int_valid)
+{
+  std::string str = "1 -10 \t  1234 1234567890 +7 123a";
+  StringRef s = str;
+  int val;
+  s = parse_int(s, 0, val);
+  EXPECT_EQ(1, val);
+  s = parse_int(s, 0, val);
+  EXPECT_EQ(-10, val);
+  s = parse_int(s, 0, val);
+  EXPECT_EQ(1234, val);
+  s = parse_int(s, 0, val);
+  EXPECT_EQ(1234567890, val);
+  s = parse_int(s, 0, val);
+  EXPECT_EQ(7, val);
+  s = parse_int(s, 0, val);
+  EXPECT_EQ(123, val);
+  EXPECT_STRREF_EQ("a", s);
+}
+
+TEST(obj_import_string_utils, parse_int_invalid)
+{
+  int val;
+  /* Invalid syntax */
+  EXPECT_STRREF_EQ("--123", parse_int("--123", -1, val));
+  EXPECT_EQ(val, -1);
+  EXPECT_STRREF_EQ("foobar", parse_int("foobar", -2, val));
+  EXPECT_EQ(val, -2);
+  /* Out of integer range */
+  EXPECT_STRREF_EQ(" a", parse_int("1234567890123 a", -3, val));
+  EXPECT_EQ(val, -3);
+  /* Has leading white-space when we don't expect it */
+  EXPECT_STRREF_EQ(" 1", parse_int(" 1", -4, val, false));
+  EXPECT_EQ(val, -4);
+}
+
+TEST(obj_import_string_utils, parse_float_valid)
+{
+  std::string str = "1 -10 123.5 -17.125 0.1 1e6 50.0e-1";
+  StringRef s = str;
+  float val;
+  s = parse_float(s, 0, val);
+  EXPECT_EQ(1.0f, val);
+  s = parse_float(s, 0, val);
+  EXPECT_EQ(-10.0f, val);
+  s = parse_float(s, 0, val);
+  EXPECT_EQ(123.5f, val);
+  s = parse_float(s, 0, val);
+  EXPECT_EQ(-17.125f, val);
+  s = parse_float(s, 0, val);
+  EXPECT_EQ(0.1f, val);
+  s = parse_float(s, 0, val);
+  EXPECT_EQ(1.0e6f, val);
+  s = parse_float(s, 0, val);
+  EXPECT_EQ(5.0f, val);
+  EXPECT_TRUE(s.is_empty());
+}
+
+TEST(obj_import_string_utils, parse_float_invalid)
+{
+  float val;
+  /* Invalid syntax */
+  EXPECT_STRREF_EQ("_0", parse_float("_0", -1.0f, val));
+  EXPECT_EQ(val, -1.0f);
+  EXPECT_STRREF_EQ("..5", parse_float("..5", -2.0f, val));
+  EXPECT_EQ(val, -2.0f);
+  /* Out of float range. Current float parser (fast_float)
+   * clamps out of range numbers to +/- infinity, so this
+   * one gets a +inf instead of fallback -3.0. */
+  EXPECT_STRREF_EQ(" a", parse_float("9.0e500 a", -3.0f, val));
+  EXPECT_EQ(val, std::numeric_limits<float>::infinity());
+  /* Has leading white-space when we don't expect it */
+  EXPECT_STRREF_EQ(" 1", parse_float(" 1", -4.0f, val, false));
+  EXPECT_EQ(val, -4.0f);
+}
+
+}  // namespace blender::io::obj