Various updates

Various updates, check for empty field names bugfix, make it easier to enter fixed width column positions, for issue #46 render skipped lines correctly

CSVLintNppPlugin/CsvLint/CsvAnalyze.cs

@@ -12,6 +12,7 @@
 using CSVLint.Tools;
 using CsvQuery.PluginInfrastructure;
 using Kbg.NppPluginNET.PluginInfrastructure;
+using System.Diagnostics;
 
 namespace CSVLint
 {
@@ -315,7 +316,7 @@ public static CsvDefinition InferFromData(bool autodetect, char mansep, string m
                 foundfieldWidths.Sort();
                 if (foundfieldWidths.Count < 3) return result; // unlikely fixed width
 
-                // widths contain line positions, convert to individual column widths, example pos [8, 14, 15, 22, 25] -> widths [8, 6, 1, 7, 3]
+                // widths now contain column end positions, convert to individual column widths, example pos [8, 14, 15, 22, 25] -> widths [8, 6, 1, 7, 3]
                 var pos1 = 0;
                 for (var i = 0; i < foundfieldWidths.Count; i++)
                 {
@@ -408,11 +409,11 @@ public static CsvDefinition InferFromData(bool autodetect, char mansep, string m
                     // if value in first row (=Names) is not of valid datatype, then first row probably contains column names
                     var str = csvvalid.EvaluateDataValue(namcol.Name, namcol, namcol.Index);
                     if (str != "") count++;
-
-                    // if value in first row (=Names) is empty then probably not header names
-                    if (namcol.Name == "") emptyname = true;
                 }
 
+                // if value in first row (=Names) is empty then probably not header names
+                if (namcol.Name == "") emptyname = true;
+
                 // TODO: carriage returns in header name not supported
                 // replace with space so that schema.ini can at least be validated
                 namcol.Name = namcol.Name.Replace("\r\n", " ").Replace('\r', ' ').Replace('\n', ' ');

CSVLintNppPlugin/Forms/DetectColumnsForm.cs

@@ -4,6 +4,7 @@
 using System.ComponentModel;
 using System.Data;
 using System.Drawing;
+using System.Linq;
 using System.Text;
 using System.Windows.Forms;
 
@@ -30,11 +31,61 @@ public void InitialiseSetting()
             numSkipLines.Value = Main.Settings.DetectSkipLines;
         }
 
+        private string GetProcessedColWidths()
+        {
+            // Quality-of-life extras, pre-process the list of col positions:
+            // 1) allow both comma and space separated
+            // 2) expect absolute positions, but also allow column widths
+            // 3) remove 0 as start position, because it is always implicitly expected
+            var ret = "";
+
+            // 1) Replace = allow comma separated "10, 12, 15, 20" and space separated "10 12 15 20" and semicolon separated
+            var strvalues = txtFixedWidthPos.Text.Replace(' ', ',').Replace(';', ',').Replace(",,", ",");
+
+            // 2) widths should contain column end positions, check if user entered column widths instead
+            // example expected pos [8, 14, 15, 22, 25] -> but user entered [8, 6, 1, 7, 3]
+            // This is recognisable, due to larger values coming before smaller values
+
+            // filter out invalid integers https://stackoverflow.com/a/2959329/1745616
+            List<int> ints = (from field in strvalues.Split(',').Where((x) => { int dummy; return Int32.TryParse(x, out dummy); }) select Int32.Parse(field)).ToList();
+
+            // This is recognisable, due to larger values coming before smaller values
+            var cl = 0; //count larger before smaller
+            for (int i = 0; i < ints.Count - 1; i++)
+            {
+                for (int j = i + 1; j < ints.Count; j++)
+                {
+                    if (ints[i] >= ints[j]) cl++;
+                }
+            }
+
+            // if more than 0 larger values before smaller values
+            if (cl > 0)
+            {
+                // assume user entered column widths instead of positions, change all width to positions
+                var endpos = 0;
+                for (int i = 0; i < ints.Count; i++)
+                {
+                    endpos += ints[i];
+                    ints[i] = endpos;
+                }
+            }
+
+            // 3) remove 0 as start position, because it is always implicitly expected
+            if (ints[0] == 0) ints.RemoveAt(0);
+            for (int i = 0; i < ints.Count; i++)
+            {
+                ret = ret + (i > 0 ? "," : "") + ints[i].ToString();
+            }
+
+            return ret;
+        }
+
         private void DetectColumnsForm_FormClosing(object sender, FormClosingEventArgs e)
         {
             // pass new values to previous form
             Separator = (cmbColumnSeparator.Text.Length > 0 ? cmbColumnSeparator.Text[0] : '\0');
-            ManWidths = txtFixedWidthPos.Text.Replace(' ', ',').Replace(",,", ","); // Replace = allow both comma separated "10, 12, 15, 20" and space separated "10 12 15 20"
+            ManWidths = GetProcessedColWidths();
             HeaderNames = chkHeaderNames.Checked;
             SkipLines = Convert.ToInt32(numSkipLines.Value);
 

CSVLintNppPlugin/PluginInfrastructure/Lexer.cs

@@ -535,33 +535,11 @@ public static void Lex(IntPtr instance, UIntPtr start_pos, IntPtr length_doc, in
             if (lineCurrent < skipLines)
             {
                 //i = (int)vtable.PositionFromLine(p_access, (IntPtr)skipLines);
-
-                var skipcount = skipLines;
-                i = start;
-                isEOL = false;
-
-                // skip the first X lines
-                while ((skipcount > 0) && (i < length))
-                {
-                    // next character
-                    byte cur = contentBytes[i];
-                    i++;
-
-                    // new line can be single character \r or \n or two characters \r\n
-                    if ((cur == '\n') || (cur == '\r'))
-                    {
-                        if (!isEOL) skipcount--;
-                        isEOL = true;
-                    }
-                    else
-                    {
-                        isEOL = false;
-                    }
-                }
+                i = (int)vtable.LineStart(p_access, (IntPtr)skipLines);
 
                 // set no color
                 vtable.StartStyling(p_access, (IntPtr)(start));
-                vtable.SetStyleFor(p_access, (IntPtr)(i), (char)0);
+                vtable.SetStyleFor(p_access, (IntPtr)(i-start), (char)0);
             }
 
 
@@ -640,7 +618,7 @@ public static void Lex(IntPtr instance, UIntPtr start_pos, IntPtr length_doc, in
                 bool whitespace = true; // to catch where value is just two quotes "" right at start of line
 
                 // fixed widths
-                while (i < length-1)
+                while (i < length - 1)
                 {
                     byte cur = contentBytes[i];
                     byte next = contentBytes[i + 1];
@@ -689,7 +667,7 @@ public static void Lex(IntPtr instance, UIntPtr start_pos, IntPtr length_doc, in
                         if (!quote)
                         {
                             // next color
-                            if ((idx++ > IDX_MAX) || isEOL) idx = 1; // reset end of line
+                            if ((++idx > IDX_MAX) || isEOL) idx = 1; // reset end of line
                         }
 
                         if (isEOL)
@@ -724,7 +702,6 @@ public static void Lex(IntPtr instance, UIntPtr start_pos, IntPtr length_doc, in
                 }
             }
 
-
             // free allocated buffer
             Marshal.FreeHGlobal(buffer_ptr);
         }