# 
# patch "format.cc"
#  from [ce1a992b9a4cadbc1d391e705cacfbcb58dc90ef]
#    to [e231ce6389ace3b897d872bab11e025198ea56d0]
# 
# patch "monotone.1"
#  from [2f053fc3c1b544dc2cfd523d443a4968966b721c]
#    to [d54c9cd556f349dca1e5788897128023529520b0]
# 
# patch "monotone.texi"
#  from [1cd3647b4cc2c03fc45c6abfe8072a6ada288f8e]
#    to [725e706a5c15ce2db80975cbdc4fc9887ddc8149]
# 
--- format.cc
+++ format.cc
@@ -289,9 +289,15 @@
     case '\\':
       out << '\\';
       break;
-    case '@':
-      out << '@';
+    case '%':
+      out << '%';
       break;
+    case '{':
+      out << '{';
+      break;
+    case '}':
+      out << '}';
+      break;
     case 'n':
       out << std::endl;
       break;
--- monotone.1
+++ monotone.1
@@ -295,11 +295,14 @@
 The format string is made by a common part, applied to each revision,
 and one or more optional changeset subformats, applied to changeset
 data.
-Special changeset placeholders put in the common expression are
-substituted by the formatted changeset informations.
-The common formatting expression is separated from the changeset ones by apposite markers. 
+Changeset subformats are enclosed beetween braces and start with an
+uppercase formatting specifier, e.g. an expression using the 'added
+files' subformat is %A{...}
+Everything \fIinside\fP the braces is repeated for every matching
+changeset item.  For example, %A{add: %f\\n} prints every added file
+prefixed by 'add: ' and followed by a newline.
  
-Common formatting specifiers:
+Base formatting specifiers:
 
 %a  : value of author certificate
 .br
@@ -331,6 +334,8 @@
 .br
 %+  : terminates formatting for the current revision. The next revision
 will be formatted by the specifiers following %+
+.br
+%%  : the % character. It could also be obtained with '\\%' (see below)
 
 
 The short forms have the following meanings:  For author, everything 
@@ -338,30 +343,30 @@
 character.  For date, everything before the first 'T' character.  
 For ids, the first 8 characters.  
 
-Changeset Placeholders:
+Changeset formatting specifiers:
 
-%P : ancestors (marker @P) 
+%P : ancestors
 .br
-%A : added files (marker @A)
+%A : added files
 .br
-%D : deleted files (marker @D)
+%D : deleted files
 .br
-%E : deleted dirs (marker @E)
+%E : deleted dirs
 .br
-%R : renamed files (marker @R)
+%R : renamed files
 .br
-%C : renamed dirs (marker @C)
+%C : renamed dirs
 .br
-%M : modified files (marker @M)
+%M : modified files
 
-The %P,%A,%D,%E,%M changeset placeholders supports the private
-specifier %f, substituted with the file/directory/ancestor name/id
-The %R and %M placeholders supports the specifiers %f, current
-file/directory name and %o, old file/directory name.
+%P,%A,%D,%E,%M supports the private specifier %f, substituted with the
+file/directory/ancestor name/id. 
+%R and %M supports the specifiers %f, current file/directory name and
+%o, old file/directory name. 
  
 Plus, all format string support the  \\n, \\r, \\a, \\t, \\b, \\f, \\v
 modifiers with the usual c-like meaning.
-\\\\, \\% and \\@ are used to obtain the \\, % and @ char respectively.
+\\\\, \\%, \\} are used to obtain the \\, % and } chars respectively.
 
 The default format string is '%i\\n'. 
 
--- monotone.texi
+++ monotone.texi
@@ -4856,15 +4856,17 @@
 The format string is made by a common part, applied to each revision,
 and one or more optional changeset subformats, applied to changeset
 data.
-Special changeset placeholders put in the common expression are
-substituted by the formatted changeset informations.
-The separate the common formatting expression is separated from  the
-changeset ones by apposite markers. 
+Changeset subformats are enclosed beetween braces and start with an
+uppercase formatting specifier, e.g. an expression using the 'added
+files' subformat is address@hidden@} .
+Everything \fIinside\fP the braces is repeated for every matching
+changeset item.  For example, address@hidden: address@hidden prints every added file
+prefixed by 'add: ' and followed by a newline.
 Some formatting specifier have also a short form, obtained prepending
 an 's' to the specifier code. As an example, the short form of the %d
 specifier is %sd.
 
-Common formatting specifiers:
+Base formatting specifiers:
 @verbatim
 %[s]a : value of author certificate
 %[s]b : value of branch certificate
@@ -4877,17 +4879,18 @@
 %[s]m : manifest id
 %+ : terminates formatting for the current revision. The next revision
 will be formatted by the specifiers following %+
+%% : the '%' character. Could also be obtained by \% (see below)
 @end verbatim
 
-Changeset Placeholders:
+Changeset formatting specifiers:
 @verbatim
-%P : ancestors (marker @P) 
-%A : added files (marker @A)
-%D : deleted files (marker @D)
-%E : deleted dirs (marker @E)
-%R : renamed files (marker @R)
-%C : renamed dirs (marker @C)
-%M : modified files (marker @M)
+%P : ancestors
+%A : added files 
+%D : deleted files 
+%E : deleted dirs 
+%R : renamed files 
+%C : renamed dirs 
+%M : modified files 
 @end verbatim
 
 The short forms have the following meanings: for ids, the first 8
@@ -4895,16 +4898,16 @@
 branch, everything after the last '.' character.  For date, everything
 before the first 'T' character.
 
-The %P,%A,%D,%E,%M changeset placeholders supports the private
-specifier %f, substituted with the file/directory/ancestor name/id
-The %R and %M placeholders supports the specifiers %f, current
-file/directory name and %o, old file/directory name.
- 
-Plus, all format string support the  \\n, \\r, \\a, \\t, \\b, \\f, \\v
+%P,%A,%D,%E,%M supports the private specifier %f, substituted with the
+file/directory/ancestor name/id. 
+%R and %M supports the specifiers %f, current file/directory name and
+%o, old file/directory name. 
+
+Plus, all format string support the \n, \r, \a, \t, \b, \f, \v
 modifiers with the usual c-like meaning.
-\\\\, \\% and \\@ are used to obtain the \\, % and @ char respectively.
+\\, \% and address@hidden are used to obtain the \, % and @} char respectively.
 
-The default format string is '%i\\n'. 
+The default format string is '%i\n'. 
 
 The specifier %+ is used to apply alternating format strings to
 different revisions. When found, a %+ specifiers terminates the
@@ -6630,6 +6633,7 @@
 %sm : shortened manifest id
 %+  : terminates formatting for the current revision. The next revision
 will be formatted by the specifiers following %+
+%%  : the % character. It could also be obtained with '\%' (see below)
 @end verbatim
 
 The short forms have the following meanings: for ids, the first 8
@@ -6637,27 +6641,27 @@
 branch, everything after the last '.' character.  For date, everything
 before the first 'T' character.
 
-Changeset Placeholders:
+Changeset formatting specifiers:
 @verbatim
-%P : ancestors (marker @P) 
-%A : added files (marker @A)
-%D : deleted files (marker @D)
-%E : deleted dirs (marker @E)
-%R : renamed files (marker @R)
-%C : renamed dirs (marker @C)
-%M : modified files (marker @M)
+%P : ancestors
+%A : added files 
+%D : deleted files 
+%E : deleted dirs 
+%R : renamed files 
+%C : renamed dirs 
+%M : modified files 
 @end verbatim
 
-The %P,%A,%D,%E,%M changeset placeholders supports the private
-specifier %f, substituted with the file/directory/ancestor name/id
-The %R and %M placeholders supports the specifiers %f, current
-file/directory name and %o, old file/directory name.
- 
-Plus, all format string support the  \\n, \\r, \\a, \\t, \\b, \\f, \\v
+%P,%A,%D,%E,%M supports the private specifier %f, substituted with the
+file/directory/ancestor name/id. 
+%R and %M supports the specifiers %f, current file/directory name and
+%o, old file/directory name. 
+
+Plus, all format string support the  \n, \r, \a, \t, \b, \f, \v
 modifiers with the usual c-like meaning.
-\\\\, \\% and \\@ are used to obtain the \\, % and @ char respectively.
+\\, \% and address@hidden are used to obtain the \, % and @} chars respectively.
 
-The default format string is '%i\\n'. 
+The default format string is '%i\n'. 
 
 @item @b{--xml}
 Generate an xml document based on automate output and containing full