smarter_csv 1.14.1 → 1.14.3

checksums.yaml

@@ -1,7 +1,7 @@
 ---
 SHA256:
-  metadata.gz: c4619533008fc05b02a009b1409d1368d56245584baf52134511b45ff505f011
-  data.tar.gz: 051fa13106a0dfe41486accfee972a07c7d80670a65dd1066f07dfc600489be3
+  metadata.gz: 79d1b26219dfb35f385c30eea361bc90d4b1b6cfc1030b6abb9339cc99eb39de
+  data.tar.gz: ab0915f4193b657c7e8380ea3bc924a4efb40f9171fb3f120aef462d6f464370
 SHA512:
-  metadata.gz: 5ee274a0a485e87f356940eeaf5e68d6af3d83dd72cd5199eca9729f6c760d0528ac2db8931c2d741521edb69251a4ff3d0c4d458f449a824f2b8110e7ad3a51
-  data.tar.gz: 87ac821280845041514a0418b94389296e6b03faf5c5fded8e2b0a0840015c03c477ad57d6a8268316b110adffb295882c81b7876c6f9f76a444531faed3c431
+  metadata.gz: 722a598f1581423d4295eb3ce6f46bf556fbf0206dbd4229eceff68625f60ad83d7e3f2dda3a5b6d61c18be2ff6d3d0fb495c46427bb2cc20fcf5f72fa58e317
+  data.tar.gz: f61d37aa5a64e37601a7b12ce57a528bc26fec8ee643e4171040f0e9021a2fe4fdeec71698638e98e29da27248e51e77792ed5053bb252716298219f4b095e85

data/CHANGELOG.md

@@ -1,9 +1,21 @@
 
 # SmarterCSV 1.x Change Log
 
+## 1.14.3 (2025-05-04)
+ * Improved C-extension parsing logic:
+   - Added fast path for unquoted fields to avoid unnecessary quote checks.
+   - Aded inline whitespace stripping inside the C parser
+ * Performance
+   -  Significantly reduced per-line overhead in non-quoted, wide CSVs (e.g. fixed-width data exports).
+   - Benchmarks show ~10–40% speedup over v1.14.2 depending on structure and quoting.
+
+## 1.14.2 (2025-04-10)
+ * bugfix: SmarterCSV::Writer fixing corner case with `quote_headers: true` ([issue 301](https://github.com/tilo/smarter_csv/issues/301))
+ * new option: `header_converter` allows to programatically modify the headers
+
 ## 1.14.1 (2025-04-09)
- * bugfix: empty hash results in a blank line ([issue 299](https://github.com/tilo/smarter_csv/issues/299))
- * bugfix: automatically quote problematic headers ([issue #300](https://github.com/tilo/smarter_csv/issues/300))
+ * bugfix: SmarterCSV::Writer empty hash results in a blank line ([issue 299](https://github.com/tilo/smarter_csv/issues/299))
+ * bugfix: SmarterCSV::Writer need to automatically quote problematic headers ([issue #300](https://github.com/tilo/smarter_csv/issues/300))
  * new option: `quote_headers` allows to explicitly quote all headers
 
 ## 1.14.0 (2025-04-07)

data/README.md

@@ -47,6 +47,7 @@ Or install it yourself as:
     
 # Articles
   * [Parsing CSV Files in Ruby with SmarterCSV](https://tilo-sloboda.medium.com/parsing-csv-files-in-ruby-with-smartercsv-6ce66fb6cf38)
+  * [CSV Writing with SmarterCSV](https://tilo-sloboda.medium.com/csv-writing-with-smartercsv-26136d47ad0c)
   * [Processing 1.4 Million CSV Records in Ruby, fast ](https://lcx.wien/blog/processing-14-million-csv-records-in-ruby/)
   * [Faster Parsing CSV with Parallel Processing](http://xjlin0.github.io/tech/2015/05/25/faster-parsing-csv-with-parallel-processing) by [Jack lin](https://github.com/xjlin0/)
   * The original [Stackoverflow Question](https://stackoverflow.com/questions/7788618/update-mongodb-with-array-from-csv-join-table/7788746#7788746) that inspired SmarterCSV

data/docs/options.md

@@ -28,10 +28,15 @@
      |                             |           | ⚠️ This disables automatic header detection!                    |
      | :map_headers                |    {}     | Similar to `headers`, but also maps each desired key to a user-specified value that is uesd as the header. | 
      |                             |           | ⚠️ This disables automatic header detection!                    |
+     | :value_converters           |    nil    | allows to define lambdas to programmatically modify values       |
+     |                             |           | * either for specific `key` names                                |
+     |                             |           | * or using `_all` for all fields                                |
+     | :header_converter           |    nil    | allows to define one lambda to programmatically modify the headers |
      | :discover_headers           |   true    | Automatically detects all keys in the input before writing the header |
      |                             |           | Do not manually set this to `false`. ⚠️                         |
      |                             |           | But you can set this to `true` when using `map_headers` option. |
      | :disable_auto_quoting       |  false    | To manually disable auto-quoting of special characters. ⚠️ Be careful with this! |
+     | :quote_headers              |  false    | To force quoting all headers (only needed in rare cases)        |
 
 
 ## CSV Reading

data/ext/smarter_csv/Makefile

@@ -0,0 +1,270 @@
+
+SHELL = /bin/sh
+
+# V=0 quiet, V=1 verbose.  other values don't work.
+V = 0
+V0 = $(V:0=)
+Q1 = $(V:1=)
+Q = $(Q1:0=@)
+ECHO1 = $(V:1=@ :)
+ECHO = $(ECHO1:0=@ echo)
+NULLCMD = :
+
+#### Start of system configuration section. ####
+
+srcdir = .
+topdir = /Users/tilo/.rvm/rubies/ruby-3.2.2/include/ruby-3.2.0
+hdrdir = $(topdir)
+arch_hdrdir = /Users/tilo/.rvm/rubies/ruby-3.2.2/include/ruby-3.2.0/arm64-darwin23
+PATH_SEPARATOR = :
+VPATH = $(srcdir):$(arch_hdrdir)/ruby:$(hdrdir)/ruby
+prefix = $(DESTDIR)/Users/tilo/.rvm/rubies/ruby-3.2.2
+rubysitearchprefix = $(rubylibprefix)/$(sitearch)
+rubyarchprefix = $(rubylibprefix)/$(arch)
+rubylibprefix = $(libdir)/$(RUBY_BASE_NAME)
+exec_prefix = $(prefix)
+vendorarchhdrdir = $(vendorhdrdir)/$(sitearch)
+sitearchhdrdir = $(sitehdrdir)/$(sitearch)
+rubyarchhdrdir = $(rubyhdrdir)/$(arch)
+vendorhdrdir = $(rubyhdrdir)/vendor_ruby
+sitehdrdir = $(rubyhdrdir)/site_ruby
+rubyhdrdir = $(includedir)/$(RUBY_VERSION_NAME)
+vendorarchdir = $(vendorlibdir)/$(sitearch)
+vendorlibdir = $(vendordir)/$(ruby_version)
+vendordir = $(rubylibprefix)/vendor_ruby
+sitearchdir = $(sitelibdir)/$(sitearch)
+sitelibdir = $(sitedir)/$(ruby_version)
+sitedir = $(rubylibprefix)/site_ruby
+rubyarchdir = $(rubylibdir)/$(arch)
+rubylibdir = $(rubylibprefix)/$(ruby_version)
+sitearchincludedir = $(includedir)/$(sitearch)
+archincludedir = $(includedir)/$(arch)
+sitearchlibdir = $(libdir)/$(sitearch)
+archlibdir = $(libdir)/$(arch)
+ridir = $(datarootdir)/$(RI_BASE_NAME)
+mandir = $(datarootdir)/man
+localedir = $(datarootdir)/locale
+libdir = $(exec_prefix)/lib
+psdir = $(docdir)
+pdfdir = $(docdir)
+dvidir = $(docdir)
+htmldir = $(docdir)
+infodir = $(datarootdir)/info
+docdir = $(datarootdir)/doc/$(PACKAGE)
+oldincludedir = $(DESTDIR)/usr/include
+includedir = $(SDKROOT)$(prefix)/include
+runstatedir = $(localstatedir)/run
+localstatedir = $(prefix)/var
+sharedstatedir = $(prefix)/com
+sysconfdir = $(prefix)/etc
+datadir = $(datarootdir)
+datarootdir = $(prefix)/share
+libexecdir = $(exec_prefix)/libexec
+sbindir = $(exec_prefix)/sbin
+bindir = $(exec_prefix)/bin
+archdir = $(rubyarchdir)
+
+
+CC_WRAPPER = 
+CC = gcc
+CXX = g++
+LIBRUBY = $(LIBRUBY_SO)
+LIBRUBY_A = lib$(RUBY_SO_NAME)-static.a
+LIBRUBYARG_SHARED = -l$(RUBY_SO_NAME)
+LIBRUBYARG_STATIC = -l$(RUBY_SO_NAME)-static -framework CoreFoundation $(MAINLIBS)
+empty =
+OUTFLAG = -o $(empty)
+COUTFLAG = -o $(empty)
+CSRCFLAG = $(empty)
+
+RUBY_EXTCONF_H = 
+cflags   = -fdeclspec $(optflags) $(debugflags) $(warnflags)
+cxxflags = 
+optflags = -O3
+debugflags = -ggdb3
+warnflags = -Wall -Wextra -Wextra-tokens -Wdeprecated-declarations -Wdivision-by-zero -Wdiv-by-zero -Wimplicit-function-declaration -Wimplicit-int -Wmisleading-indentation -Wpointer-arith -Wshorten-64-to-32 -Wwrite-strings -Wold-style-definition -Wmissing-noreturn -Wno-cast-function-type -Wno-constant-logical-operand -Wno-long-long -Wno-missing-field-initializers -Wno-overlength-strings -Wno-parentheses-equality -Wno-self-assign -Wno-tautological-compare -Wno-unused-parameter -Wno-unused-value -Wunused-variable -Wundef
+cppflags = 
+CCDLFLAGS = -fno-common
+CFLAGS   = $(CCDLFLAGS) -O3 -I/opt/homebrew/opt/libyaml/include -I/opt/homebrew/opt/libksba/include -I/opt/homebrew/opt/readline/include -I/opt/homebrew/opt/zlib/include -I/opt/homebrew/opt/openssl@1.1/include $(cflags) -fno-common -pipe $(ARCH_FLAG)
+INCFLAGS = -I. -I$(arch_hdrdir) -I$(hdrdir)/ruby/backward -I$(hdrdir) -I$(srcdir)
+DEFS     = 
+CPPFLAGS =  -D_XOPEN_SOURCE -D_DARWIN_C_SOURCE -D_DARWIN_UNLIMITED_SELECT -D_REENTRANT $(DEFS) $(cppflags)
+CXXFLAGS = $(CCDLFLAGS) -fdeclspec $(ARCH_FLAG)
+ldflags  = -L. -L/opt/homebrew/opt/libyaml/lib -L/opt/homebrew/opt/libksba/lib -L/opt/homebrew/opt/readline/lib -L/opt/homebrew/opt/zlib/lib -L/opt/homebrew/opt/openssl@1.1/lib -fstack-protector-strong
+dldflags = -L/opt/homebrew/opt/libyaml/lib -L/opt/homebrew/opt/libksba/lib -L/opt/homebrew/opt/readline/lib -L/opt/homebrew/opt/zlib/lib -L/opt/homebrew/opt/openssl@1.1/lib -Wl,-undefined,dynamic_lookup $(LIBRUBYARG_SHARED)
+ARCH_FLAG = 
+DLDFLAGS = $(ldflags) $(dldflags) $(ARCH_FLAG)
+LDSHARED = $(CC) -dynamic -bundle
+LDSHAREDXX = $(CXX) -dynamic -bundle
+AR = ar
+EXEEXT = 
+
+RUBY_INSTALL_NAME = $(RUBY_BASE_NAME)
+RUBY_SO_NAME = ruby.3.2
+RUBYW_INSTALL_NAME = 
+RUBY_VERSION_NAME = $(RUBY_BASE_NAME)-$(ruby_version)
+RUBYW_BASE_NAME = rubyw
+RUBY_BASE_NAME = ruby
+
+arch = arm64-darwin23
+sitearch = $(arch)
+ruby_version = 3.2.0
+ruby = $(bindir)/$(RUBY_BASE_NAME)
+RUBY = $(ruby)
+BUILTRUBY = $(bindir)/$(RUBY_BASE_NAME)
+ruby_headers = $(hdrdir)/ruby.h $(hdrdir)/ruby/backward.h $(hdrdir)/ruby/ruby.h $(hdrdir)/ruby/defines.h $(hdrdir)/ruby/missing.h $(hdrdir)/ruby/intern.h $(hdrdir)/ruby/st.h $(hdrdir)/ruby/subst.h $(arch_hdrdir)/ruby/config.h
+
+RM = rm -f
+RM_RF = rm -fr
+RMDIRS = rmdir -p
+MAKEDIRS = /opt/homebrew/opt/coreutils/bin/gmkdir -p
+INSTALL = /opt/homebrew/opt/coreutils/bin/ginstall -c
+INSTALL_PROG = $(INSTALL) -m 0755
+INSTALL_DATA = $(INSTALL) -m 644
+COPY = cp
+TOUCH = exit >
+
+#### End of system configuration section. ####
+
+preload = 
+libpath = . $(libdir)
+LIBPATH =  -L. -L$(libdir)
+DEFFILE = 
+
+CLEANFILES = mkmf.log
+DISTCLEANFILES = 
+DISTCLEANDIRS = 
+
+extout = 
+extout_prefix = 
+target_prefix = /smarter_csv
+LOCAL_LIBS = 
+LIBS = $(LIBRUBYARG_SHARED)  -lpthread  
+ORIG_SRCS = smarter_csv.c
+SRCS = $(ORIG_SRCS) 
+OBJS = smarter_csv.o
+HDRS = 
+LOCAL_HDRS = 
+TARGET = smarter_csv
+TARGET_NAME = smarter_csv
+TARGET_ENTRY = Init_$(TARGET_NAME)
+DLLIB = $(TARGET).bundle
+EXTSTATIC = 
+STATIC_LIB = 
+
+TIMESTAMP_DIR = .
+BINDIR        = $(bindir)
+RUBYCOMMONDIR = $(sitedir)$(target_prefix)
+RUBYLIBDIR    = $(sitelibdir)$(target_prefix)
+RUBYARCHDIR   = $(sitearchdir)$(target_prefix)
+HDRDIR        = $(sitehdrdir)$(target_prefix)
+ARCHHDRDIR    = $(sitearchhdrdir)$(target_prefix)
+TARGET_SO_DIR =
+TARGET_SO     = $(TARGET_SO_DIR)$(DLLIB)
+CLEANLIBS     = $(TARGET_SO) $(TARGET_SO).dSYM
+CLEANOBJS     = $(OBJS) *.bak
+TARGET_SO_DIR_TIMESTAMP = $(TIMESTAMP_DIR)/.sitearchdir.-.smarter_csv.time
+
+all:    $(DLLIB)
+static: $(STATIC_LIB)
+.PHONY: all install static install-so install-rb
+.PHONY: clean clean-so clean-static clean-rb
+
+clean-static::
+clean-rb-default::
+clean-rb::
+clean-so::
+clean: clean-so clean-static clean-rb-default clean-rb
+		-$(Q)$(RM_RF) $(CLEANLIBS) $(CLEANOBJS) $(CLEANFILES) .*.time
+
+distclean-rb-default::
+distclean-rb::
+distclean-so::
+distclean-static::
+distclean: clean distclean-so distclean-static distclean-rb-default distclean-rb
+		-$(Q)$(RM) Makefile $(RUBY_EXTCONF_H) conftest.* mkmf.log
+		-$(Q)$(RM) core ruby$(EXEEXT) *~ $(DISTCLEANFILES)
+		-$(Q)$(RMDIRS) $(DISTCLEANDIRS) 2> /dev/null || true
+
+realclean: distclean
+install: install-so install-rb
+
+install-so: $(DLLIB) $(TARGET_SO_DIR_TIMESTAMP)
+	$(INSTALL_PROG) $(DLLIB) $(RUBYARCHDIR)
+clean-static::
+	-$(Q)$(RM) $(STATIC_LIB)
+install-rb: pre-install-rb do-install-rb install-rb-default
+install-rb-default: pre-install-rb-default do-install-rb-default
+pre-install-rb: Makefile
+pre-install-rb-default: Makefile
+do-install-rb:
+do-install-rb-default:
+pre-install-rb-default:
+	@$(NULLCMD)
+$(TARGET_SO_DIR_TIMESTAMP):
+	$(Q) $(MAKEDIRS) $(@D) $(RUBYARCHDIR)
+	$(Q) $(TOUCH) $@
+
+site-install: site-install-so site-install-rb
+site-install-so: install-so
+site-install-rb: install-rb
+
+.SUFFIXES: .c .m .cc .mm .cxx .cpp .o .S
+
+.cc.o:
+	$(ECHO) compiling $(<)
+	$(Q) $(CXX) $(INCFLAGS) $(CPPFLAGS) $(CXXFLAGS) $(COUTFLAG)$@ -c $(CSRCFLAG)$<
+
+.cc.S:
+	$(ECHO) translating $(<)
+	$(Q) $(CXX) $(INCFLAGS) $(CPPFLAGS) $(CXXFLAGS) $(COUTFLAG)$@ -S $(CSRCFLAG)$<
+
+.mm.o:
+	$(ECHO) compiling $(<)
+	$(Q) $(CXX) $(INCFLAGS) $(CPPFLAGS) $(CXXFLAGS) $(COUTFLAG)$@ -c $(CSRCFLAG)$<
+
+.mm.S:
+	$(ECHO) translating $(<)
+	$(Q) $(CXX) $(INCFLAGS) $(CPPFLAGS) $(CXXFLAGS) $(COUTFLAG)$@ -S $(CSRCFLAG)$<
+
+.cxx.o:
+	$(ECHO) compiling $(<)
+	$(Q) $(CXX) $(INCFLAGS) $(CPPFLAGS) $(CXXFLAGS) $(COUTFLAG)$@ -c $(CSRCFLAG)$<
+
+.cxx.S:
+	$(ECHO) translating $(<)
+	$(Q) $(CXX) $(INCFLAGS) $(CPPFLAGS) $(CXXFLAGS) $(COUTFLAG)$@ -S $(CSRCFLAG)$<
+
+.cpp.o:
+	$(ECHO) compiling $(<)
+	$(Q) $(CXX) $(INCFLAGS) $(CPPFLAGS) $(CXXFLAGS) $(COUTFLAG)$@ -c $(CSRCFLAG)$<
+
+.cpp.S:
+	$(ECHO) translating $(<)
+	$(Q) $(CXX) $(INCFLAGS) $(CPPFLAGS) $(CXXFLAGS) $(COUTFLAG)$@ -S $(CSRCFLAG)$<
+
+.c.o:
+	$(ECHO) compiling $(<)
+	$(Q) $(CC) $(INCFLAGS) $(CPPFLAGS) $(CFLAGS) $(COUTFLAG)$@ -c $(CSRCFLAG)$<
+
+.c.S:
+	$(ECHO) translating $(<)
+	$(Q) $(CC) $(INCFLAGS) $(CPPFLAGS) $(CFLAGS) $(COUTFLAG)$@ -S $(CSRCFLAG)$<
+
+.m.o:
+	$(ECHO) compiling $(<)
+	$(Q) $(CC) $(INCFLAGS) $(CPPFLAGS) $(CFLAGS) $(COUTFLAG)$@ -c $(CSRCFLAG)$<
+
+.m.S:
+	$(ECHO) translating $(<)
+	$(Q) $(CC) $(INCFLAGS) $(CPPFLAGS) $(CFLAGS) $(COUTFLAG)$@ -S $(CSRCFLAG)$<
+
+$(TARGET_SO): $(OBJS) Makefile
+	$(ECHO) linking shared-object smarter_csv/$(DLLIB)
+	-$(Q)$(RM) $(@)
+	$(Q) $(LDSHARED) -o $@ $(OBJS) $(LIBPATH) $(DLDFLAGS) $(LOCAL_LIBS) $(LIBS)
+	$(Q) $(POSTLINK)
+
+
+
+$(OBJS): $(HDRS) $(ruby_headers)

data/ext/smarter_csv/extconf.rb

@@ -9,6 +9,8 @@ if RbConfig::MAKEFILE_CONFIG["CFLAGS"].include?("-g -O3")
   RbConfig::MAKEFILE_CONFIG["CFLAGS"] = fixed_CFLAGS
 end
 
-CONFIG["optflags"] = "-O3"
+# CONFIG["optflags"] = "-O3 -march=native -flto"
+CONFIG["optflags"] = "-O3 -march=native -flto -fomit-frame-pointer -DNDEBUG"
+CONFIG["debugflags"] = ""
 
 create_makefile('smarter_csv/smarter_csv')

data/ext/smarter_csv/smarter_csv.c

@@ -2,6 +2,7 @@
 #include "ruby/encoding.h"
 #include <stdio.h>
 #include <stdbool.h>
+#include <string.h>
 
 #ifndef bool
   #define bool int
@@ -12,8 +13,25 @@
 VALUE SmarterCSV = Qnil;
 VALUE eMalformedCSVError = Qnil;
 VALUE Parser = Qnil;
+VALUE Qempty_string = Qnil; // shared frozen empty string
+
+static VALUE unescape_quotes(char *str, long len, char quote_char, rb_encoding *encoding) {
+  char *buf = ALLOC_N(char, len);
+  long j = 0;
+  for (long i = 0; i < len; i++) {
+    if (str[i] == quote_char && i + 1 < len && str[i + 1] == quote_char) {
+      buf[j++] = quote_char;
+      i++; // skip second quote
+    } else {
+      buf[j++] = str[i];
+    }
+  }
+  VALUE out = rb_enc_str_new(buf, j, encoding);
+  xfree(buf);
+  return out;
+}
 
-static VALUE rb_parse_csv_line(VALUE self, VALUE line, VALUE col_sep, VALUE quote_char, VALUE max_size) {
+static VALUE rb_parse_csv_line(VALUE self, VALUE line, VALUE col_sep, VALUE quote_char, VALUE max_size, VALUE has_quotes_val, VALUE strip_ws_val) {
   if (RB_TYPE_P(line, T_NIL) == 1) {
     return rb_ary_new();
   }
@@ -22,74 +40,180 @@ static VALUE rb_parse_csv_line(VALUE self, VALUE line, VALUE col_sep, VALUE quot
     rb_raise(rb_eTypeError, "ERROR in SmarterCSV.parse_line: line has to be a string or nil");
   }
 
-  rb_encoding *encoding = rb_enc_get(line); /* get the encoding from the input line */
-  char *startP = RSTRING_PTR(line); /* may not be null terminated */
+  rb_encoding *encoding = rb_enc_get(line);
+  char *startP = RSTRING_PTR(line);
   long line_len = RSTRING_LEN(line);
-  char *endP = startP + line_len; /* points behind the string */
+  char *endP = startP + line_len;
   char *p = startP;
 
   char *col_sepP = RSTRING_PTR(col_sep);
   long col_sep_len = RSTRING_LEN(col_sep);
 
   char *quoteP = RSTRING_PTR(quote_char);
-  long quote_count = 0;
-
-  bool col_sep_found = true;
+  char quote_char_val = quoteP[0];
+  size_t quote_len = strlen(quoteP);
 
   VALUE elements = rb_ary_new();
   VALUE field;
-  long i;
 
-  /* Variables for escaped quote handling */
+  long element_count = 0;
+  int max_fields = -1;
+  if (max_size != Qnil) {
+    max_fields = NUM2INT(max_size);
+    if (max_fields < 0) {
+      return rb_ary_new();
+    }
+  }
+
+  bool has_quotes = RTEST(has_quotes_val);
+  bool strip_ws = RTEST(strip_ws_val);
+
+  // === FAST PATH: No quotes and single-character separator ===
+  if (__builtin_expect(!has_quotes && col_sep_len == 1, 1)) {
+    char sep = *col_sepP;
+    char *sep_pos = NULL;
+
+    while ((sep_pos = memchr(p, sep, endP - p))) {
+      if ((max_fields >= 0) && (element_count >= max_fields)) {
+        break;
+      }
+
+      long field_len = sep_pos - startP;
+      char *raw_field = startP;
+      char *trim_start = raw_field;
+      char *trim_end = raw_field + field_len - 1;
+
+      if (strip_ws) {
+        while (trim_start <= trim_end && (*trim_start == ' ' || *trim_start == '\t')) trim_start++;
+        while (trim_end >= trim_start && (*trim_end == ' ' || *trim_end == '\t')) trim_end--;
+      }
+
+      long trimmed_len = trim_end - trim_start + 1;
+
+      field = rb_enc_str_new(trim_start, trimmed_len, encoding);
+      rb_ary_push(elements, field);
+      element_count++;
+
+      p = sep_pos + 1;
+      startP = p;
+    }
+
+    if ((max_fields < 0) || (element_count < max_fields)) {
+      long field_len = endP - startP;
+      char *raw_field = startP;
+      char *trim_start = raw_field;
+      char *trim_end = raw_field + field_len - 1;
+
+      if (strip_ws) {
+        while (trim_start <= trim_end && (*trim_start == ' ' || *trim_start == '\t')) trim_start++;
+        while (trim_end >= trim_start && (*trim_end == ' ' || *trim_end == '\t')) trim_end--;
+      }
+
+      long trimmed_len = trim_end - trim_start + 1;
+
+      field = rb_enc_str_new(trim_start, trimmed_len, encoding);
+      rb_ary_push(elements, field);
+    }
+
+    return elements;
+  }
+
+  // === SLOW PATH: Quoted fields or multi-char separator ===
+  long i;
   long backslash_count = 0;
   bool in_quotes = false;
+  bool col_sep_found = true;
 
   while (p < endP) {
-    /* does the remaining string start with col_sep ? */
     col_sep_found = true;
-    for(i=0; (i < col_sep_len) && (p+i < endP); i++) {
-      col_sep_found = col_sep_found && (*(p+i) == *(col_sepP+i));
+    for (i = 0; (i < col_sep_len) && (p + i < endP); i++) {
+      if (*(p + i) != *(col_sepP + i)) {
+        col_sep_found = false;
+        break;
+      }
     }
-    /* if col_sep was found and we're not inside quotes */
+
     if (col_sep_found && !in_quotes) {
-      /* if max_size != nil && elements.size >= header_size */
-      if ((max_size != Qnil) && RARRAY_LEN(elements) >= NUM2INT(max_size)) {
+      if ((max_fields >= 0) && (element_count >= max_fields)) {
         break;
-      } else {
-        /* push that field with original encoding onto the results */
-        field = rb_enc_str_new(startP, p - startP, encoding);
-        rb_ary_push(elements, field);
+      }
+
+      long field_len = p - startP;
+      char *raw_field = startP;
 
-        p += col_sep_len;
-        startP = p;
-        backslash_count = 0; // Reset backslash count at the start of a new field
+      bool quoted = (field_len >= 2 && raw_field[0] == quote_char_val && raw_field[field_len - 1] == quote_char_val);
+      if (quoted) {
+        raw_field++;
+        field_len -= 2;
+      }
+
+      char *trim_start = raw_field;
+      char *trim_end = raw_field + field_len - 1;
+
+      if (strip_ws) {
+        while (trim_start <= trim_end && (*trim_start == ' ' || *trim_start == '\t')) trim_start++;
+        while (trim_end >= trim_start && (*trim_end == ' ' || *trim_end == '\t')) trim_end--;
+      }
+
+      long trimmed_len = trim_end - trim_start + 1;
+
+      if (quoted || memchr(trim_start, quote_char_val, trimmed_len)) {
+        field = unescape_quotes(trim_start, trimmed_len, quote_char_val, encoding);
+      } else {
+        field = rb_enc_str_new(trim_start, trimmed_len, encoding);
       }
+
+      rb_ary_push(elements, field);
+      element_count++;
+
+      p += col_sep_len;
+      startP = p;
+      backslash_count = 0;
     } else {
       if (*p == '\\') {
         backslash_count++;
       } else {
-        if (*p == *quoteP) {
+        if (*p == quote_char_val) {
           if (backslash_count % 2 == 0) {
-            /* Even number of backslashes means quote is not escaped */
             in_quotes = !in_quotes;
           }
-          /* Else, quote is escaped; do nothing */
         }
-        backslash_count = 0; // Reset after any character other than backslash
+        backslash_count = 0;
       }
       p++;
     }
-  } /* while */
+  }
 
-  /* Check for unclosed quotes at the end of the line */
   if (in_quotes) {
     rb_raise(eMalformedCSVError, "Unclosed quoted field detected in line: %s", StringValueCStr(line));
   }
 
-  /* check if the last part of the line needs to be processed */
-  if ((max_size == Qnil) || RARRAY_LEN(elements) < NUM2INT(max_size)) {
-    /* copy the remaining line as a field with original encoding onto the results */
-    field = rb_enc_str_new(startP, endP - startP, encoding);
+  if ((max_fields < 0) || (element_count < max_fields)) {
+    long field_len = endP - startP;
+    char *raw_field = startP;
+
+    bool quoted = (field_len >= 2 && raw_field[0] == quote_char_val && raw_field[field_len - 1] == quote_char_val);
+    if (quoted) {
+      raw_field++;
+      field_len -= 2;
+    }
+
+    char *trim_start = raw_field;
+    char *trim_end = raw_field + field_len - 1;
+
+    if (strip_ws) {
+      while (trim_start <= trim_end && (*trim_start == ' ' || *trim_start == '\t')) trim_start++;
+      while (trim_end >= trim_start && (*trim_end == ' ' || *trim_end == '\t')) trim_end--;
+    }
+
+    long trimmed_len = trim_end - trim_start + 1;
+
+    if (quoted || memchr(trim_start, quote_char_val, trimmed_len)) {
+      field = unescape_quotes(trim_start, trimmed_len, quote_char_val, encoding);
+    } else {
+      field = rb_enc_str_new(trim_start, trimmed_len, encoding);
+    }
+
     rb_ary_push(elements, field);
   }
 
@@ -97,10 +221,10 @@ static VALUE rb_parse_csv_line(VALUE self, VALUE line, VALUE col_sep, VALUE quot
 }
 
 void Init_smarter_csv(void) {
-  // these modules and the error class are already defined in Ruby code, make them accessible:
   SmarterCSV = rb_const_get(rb_cObject, rb_intern("SmarterCSV"));
   Parser = rb_const_get(SmarterCSV, rb_intern("Parser"));
   eMalformedCSVError = rb_const_get(SmarterCSV, rb_intern("MalformedCSV"));
-
-  rb_define_module_function(Parser, "parse_csv_line_c", rb_parse_csv_line, 4);
+  Qempty_string = rb_str_new_literal("");
+  rb_gc_register_address(&Qempty_string);
+  rb_define_module_function(Parser, "parse_csv_line_c", rb_parse_csv_line, 6);
 }