[PATCH 06/11] diagnostics: learn how to count column number with multibyte chars

[Akim Demaille]

So far diagnostics were cheating: in addition to the 'column' field of
locations (based on actual screen width per multibyte characters and
on tabulation expansion), the scanner sets the 'byte' field.
Diagnostics used this byte count to decide where to insert (color)
style.

We want to be able to truncate the quoted lines when there are too
wide to fit the screen.  This requires that the diagnostics learn how
to count columns, the byte-in-boundary trick no longer works.

Bytes are still used for fix-its.

* bootstrap.conf: We need mbfile for mbf_getc.
* src/location.c (caret_info): We need an mbfile.
(caret_set_file): Initialize it.
(caret_getc): Convert to mbfile.
(location_caret): Instead of relying on the byte position to decide
where to insert the color style, count the current column using
boundary_compute.
---
 bootstrap.conf |  2 +-
 lib/.gitignore |  5 ++++
 m4/.gitignore  |  3 +++
 src/location.c | 62 ++++++++++++++++++++++++++++----------------------
 4 files changed, 44 insertions(+), 28 deletions(-)

diff --git a/bootstrap.conf b/bootstrap.conf
index 44469581..756d798d 100644
--- a/bootstrap.conf
+++ b/bootstrap.conf
@@ -30,7 +30,7 @@ gnulib_modules='
   ldexpl
   libtextstyle-optional
   malloc-gnu
-  mbswidth
+  mbfile mbswidth
   non-recursive-gnulib-prefix-hack
   obstack
   obstack-printf
diff --git a/lib/.gitignore b/lib/.gitignore
index 70f13cd6..533d5dbc 100644
--- a/lib/.gitignore
+++ b/lib/.gitignore
@@ -132,6 +132,7 @@
 /isnanf.c
 /isnanl-nolibm.h
 /isnanl.c
+/iswblank.c
 /itold.c
 /ldexpl.c
 /libc-config.h
@@ -148,6 +149,10 @@
 /math.c
 /math.h
 /math.in.h
+/mbchar.c
+/mbchar.h
+/mbfile.c
+/mbfile.h
 /mbrtowc.c
 /mbsinit.c
 /mbswidth.c
diff --git a/m4/.gitignore b/m4/.gitignore
index 0d5b7472..286437d0 100644
--- a/m4/.gitignore
+++ b/m4/.gitignore
@@ -67,6 +67,7 @@
 /isnand.m4
 /isnanf.m4
 /isnanl.m4
+/iswblank.m4
 /javacomp.m4
 /javaexec.m4
 /largefile.m4
@@ -91,6 +92,8 @@
 /malloc.m4
 /malloca.m4
 /math_h.m4
+/mbchar.m4
+/mbfile.m4
 /mbrtowc.m4
 /mbsinit.m4
 /mbstate_t.m4
diff --git a/src/location.c b/src/location.c
index ca2d0b7f..66361fc0 100644
--- a/src/location.c
+++ b/src/location.c
@@ -21,6 +21,7 @@
 #include <config.h>
 #include "system.h"
 
+#include <mbfile.h>
 #include <mbswidth.h>
 #include <quotearg.h>
 #include <stdio.h>    /* fileno */
@@ -167,6 +168,8 @@ static struct
 {
   /* Raw input file.  */
   FILE *file;
+  /* Input file as a stream of multibyte characters.  */
+  mb_file_t mbfile;
   /* The position within the last file we quoted.  If POS.FILE is non
      NULL, but FILE is NULL, it means this file is special and should
      not be quoted. */
@@ -201,7 +204,7 @@ caret_set_file (const char *file)
               && buf.st_mode & S_IFREG)
             {
               caret_info.pos.line = 1;
-              caret_info.offset = 0;
+              mbf_init (caret_info.mbfile, caret_info.file);
             }
           else
             caret_free ();
@@ -221,22 +224,23 @@ caret_free (void)
 }
 
 /* Getc, but smash \r\n as \n.  */
-static int
-caret_getc (void)
+static void
+caret_getc_internal (mbchar_t *res)
 {
-  FILE *f = caret_info.file;
-  int res = getc (f);
-  if (res == '\r')
+  mbf_getc (*res, caret_info.mbfile);
+  if (mb_iseq (*res, '\r'))
     {
-      int c = getc (f);
-      if (c == '\n')
-        res = c;
+      mbchar_t c;
+      mbf_getc (c, caret_info.mbfile);
+      if (mb_iseq (c, '\n'))
+        mb_copy (res, &c);
       else
-        ungetc (c, f);
+        mbf_ungetc (c, caret_info.mbfile);
     }
-  return res;
 }
 
+#define caret_getc(Var) caret_getc_internal(&Var)
+
 void
 location_caret (location loc, const char *style, FILE *out)
 {
@@ -259,19 +263,25 @@ location_caret (location loc, const char *style, FILE 
*out)
   /* Advance to the line's position, keeping track of the offset.  */
   while (caret_info.pos.line < loc.start.line)
     {
-      int c = caret_getc ();
-      if (c == EOF)
+      mbchar_t c;
+      caret_getc (c);
+      if (mb_iseof (c))
         /* Something is wrong, that line number does not exist.  */
         return;
-      caret_info.pos.line += c == '\n';
+      caret_info.pos.line += mb_iseq (c, '\n');
     }
   caret_info.offset = ftell (caret_info.file);
+  caret_info.pos.column = 1;
+  /* Reset mbf's internal state.
+     FIXME: should be done in mbfile.  */
+  caret_info.mbfile.eof_seen = 0;
 
   /* Read the actual line.  Don't update the offset, so that we keep a pointer
      to the start of the line.  */
   {
-    int c = caret_getc ();
-    if (c != EOF)
+    mbchar_t c;
+    caret_getc (c);
+    if (!mb_iseof (c))
       {
         bool single_line = loc.start.line == loc.end.line;
         /* Quote the file (at most the first line in the case of
@@ -280,29 +290,27 @@ location_caret (location loc, const char *style, FILE 
*out)
           fprintf (out, "%5d | ", loc.start.line);
           /* Consider that single point location (with equal boundaries)
              actually denote the character that they follow.  */
-          int byte_end = loc.end.byte +
-            (single_line && loc.start.byte == loc.end.byte);
-          /* Byte number.  */
-          int byte = 1;
+          int col_end = loc.end.column +
+            (single_line && loc.start.column == loc.end.column);
           /* Whether we opened the style.  If the line is not as
              expected (maybe the file was changed since the scanner
              ran), we might reach the end before we actually saw the
              opening column.  */
           bool opened = false;
-          while (c != EOF && c != '\n')
+          while (!mb_iseof (c) && !mb_iseq (c, '\n'))
             {
-              if (byte == loc.start.byte)
+              if (caret_info.pos.column == loc.start.column)
                 {
                   begin_use_class (style, out);
                   opened = true;
                 }
-              fputc (c, out);
-              c = caret_getc ();
-              ++byte;
+              mb_putc (c, out);
+              boundary_compute (&caret_info.pos, mb_ptr (c), mb_len (c));
+              caret_getc (c);
               if (opened
                   && (single_line
-                      ? byte == byte_end
-                      : c == '\n' || c == EOF))
+                      ? caret_info.pos.column == col_end
+                      : mb_iseq (c, '\n') || mb_iseof (c)))
                 end_use_class (style, out);
             }
           putc ('\n', out);
-- 
2.23.0