From ce4b7363932c1152060ba4b24d9a7e7bf438d4e4 Mon Sep 17 00:00:00 2001
From: albert-github <albert.tests@gmail.com>
Date: Tue, 5 Nov 2019 18:52:56 +0100
Subject: Simple nospam email address

When having a simple nospam email address coded like:
```
/** \file
 * Big Unknown <big at none dot com>
 */
```
this results in the warning:
```
aa.c:2: warning: Unsupported xml/html tag <big> found
```
and in the HTML output it looks like:
```
Big Unknown <big at none dot com>
```
instead of:
```
Big Unknown <big at none dot com>
```

(Found in a couple of open source projects).
---
 src/doctokenizer.l | 5 +++++
 1 file changed, 5 insertions(+)

diff --git a/src/doctokenizer.l b/src/doctokenizer.l
index 22b14a0..59a0a9d 100644
--- a/src/doctokenizer.l
+++ b/src/doctokenizer.l
@@ -345,6 +345,7 @@ CITESCHAR [a-z_A-Z0-9\x80-\xFF\-\?]
 CITEECHAR [a-z_A-Z0-9\x80-\xFF\-\+:\/\?]
 CITEID    {CITESCHAR}{CITEECHAR}*("."{CITESCHAR}{CITEECHAR}*)*|"\""{CITESCHAR}{CITEECHAR}*("."{CITESCHAR}{CITEECHAR}*)*"\""
 MAILADR   ("mailto:")?[a-z_A-Z0-9.+-]+"@"[a-z_A-Z0-9-]+("."[a-z_A-Z0-9\-]+)+[a-z_A-Z0-9\-]+
+MAILADR2  {BLANK}*[a-z_A-Z0-9+-]+({BLANK}*[Dd][Oo][Tt]{BLANK}*[a-z_A-Z0-9+-]+)?{BLANK}*[Aa][Tt]{BLANK}*[a-z_A-Z0-9-]+({BLANK}*[Dd][Oo][Tt]{BLANK}*[a-z_A-Z0-9\-]+)+{BLANK}*[a-z_A-Z0-9\-]+{BLANK}*
 OPTSTARS  ("//"{BLANK}*)?"*"*{BLANK}*
 LISTITEM  {BLANK}*[-]("#")?{WS}
 MLISTITEM {BLANK}*[+*]{WS}
@@ -655,6 +656,10 @@ REFWORD_NOCV   {FILEMASK}|{LABELID}|{REFWORD2_NOCV}|{REFWORD3}|{REFWORD4_NOCV}
 			 g_token->isEMailAddr=TRUE;
 			 return TK_URL;
                        }
+<St_Para>"<"{MAILADR2}">" { // anti spame mail address
+                         g_token->name=yytext;
+			 return TK_WORD;
+                       }
 <St_Para>"$"{ID}":"[^:\n$][^\n$]*"$" { /* RCS tag */
                          QCString tagName(yytext+1);
 			 int index=tagName.find(':');
-- 
cgit v0.12


From c68ae6ebd44ae66ba35154d868c686cf14c7f6b8 Mon Sep 17 00:00:00 2001
From: Dimitri van Heesch <doxygen@gmail.com>
Date: Wed, 18 Dec 2019 10:17:55 +0100
Subject: Made regexp more strict

---
 src/doctokenizer.l | 11 ++++++-----
 1 file changed, 6 insertions(+), 5 deletions(-)

diff --git a/src/doctokenizer.l b/src/doctokenizer.l
index bf9ccf6..a527672 100644
--- a/src/doctokenizer.l
+++ b/src/doctokenizer.l
@@ -344,8 +344,9 @@ PHPTYPE   [\\:a-z_A-Z0-9\x80-\xFF\-]+
 CITESCHAR [a-z_A-Z0-9\x80-\xFF\-\?]
 CITEECHAR [a-z_A-Z0-9\x80-\xFF\-\+:\/\?]
 CITEID    {CITESCHAR}{CITEECHAR}*("."{CITESCHAR}{CITEECHAR}*)*|"\""{CITESCHAR}{CITEECHAR}*("."{CITESCHAR}{CITEECHAR}*)*"\""
-MAILADR   ("mailto:")?[a-z_A-Z0-9.+-]+"@"[a-z_A-Z0-9-]+("."[a-z_A-Z0-9\-]+)+[a-z_A-Z0-9\-]+
-MAILADR2  {BLANK}*[a-z_A-Z0-9+-]+({BLANK}*[Dd][Oo][Tt]{BLANK}*[a-z_A-Z0-9+-]+)?{BLANK}*[Aa][Tt]{BLANK}*[a-z_A-Z0-9-]+({BLANK}*[Dd][Oo][Tt]{BLANK}*[a-z_A-Z0-9\-]+)+{BLANK}*[a-z_A-Z0-9\-]+{BLANK}*
+MAILADDR  ("mailto:")?[a-z_A-Z0-9.+-]+"@"[a-z_A-Z0-9-]+("."[a-z_A-Z0-9\-]+)+[a-z_A-Z0-9\-]+
+MAILWS    [\t a-z_A-Z0-9+-]
+MAILADDR2 {MAILWS}+{BLANK}+("at"|"AT"|"_at_"|"_AT_"){BLANK}+{MAILWS}+("dot"|"DOT"|"_dot_"|"_DOT_"){BLANK}+{MAILWS}+
 OPTSTARS  ("//"{BLANK}*)?"*"*{BLANK}*
 LISTITEM  {BLANK}*[-]("#")?{WS}
 MLISTITEM {BLANK}*[+*]{WS}
@@ -643,20 +644,20 @@ REFWORD_NOCV   {FILEMASK}|{LABELID}|{REFWORD2_NOCV}|{REFWORD3}|{REFWORD4_NOCV}
 			 g_token->isEMailAddr=FALSE;
 			 return TK_URL;
                        }
-<St_Para>{MAILADR}     { // Mail address
+<St_Para>{MAILADDR}    { // Mail address
                          g_token->name=yytext;
                          g_token->name.stripPrefix("mailto:");
 			 g_token->isEMailAddr=TRUE;
 			 return TK_URL;
                        }
-<St_Para>"<"{MAILADR}">" { // Mail address
+<St_Para>"<"{MAILADDR}">" { // Mail address
                          g_token->name=yytext;
                          g_token->name = g_token->name.mid(1,g_token->name.length()-2);
                          g_token->name.stripPrefix("mailto:");
 			 g_token->isEMailAddr=TRUE;
 			 return TK_URL;
                        }
-<St_Para>"<"{MAILADR2}">" { // anti spame mail address
+<St_Para>"<"{MAILADDR2}">" { // anti spam mail address
                          g_token->name=yytext;
 			 return TK_WORD;
                        }
-- 
cgit v0.12