/usr/share/doc/tcl-tclex/html/complement.html

<!DOCTYPE HTML PUBLIC "-//W3C//DTD HTML 4.0 Transitional//EN">
<HTML>
<HEAD>
   <META HTTP-EQUIV="Content-Type" CONTENT="text/html; charset=iso-8859-1">

   <TITLE>TcLex documentation - Complementary information</TITLE>

   <META NAME="Author"      CONTENT="Frédéric BONNET">
   <META NAME="Copyright"   CONTENT="Copyright (c) Frédéric BONNET, 1998">

   <LINK REL="StyleSheet" TYPE="text/css" HREF="tcLex.css">
</HEAD>
<BODY>

<H2>
Complementary information
</H2>

	<H3><A NAME="regexp81">	
	Using regexps under Tcl 8.1
	</A></H3>

<P>
<A HREF="http://www.scriptics.com">Scriptics</A> maintains a very good 
<A HREF="http://www.scriptics.com/support/howto/regexp81.html">regexp-HOWTO</A>
with many useful information on how to use regexps under Tcl 8.1.

	<H3><A NAME="flex">	
	Flex vs tcLex: a side-by-side comparison
	</A></H3>

<P>
<STRONG>TcLex</STRONG> borrowed many of its concepts to <STRONG>flex</STRONG> and tried to adapt them to
the Tcl philosophy. There are then areas where <STRONG>tcLex</STRONG> and <STRONG>flex</STRONG> are really
similar and interchangeable, while some features are enough different
to prevent an easy move from one to another.

		<H4><A NAME="syntax">
		Syntax
		</A></H4>

<P>
The most obvious difference is the syntax: <STRONG>flex</STRONG> uses its own whereas <STRONG>tcLex</STRONG>
uses Tcl's. <STRONG>TcLex</STRONG> tried to borrow as many existing Tcl constructs as
possible to make it easy to use by Tcl programmers while allowing easy
understanding by <STRONG>flex</STRONG> programmers. For example, <STRONG>tcLex</STRONG> borrowed its syntax
to the <STRONG>switch</STRONG> command and most of its procedural model to <STRONG>proc</STRONG>. <STRONG>Flex</STRONG> uses
specific syntax for rules specifiers and C syntax for actions (ie rules
scripts). In some areas, <STRONG>tcLex</STRONG> has higher level features than <STRONG>flex</STRONG>.

<P>
<STRONG>TcLex</STRONG> uses list syntax for rules specifiers and thus requires
that a fixed number of elements (ie 4 elements) be specified. <STRONG>Flex</STRONG> is more
permissive because some specifiers can be omitted. For example, the conditions
list defaults to empty in <STRONG>flex</STRONG> whereas it must be specified as an empty
list in <STRONG>tcLex</STRONG>. Moreover, <STRONG>tcLex</STRONG> also needs a list of match variables used to
report the matched parts of the input string, the same way <STRONG>regexp</STRONG> does.
Here is a short example:

<UL>
<LI><STRONG>flex</STRONG> syntax:

<BLOCKQUOTE><PRE>
&lt;cond1&gt;"\n" {
  /* matches newlines within condition cond1 */
}
"\n" {
  /* matches newlines with empty conditions specifier */
} 
</PRE></BLOCKQUOTE>

<LI><STRONG>tcLex</STRONG> syntax:

<BLOCKQUOTE><PRE>
{cond1} "\n" {} {
  # matches newlines within condition cond1
}
{} "\n" {} {
  /* matches newlines with empty conditions specifier */
} 
</PRE></BLOCKQUOTE>

</UL>

<P>
Note the empty condition list in the second <STRONG>tcLex</STRONG> rule, as well as the (here
empty) match variables list. Apart from that, <STRONG>flex</STRONG> and <STRONG>tcLex</STRONG> have very similar
syntaxes.

<P>
The second visible difference between <STRONG>flex</STRONG> and <STRONG>tcLex</STRONG> is the latter requiring
an extra list of match variables. This is a convenient way to get the matched
parts of the input string, and is inspired by the <STRONG>regexp</STRONG> command syntax.
Getting the matched string with <STRONG>flex</STRONG> requires using the <STRONG>yytext</STRONG> variable,
and there is no subexpressions reporting (see <A HREF="#feedback">Feedback</A> below).

<P>
The third syntactical difference is <STRONG>tcLex</STRONG>'s lack of name definitions. These
are macro definitions that allows for simpler definition of <STRONG>flex</STRONG>'s regular
expressions. There is no name definition in <STRONG>tcLex</STRONG> because they can be
handled by Tcl itself using variable substitutions. For example:

<BLOCKQUOTE><PRE>
DIGIT    [0-9]
ID       [a-z][a-z0-9]*

%%

{DIGIT}+"."{DIGIT}*        {
	printf( "A float: %s (%g)\n", yytext,
			atof( yytext ) );
}

{DIGIT}+    {
	printf( "An integer: %s (%d)\n", yytext,
			atoi( yytext ) );
}
</PRE></BLOCKQUOTE>

<P>
Note the two name definitions <EM>DIGIT</EM> and <EM>ID</EM>. They are subsequently used in
the regular expressions to simplify their writing and also make them more
readable. In <STRONG>tcLex</STRONG> variables must be used:

<BLOCKQUOTE><PRE>
set DIGIT    {[0-9]}
set ID       {[a-z][a-z0-9]*}

lexer lx \
{} ${DIGIT}+.${DIGIT}* {text} {
	puts "A float: $text ([expr double($text)])"
} \
{} ${DIGIT}+ {text} {
	puts "An integer: $text ([expr int($text)])"
}
</PRE></BLOCKQUOTE>

<P>
Note the use of variable substitutions in regexps. To make variable
substitutions work, you also have to use the "rules-as-arguments"
lexer style (see <A HREF="design.html#style">Lexing With Style</A>): note the backslashes at every
end of line. The other noticeable difference is the use of a variable
named <EM>text</EM> to report the matched string. <STRONG>Flex</STRONG> uses <STRONG>yytext</STRONG> for that
purpose. Be careful to respect Tcl's syntax when defining variables:
here we have to enclose the name definitions between braces to
avoid bracket substitution. The regexps are the same except for the
added <TT>'$'</TT> for variable substition and the removal of inside double quotes.
Both lexers should work the same.

<P>
Finally, there are subtle differences in regular expressions syntax
(<STRONG>tcLex</STRONG> uses Tcl's syntax). First is the absence of trailing contexts
in <STRONG>tcLex</STRONG> (syntax: <EM>r</EM><TT>/</TT><EM>s</EM>, where <EM>r</EM> is the regular expression and <EM>s</EM> is the
trailing context). This shouldn't be a big problem though, unless
when converting from <STRONG>flex</STRONG> to <STRONG>tcLex</STRONG>. Second is line-sensitivity. <STRONG>Flex</STRONG>
uses line-sensitive regexps whereas Tcl uses line-insensitive by
default. Since there is no support for line-sensitivity in Tcl8.0
and Tcl8.1 requires specific syntax, <STRONG>tcLex</STRONG> provides its own portable
line-sensitivity through the use of the <A HREF="using.html#-lines" CLASS="keyword"><STRONG>-lines</STRONG></A> switch. Line sensitivity
changes the meaning of the special characters <TT>'^'</TT>, <TT>'$'</TT> and <TT>'.'</TT> within regexps.
With line-insensitive regexps, <TT>'^'</TT> and <TT>'$'</TT> respectively match the beginning
and end of string, and <TT>'.'</TT> any character. With line-sensitive regexps,
<TT>'^'</TT> and <TT>'$'</TT> respectively match the beginning and end of lines, and <TT>'.'</TT> any
but the newline character.


		<H4><A NAME="feedback">
		Feedback
		</A></H4>

<P>
To report the matched string, <STRONG>flex</STRONG> uses the special string variable
<STRONG>yytext</STRONG>. <STRONG>TcLex</STRONG> uses a quite different method that is closer to Tcl
philosophy (ie the <STRONG>regexp</STRONG> command) and also more powerful in term
of feedback. With <STRONG>flex</STRONG>, parentheses are only used to override
precedence in regexps. However Tcl regexps also use parentheses
for reporting subexpressions. To keep this feature, <STRONG>tcLex</STRONG> needs
variables to report matched substrings, the same way <STRONG>regexp</STRONG> does.
Thus there is no predefined <STRONG>yytext</STRONG> variable but per-rule user-defined
variables. This can allow for elegant constructs when several
rules share the same action. Let's consider the following example,
where several syntactically different but conceptually similar constructs
have to be matched:

<BLOCKQUOTE><PRE>
lexer lexCComments {
  {} "(/\\*)([^*]*)\\*/"        {text style comment} -
  -  "(//)(([^\\\\\n]|\\\\.)*)" {text style comment} {
    # C/C++ comments
    switch -- $style {
      /* {set lang C}
      // {set lang C++}
    }
    puts "$lang comment matched: $comment"
  }
}
</PRE></BLOCKQUOTE>

<P>
The two rules are used to match C and C++ comments. C comments
are any characters enclosed between <TT>"/*"</TT> and <TT>"*/"</TT>. C++ comments are any
characters after <TT>"//"</TT> until the end of a line not ending with <TT>'\'</TT> (note
the quoting hell required for <TT>'\'</TT> to work). Here the same action is used
for two different rules, but using the same variables for reporting. This
greatly simplifies rules writing as it allows using the same action
for several similar yet different rules. Note that the order of the variables may
differ.


		<H4><A NAME="behavior">
		Behavior
		</A></H4>

<P>
There are some behavioral differences between <STRONG>flex</STRONG> and <STRONG>tcLex</STRONG>, but most
can be overriden by using specific <STRONG>tcLex</STRONG> switches.

<P>
As previously said, <STRONG>flex</STRONG> is line-sensitive whereas <STRONG>tcLex</STRONG> is line-insensitive
by default (as Tcl). To enable line-sensitivity, use the <A HREF="using.html#-lines" CLASS="keyword"><STRONG>-lines</STRONG></A> flag.

<P>
Second, <STRONG>flex</STRONG> uses a longest prefered match scheme whereas <STRONG>tcLex</STRONG> uses a
first match scheme by default. Longest prefered match means that every
rule will be tried and the longest will be used. First match means that
the first matching rule will be used. This can lead to strange behavior
when converting from <STRONG>flex</STRONG> to <STRONG>tcLex</STRONG>. For example, this simple lexer taken
from the <STRONG>flex</STRONG> man page:

<BLOCKQUOTE><PRE>
DIGIT    [0-9]
ID       [a-z][a-z0-9]*

%%

{DIGIT}+    {
	printf( "An integer: %s (%d)\n", yytext,
			atoi( yytext ) );
}

{DIGIT}+"."{DIGIT}*        {
	printf( "A float: %s (%g)\n", yytext,
			atof( yytext ) );
}
</PRE></BLOCKQUOTE>

<P>
Converting straightforwardly this lexer to <STRONG>tcLex</STRONG> will give incorrect
results:

<BLOCKQUOTE><PRE>
set DIGIT    {[0-9]}
set ID       {[a-z][a-z0-9]*}

lexer lx \
{} ${DIGIT}+           {text} {
	puts "An integer: $text ([expr int($text)])"
} \
{} ${DIGIT}+.${DIGIT}* {text} {
	puts "A float: $text ([expr double($text)])"
}
</PRE></BLOCKQUOTE>

<P>
Since <STRONG>tcLex</STRONG> uses first-match by default, the string <TT>"1.5"</TT> which
would be seen as a float by <STRONG>flex</STRONG> (2nd rule) will be seen as an integer
by <STRONG>tcLex</STRONG> because first rule is matched before the second. To avoid this
behavior, either use the <A HREF="using.html#-longest" CLASS="keyword"><STRONG>-longest</STRONG></A> flag or change the order of the
rules. Using longest-match can hit performances when there is a large
number of rules, but changing the order of the rules must also be
done carefully. What to do will depend on the situation: simple lexers
(like the one above)
will behave better by changing order whereas more complex ones will
be preferabily converted with <A HREF="using.html#-longest" CLASS="keyword"><STRONG>-longest</STRONG></A> flag to avoid obscure bugs.
Note that first-match scheme can also be useful to simplify rules
definitions because one can take their precedence order into account,
and can greatly speedup the processing. First-match
scheme was chosed for <STRONG>tcLex</STRONG> because it is the way <STRONG>switch</STRONG> works.

<P>
The last behavioral difference lies in actions. With <STRONG>flex</STRONG>, actions
often end with a <STRONG>return</STRONG> statement that returns a specific value to
the calling context (typically a token returned to a <STRONG>yacc</STRONG> parser).
<STRONG>TcLex</STRONG> lexers have a completely different behavior. Since lexers are
seen as a mixture of Tcl <STRONG>switch</STRONG> and <STRONG>proc</STRONG>, a return statement has the
same consequence as with these commands: stopping the processing.
Thus it isn't possible yet to use <STRONG>tcLex</STRONG> lexers as tokenizers that return
a value every time a rule succeeds. Future versions will certainly
improve on this point and turn lexers into tokenizers. For now, the only
way lexers can return a value during processing is through the
incremental processing scheme (see <A HREF="using.html#incremental">Incremental processing</A>). This choice
was made to be consistent with <STRONG>proc</STRONG> and <STRONG>switch</STRONG>.


		<H4><A NAME="other">
		Other features
		</A></H4>

<P>
<STRONG>Flex</STRONG> provides many functions to access and/or modify the input buffer, such
as C functions <STRONG>yymore()</STRONG>, <STRONG>yyless()</STRONG>, <STRONG>input()</STRONG> and <STRONG>unput()</STRONG>. <STRONG>TcLex</STRONG> provides
similar volontarily limited features in the form of <A HREF="using.html#input" CLASS="keyword"><STRONG>input</STRONG></A> and <A HREF="using.html#unput" CLASS="keyword"><STRONG>unput</STRONG></A> lexer
subcommands. The reason for these limitations is that these features can
be kludgy or can modify the input. 

<UL>
<LI><STRONG>yymore()</STRONG> is used to append some characters to the next matched text. It
    can lead to very strange results in some cases and is very kludgy by
    nature. IMHO the fact that the <STRONG>flex</STRONG> man page uses the <TT>"mega-kludge"</TT>
    string in an example of <STRONG>yymore()</STRONG> use is not a coincidence :-)
<LI><STRONG>input()</STRONG> is available unchanged in the form of the <A HREF="using.html#input" CLASS="keyword"><STRONG>input</STRONG></A> subcommand.
<LI><STRONG>yyless()</STRONG> is available in the form of the <A HREF="using.html#unput" CLASS="keyword"><STRONG>unput</STRONG></A> subcommand. <STRONG>Unput()</STRONG> is
    also available through this subcommand provided that it cannot modify
    the input buffer, ie unput an arbitrary character. I consider this to
    be a kludge like <STRONG>yymore()</STRONG> that somewhat violates Tcl philosophy as a
    high level language. If someone convinces me of the contrary, I'll
    consider adding such features in future versions ;-) Anyway I don't think
    they are fundamental but rather denote quick'n dirty design.
</UL>


		<H4><A NAME="chart">
		Features chart
		</A></H4>

<TABLE BORDER=2>
<TR CLASS="odd">
<TH></TH>					<TH>flex</TH>					<TH>tcLex</TH>
</TR>
<TR CLASS="even">
<TH>use</TH>				<TD>generates static C source files</TD>	<TD>dynamic creation of Tcl commands</TD>
</TR>
<TR CLASS="odd">
<TH>syntax</TH>				<TD>specific</TD>					<TD>Tcl</TD>
</TR>
<TR CLASS="even">
<TH>name definitions</TH>		<TD>yes, specific syntax</TD>			<TD>via Tcl substitutions</TD>
</TR>
<TR CLASS="odd">
<TH>match scheme</TH>			<TD>longest-prefered match only</TD>	<TD>both longest-prefered and first match</TD>
</TR>
<TR CLASS="even">
<TH>line-sensitivity</TH>		<TD>line-sensitive only</TD>			<TD>both line-sensitive and -insensitive</TD>
</TR>
<TR CLASS="odd">
<TH>regular expressions</TH>		<TD>alike Tcl8.0 plus trailing contexts and characters classes</TD>	<TD>depend on Tcl version</TD>
</TR>
<TR CLASS="even">
<TH>reentrancy</TH>			<TD>limited</TD>					<TD>yes</TD>
</TR>
<TR CLASS="odd">
<TH>subexpressions reporting</TH>	<TD>no</TD>						<TD>yes</TD>
</TR>
<TR CLASS="even">
<TH>tokenization</TH>			<TD>yes</TD>					<TD>no</TD>
</TR>
<TR CLASS="odd">
<TH>incremental processing</TH>	<TD>must be specially designed</TD>		<TD>transparent</TD>
</TR>
<TR CLASS="even">
<TH>multiple input buffers</TH>	<TD>yes</TD>					<TD>no</TD>
</TR>
</TABLE>



		<H4><A NAME="converting">
		Converting <STRONG>flex</STRONG> to <STRONG>tcLex</STRONG>: Rules of Thumb
		</A></H4>

<P>
If you plan to use <STRONG>tcLex</STRONG> as a Tcl replacement for <STRONG>flex</STRONG>, it is highly
probable that you will have to convert a <STRONG>flex</STRONG> lexer to <STRONG>tcLex</STRONG> one day.
Here are several rules of thumb to save time and avoid strange
errors:

<UL>
<LI>name definitions in <STRONG>flex</STRONG> are automatically enclosed between
    parentheses before they are expanded, unless they begin with
    <TT>'^'</TT> or end with <TT>'$'</TT>, to avoid precedence problems in regexps. You
    have to be careful whether you need to add parentheses or not
    when converting to <STRONG>tcLex</STRONG>.
<LI>parentheses are only used by <STRONG>flex</STRONG> for precedence in regexps.
    In <STRONG>tcLex</STRONG> they are also used for reporting. This causes no
    special problem when directly converting from flex to <STRONG>tcLex</STRONG>,
    however be careful about parentheses if you later add
    subexpressions reporting.
<LI>be careful about Tcl substitutions and escape rules when converting regexps, else
    the evil Tcl parser will bite'ya ;-) Take care about backslash characters and braces
    (see the <A HREF="#FAQs">FAQs</A> below).
<LI><STRONG>flex</STRONG> uses longest-prefered-match scheme, <STRONG>tcLex</STRONG> uses first-match
    scheme by default. When converting, either use the <A HREF="using.html#-longest" CLASS="keyword"><STRONG>-longest</STRONG></A>
    flag (safest but less performant), or change the order of the
    rules (only if you know what you're doing!)
<LI><STRONG>flex</STRONG> uses line-sensitive regexp, <STRONG>tcLex</STRONG> is line-insensitive by
    default. Just use the <A HREF="using.html#-lines" CLASS="keyword"><STRONG>-lines</STRONG></A> flag if you see that the lexer
    uses special line-sensitive constructs (ie special characters <TT>^$.</TT>)
<LI>do not use return statements in actions: they cause the processing
    to stop. If there are return statements in your <STRONG>flex</STRONG> lexer, it
    probably means that the lexer is a tokenizer (used in conjunction
    with <STRONG>yacc</STRONG>). <STRONG>TcLex</STRONG> currently doesn't support tokenizer-style
    processing so this may need a bit work to convert.
<LI><STRONG>input()</STRONG> in <STRONG>flex</STRONG> consumes the next character. The <A HREF="using.html#input" CLASS="keyword"><STRONG>input</STRONG></A> subcommand
    in <STRONG>tcLex</STRONG> returns one character by default but can take more if
    requested. A single <STRONG>tcLex</STRONG> <A HREF="using.html#input" CLASS="keyword"><STRONG>input</STRONG></A> can then be used to replace several
    subsequent <STRONG>input()</STRONG> calls.
<LI>replace <STRONG>yyless(<EM>n</EM>)</STRONG> by <A HREF="using.html#unput" CLASS="keyword"><STRONG>unput</STRONG></A> <STRONG><EM>$n</EM></STRONG>;
it rewinds the input by <EM>n</EM> chars. Note that, contrary to <STRONG>flex</STRONG>, you cannot rewind past the beginning of the matched string.
<LI><STRONG>unput()</STRONG> puts back a single character in the input stream. If this
    character is the same as the one that was extracted (by <STRONG>input()</STRONG>), then
    use <A HREF="using.html#unput" CLASS="keyword"><STRONG>unput</STRONG></A> which by default rewinds the input by one char. Else,
    you're out of luck  :-(
<LI>there is no equivalent to <STRONG>yymore()</STRONG>. You can somewhat emulate its
    behavior by storing the <STRONG>yymore</STRONG>'d string in a variable (empty by
    default) and prepend it to the matched string in every action.
    Do it only if you can't do else.
<LI>by default, <STRONG>flex</STRONG> sends matched characters to the output. <STRONG>TcLex</STRONG>
    has no idea what the output can be (a string, a list, a tree...)
    so by default it does nothing. You can replace outputs (eg.
    <STRONG>ECHO</STRONG> statements) with <STRONG>puts</STRONG> calls, and add a default rule that
    outputs any unmatched character:

<BLOCKQUOTE><PRE>
* .|\n {c} {puts -nonewline $c}
</PRE></BLOCKQUOTE>
</UL>


	<H3><A NAME="tips">
	Tips & Tricks
	</A></H3>

[to be written]


	<H3><A NAME="howto">
	How to...
	</A></H3>

[to be written]


	<H3><A NAME="FAQs">
	Frequently Asked Questions (FAQs) & Frequently Made Mistakes (FMMs)
	</A></H3>

[to be written]


	<H3><A NAME="resources">
	Resources
	</A></H3>

<P>
Home page:
<BLOCKQUOTE>
	<A HREF="http://www.multimania.com/fbonnet/Tcl/tcLex/index.en.htm">http://www.multimania.com/fbonnet/Tcl/tcLex/index.en.htm (english)</A><BR>
	<A HREF="http://www.multimania.com/fbonnet/Tcl/tcLex/index.htm">http://www.multimania.com/fbonnet/Tcl/tcLex/index.htm (french)</A>
</BLOCKQUOTE>

<P>
Mailing list:
<BLOCKQUOTE>
	Home page: <A HREF="http://www.eGroups.com/list/tclex">http://www.eGroups.com/list/tclex</A><BR>
	To subscribe: <A HREF="mailto:tclex-subscribe@egroups.com">tclex-subscribe@egroups.com</A>
</BLOCKQUOTE>

<HR>
<ADDRESS>
Send questions, comments, requests, etc. to the author: <A HREF="mailto:frederic.bonnet@ciril.fr">Frédéric BONNET &lt;frederic.bonnet@ciril.fr&gt;</A>.
</ADDRESS>

</BODY>
</HTML>
tcl-tclex 1.2a1-16 / usr / share / doc / tcl-tclex / html / complement.html
