slang2

slang2のバグ退治

• Etch: 2.0.6-4
• Lenny: 2.0.7-3
• Sid: 2.1.3-2

現象

• 幅計算がおかしい？ #425835
  • xtermだとうまくいくような…。端末の問題？
  • btermはbogl-bterm, bf-utf-source, libbogl-devをインストールし、bdftobogl -b /usr/src/unifont.bdf > font.bgf でフォント作成。「bterm -f font.bgf -l C.UTF-8」で呼び出したのがインストーラと同じもの。

    #!/bin/sh
    if [ "$1" != "" ]; then
      LD_PRELOAD=$1
      export LD_PRELOAD
    fi
    whiptail --menu test 20 50 5 2 'sdasdsd' 3 'dfdfdsfds' 4 'dfdsfsdfdsfdfds' 1 '日本語' 2 'one more line' 3 'other line' 4 'and other more'
    

目標

• 幅計算を正す
• #399355と#369152のテストケースが通るように。現状のsid版(パッチ当たってない版)では通るので、これを壊さないように
• 端末側の問題の可能性がある。VMwareでbtermを使いテストする必要がある。unifontを作る必要ありか。

前ぶれ

• slang2 (2.0.7-1) unstable; urgency=lowで、「* Backout 080_wide_chars.patch ; improved fix in 2.0.7. Closes: #399355.」
• slang2 (2.0.6-4) unstable; urgency=highで、「* Fix regression in slutf8.c due to removing 080_wide_chars.patch that broke double-width chars. Fixed bugs in patch instead. Closes: #392987, #392942.」
• slang2 (2.0.6-3) unstable; urgency=lowで、「* Remove 080_wide_chars.patch which breaks jed. Closes: #369152.」
• #399355: de guyから、jedが壊れたとな。http://www.jedsoft.org/pipermail/slang-users_jedsoft.org/2006/000482.html

  #include <slang.h>
  #include <stdio.h>
  
  int main(void)
  {
      SLuchar_Type string[] = "\xE4";
      unsigned int dnum;
  
      SLutf8_skip_chars(string, string+1, 1, &dnum, 1);
  
      printf("%d\n", dnum);
  
      return 0;
  }
  
  

• #392987, #392942: It looks like functions SLsmg_write_string and SLsmg_write_nstring now work incorrecly. #384787と同じ？ バグ詳細あり。

  slsmg.c contains:
  void SLsmg_write_nstring (char *str, unsigned int n)
  {
  ...
          if (UTF8_Mode)
            umax = SLutf8_skip_chars (u, u+width, n, &width, 0);
  ...
          SLsmg_write_chars (u, umax);
       }
  
     while (width++ < n) SLsmg_write_chars (blank, blank+1);
  }
  
  from last line: 'n' is string width.
  
  Let's look at SLutf8_skip_chars, in slutf8.c:
  
  SLuchar_Type *SLutf8_skip_chars (SLuchar_Type *s, SLuchar_Type *smax,
                                    unsigned int num, unsigned int *dnum,
                                    int ignore_combining)
  {
     unsigned int n;
  
     n = 0;
     while ((n < num) && (s < smax))
     {
  ...
     /* here code that modifies n always adds 1, nothing depends on
   * character width */
  ...
  
  
  So here num, aka n in previous function, is used as number of
  characters and not as string width.
  

• #384787:newt: incorrectly handles non-ASCII characters in input boxes under UTF-8 locale: whiptail --inputbox text 10 40 On my system I get input field that is wider than dialog box: これは直ってるような？
• #369152: libslang2: 080_wide_chars breakes slang: you've added the patch 080_wide_chars (why?) which breaks the function SLutf8_skip_chars().This example should skip 1 character, but returns in dnum it has skipped 2 characters. Confusing.

  #include <slang.h>
  #include <stdio.h>
  
  int main(void)
  {
      SLuchar_Type string[] = "\xE3\x82\xA6\xE3\x82\xA7\xE3\x83\x96\xE3";
      unsigned int dnum;
  
      SLutf8_skip_chars(string, string+9, 1, &dnum, 1);
  
      printf("%d\n", dnum);
  
      return 0;
  }
  

080_wide_charsのパッチ

#

diff -ruN slang-2.0.4-old/src/slutf8.c slang-2.0.4/src/slutf8.c
--- slang-2.0.4-old/src/slutf8.c        2005-06-20 18:56:22.000000000 +0300
+++ slang-2.0.4/src/slutf8.c    2005-07-14 18:23:07.456402952 +0300
@@ -148,19 +148,15 @@
             continue;
          }
        
-       if (ignore_combining)
-         {
+             int wi;
             SLwchar_Type w = fast_utf8_decode (s, len);
-            if (0 != SLwchar_wcwidth (w))
-              n++;
+            wi = SLwchar_wcwidth (w);
+            n+=wi;
             s += len;
             continue;
-         }
-
-       n++;
-       s += len;
      }
-
+#if 0
+   /* XXX how this should work???? */
    if (ignore_combining)
      {
        while (s < smax)
@@ -176,7 +172,7 @@
             s += nconsumed;
          }
      }
-
+#endif
    if (dnum != NULL)
      *dnum = n;
    return s;