From b2edeb4ad108e03f1311ef67f6bed72c49ada460 Mon Sep 17 00:00:00 2001
From: Stephen Heumann <stephenheumann@gmail.com>
Date: Fri, 25 Mar 2022 18:10:13 -0500
Subject: [PATCH] Properly stringize tokens that start with a trigraph.

This did not work correctly before, because such tokens were recorded as starting with the third character of the trigraph.

Here is an example affected by this:

#define mkstr(a) # a
#include <stdio.h>
int main(void) {
        puts(mkstr(??!));
        puts(mkstr(??!??!));
        puts(mkstr('??<'));
        puts(mkstr(+??!));
        puts(mkstr(+??'));
}
---
 Scanner.asm | 11 ++++++++---
 Scanner.pas | 10 +++++++---
 cc.notes    |  2 +-
 3 files changed, 16 insertions(+), 7 deletions(-)

diff --git a/Scanner.asm b/Scanner.asm
index 5409001..436e9be 100644
--- a/Scanner.asm
+++ b/Scanner.asm
@@ -446,6 +446,7 @@ rkModifiers ds 2
 *
 *  Outputs:
 *        ch - character read
+*        currentChPtr - pointer to ch in source file
 *
 ****************************************************************
 *
@@ -493,12 +494,15 @@ pf1      dey
 pf2      sty   lastWasReturn
 ! 1:
 lab1     anop
+! currentChPtr := chPtr;
 ! if chPtr = eofPtr then begin          {flag end of file if we're there}
          lda   chPtr
+         sta   currentChPtr
+         ldx   chPtr+2
+         stx   currentChPtr+2
          cmp   eofPtr
          bne   la1
-         lda   chPtr+2
-         cmp   eofPtr+2
+         cpx   eofPtr+2
          beq   la2
 la1      brl   lb5
 la2      anop
@@ -621,7 +625,8 @@ lb4      lda   [p1],Y
 ! else begin
 lb5      anop
 !    ch := chr(chPtr^);                 {fetch the character}
-         move4 chPtr,p1
+         sta   p1
+         stx   p1+2
          lda   [p1]
          and   #$00FF
          sta   ch
diff --git a/Scanner.pas b/Scanner.pas
index 3e86b1f..88abf9b 100644
--- a/Scanner.pas
+++ b/Scanner.pas
@@ -144,6 +144,7 @@ procedure NextCh; extern;
 {                                                               }
 { Globals:                                                      }
 {       ch - character read                                     }
+{       currentChPtr - pointer to ch in source file             }
 
 
 procedure NextToken;
@@ -240,6 +241,7 @@ type
 
 var
    charStrPrefix: charStrPrefixEnum;    {prefix of character/string literal}
+   currentChPtr: ptr;                   {pointer to current character in source file}
    customDefaultName: stringPtr;        {name of custom pre-included default file}
    dateStr: longStringPtr;              {macro date string}
    doingCommandLine: boolean;           {are we processing the cc= command line?}
@@ -2269,6 +2271,7 @@ if gotName then begin			{read the file name from the line}
    changedSourceFile := true;
    ReadFile;				{read the file}
    chPtr := bofPtr;			{set the start, end pointers}
+   currentChPtr := bofPtr;
    eofPtr := pointer(ord4(bofPtr)+ffDCBGS.fileLength);
    firstPtr := chPtr;			{first char in line}
    ch := chr(RETURN);			{set the initial character}
@@ -4188,6 +4191,7 @@ expandMacros := true;                   {enable macro expansion}
 reportEOL := false;                     {report eolsy as a token?}
 lineNumber := 1;                        {start the line counter}
 chPtr := start;                         {set the start, end pointers}
+currentChPtr := start;
 eofPtr := endPtr;
 firstPtr := start;                      {first char in line}
 numErr := 0;                            {no errors so far}
@@ -4942,8 +4946,8 @@ while charKinds[ord(ch)] in [illegal,ch_white,ch_eol] do begin
       end;
    end; {while}
 tokenLine := lineNumber;                {record the position of the token}
-tokenColumn := ord(ord4(chPtr)-ord4(firstPtr));
-tokenStart := pointer(ord4(chPtr)-1);
+tokenColumn := ord(ord4(currentChPtr)-ord4(firstPtr)+1);
+tokenStart := currentChPtr;
 6:
 token.class := reservedSymbol;          {default to the most common class}
 case charKinds[ord(ch)] of
@@ -5366,7 +5370,7 @@ case charKinds[ord(ch)] of
 
    otherwise: Error(57);
    end; {case}
-tokenEnd := pointer(ord4(chPtr)-1);     {record the end of the token}
+tokenEnd := currentChPtr;               {record the end of the token}
 2:
 if skipping then                        {conditional compilation branch}
    if not (token.kind in [eofsy,eolsy]) then
diff --git a/cc.notes b/cc.notes
index abd6421..7de66df 100644
--- a/cc.notes
+++ b/cc.notes
@@ -1784,7 +1784,7 @@ int foo(int[42]);
 
 182. #pragma path directives were not saved in .sym files.  This could cause ORCA/C not to search the proper paths for include files that were not represented in the .sym file (e.g. because they were included after a function).
 
-183. The # preprocessor operator would not work correctly on tokens that had been produced by the ## preprocessor operator, or on tokens that were split over two or more lines using line continuations.
+183. The # preprocessor operator would not work correctly on tokens that had been produced by the ## preprocessor operator, tokens that were split over two or more lines using line continuations, or tokens represented using trigraphs.
 
 -- Bugs from C 2.1.0 that have been fixed -----------------------------------