#!/usr/bin/perl ############################################################################### # # # ca65html # # # # Convert a ca65 source into HTML # # # # # # # # (C) 2000-2007 Ullrich von Bassewitz # # Roemerstrasse 52 # # D-70794 Filderstadt # # EMail: uz@cc65.org # # # # # # This software is provided 'as-is', without any expressed or implied # # warranty. In no event will the authors be held liable for any damages # # arising from the use of this software. # # # # Permission is granted to anyone to use this software for any purpose, # # including commercial applications, and to alter it and redistribute it # # freely, subject to the following restrictions: # # # # 1. The origin of this software must not be misrepresented; you must not # # claim that you wrote the original software. If you use this software # # in a product, an acknowledgment in the product documentation would be # # appreciated but is not required. # # 2. Altered source versions must be plainly marked as such, and must not # # be misrepresented as being the original software. # # 3. This notice may not be removed or altered from any source # # distribution. # # # ############################################################################### # Things currently missing: # # - Scoping with .proc/.endproc, .scope/.endscope, .enum/.endenum, # .struct/.endstruct, .union/endunion, .repeat/.endrep, .local # - .global is ignored # - .case is ignored, labels are always case-sensitive # - .include handling (difficult) # - The global namespace operator :: # use strict 'vars'; use warnings; # Modules use Getopt::Long; #-----------------------------------------------------------------------------# # Variables # # ----------------------------------------------------------------------------# # Global variables my %Files = (); # List of all files. my $FileCount = 0; # Number of input files my %Exports = (); # List of exported symbols. my %Imports = (); # List of imported symbols. my %Labels = (); # List of all labels my $LabelNum = 0; # Counter to generate unique labels # Command line options my $BGColor = "#FFFFFF"; # Background color my $Colorize = 0; # Colorize the output my $CommentColor = "#B22222"; # Color for comments my $CRefs = 0; # Add references to the C file my $CtrlColor = "#228B22"; # Color for control directives my $CvtTabs = 0; # Convert tabs to spaces my $TabSize = 8; # This is how god created them my $Debug = 0; # No debugging my $Help = 0; # Help flag my $HTMLDir = ""; # Directory in which to create the files my $IndexCols = 6; # Columns in the file listing my $IndexTitle = "Index"; # Title of index page my $IndexName = "index.html"; # Name of index page my $IndexPage = 0; # Create an index page my $KeywordColor = "#A020F0"; # Color for keywords my $LineLabels = 0; # Add a HTML label to each line my $LineNumbers = 0; # Add line numbers to the output my $LinkStyle = 0; # Default link style my $ReplaceExt = 0; # Replace extension instead of appending my $StringColor = "#6169C1"; # Color for strings my $TextColor = "#000000"; # Text color my $Verbose = 0; # Be quiet # Table used to convert the label number into names my @NameTab = ('A' .. 'Z', '0' .. '9'); #-----------------------------------------------------------------------------# # Helper functions # # ----------------------------------------------------------------------------# # Terminate with an error sub Abort { print STDERR "ca65html: @_\n"; exit 1; } # Print a message if verbose is true sub Gabble { if ($Verbose) { print "ca65html: @_\n"; } } # Generate a label and return it sub GenLabel { my $I; my $L = "";; my $Num = $LabelNum++; # Generate the label for ($I = 0; $I < 4; $I++) { $L = $NameTab[$Num % 36] . $L; $Num /= 36; } return $L; } # Make an output file name from an input file name sub GetOutName { # Input name is parameter my $InName = $_[0]; # Create the output file name from the input file name if ($ReplaceExt && $InName =~ /^(.+)\.([^\.\/]*)$/) { return "$1.html"; } else { return "$InName.html"; } } # Translate some HTML characters into harmless names. sub Cleanup { my $S = shift (@_); $S =~ s/&/&/g; $S =~ s/</g; $S =~ s/>/>/g; $S =~ s/\"/"/g; return $S; } # Strip a path from a filename and return just the name sub StripPath { # Filename is argument my $FileName = $_[0]; # Remove a path name if we have one $FileName =~ /^(.*?)([^\/]*)$/; return $2; } #-----------------------------------------------------------------------------# # Document header and footer # # ----------------------------------------------------------------------------# # Print the document header sub DocHeader { my $OUT = shift (@_); my $Asm = shift (@_); print $OUT "\n"; print $OUT <<"EOF";
\n";
# Keep the user happy
Gabble ("$FileName => $OutName");
# The instructions that will have hyperlinks if a label is used.
# And, they will be highlighted when color is used.
my $LabelIns = "adc|add|and|asl|bb[rs][0-7]|b[cv][cs]|beq|bge|bit|blt|".
"bmi|bne|bpl|br[akl]|bsr|cmp|cop|cp[axy]|dec|eor|inc|jml|".
"jmp|jsl|jsr|ld[axy]|lsr|mvn|mvp|ora|pe[air]|rep|".
"[rs]mb[0-7]|rol|ror|sbc|sep|st[012axyz]|sub|tai|tam|tdd|".
"ti[ain]|tma|trb|tsb|tst";
# Instructions that have only the implied-addressing mode -- therefore,
# no hyperlinking. They will be highlighted only, when color is used.
my $OtherIns = "cl[acdivxy]|csh|csl|de[axy]|in[axy]|nop|ph[abdkpxy]|".
"pl[abdpxy]|rt[ils]|sax|say|se[cdit]|stp|swa|sxy|ta[dsxy]|".
"tam[0-7]|tcd|tcs|tda|tdc|tma[0-7]|ts[acx]|tx[asy]|tya|tyx|".
"wai|xba|xce";
# Read the input file, replacing references with hyperlinks; and, mark
# labels as link targets.
my $LineNo = 0;
LINE: while ($Line = ) {
# Count input lines
$LineNo++;
# Remove the newline at the end of line. Don't use chomp to be able to
# read dos/windows sources on unices.
$Line =~ s/[\r\n]*$//;
# If requested, convert tabs to spaces
if ($CvtTabs) {
# Don't ask me - this is from the perl manual page
1 while ($Line =~ s/\t+/' ' x (length($&) * $TabSize - length($`) % $TabSize)/e) ;
}
# Clear the output line
$OutLine = "";
# If requested, add a html label to each line with a name "linexxx",
# so it can be referenced from the outside (this is the same convention
# that is used by c2html). If we have line numbers enabled, add them.
if ($LineLabels && $LineNumbers) {
$OutLine .= sprintf ("%6d: ", $LineNo, $LineNo);
} elsif ($LineLabels) {
$OutLine .= sprintf ("", $LineNo);
} elsif ($LineNumbers) {
$OutLine .= sprintf ("%6d: ", $LineNo);
}
# Cut off a comment from the input line. Beware: We have to check for
# strings, since these may contain a semicolon that is no comment
# start.
($Line, $Comment) = $Line =~ /^((?:[^"';]+|".*?"|'.*?')*)(.*)$/;
if ($Comment =~ /^["']/) {
# Line with invalid syntax - there's a string start but
# no string end.
Abort (sprintf ("Invalid input at %s(%d)", $FileName, $LineNo));
}
# Remove trailing whitespace and move it together with the comment
# into the $Trailer variable.
$Line =~ s/\s*$//;
$Trailer = $& . ColorizeComment (Cleanup ($Comment));
# Check for a label at the start of the line. If we have one, process
# it, and remove it from the line.
if ($Line =~ s/^\s*?(([\@?]?)[_a-zA-Z]\w*)(\s*(?::=?|=))//) {
# Is this a local label?
if ($2 ne "") {
# Use the prefix
$Id = "$CheapPrefix$1";
} else {
# Use as is
$Id = $1;
# Remember the id as new cheap local prefix
$CheapPrefix = $Id;
}
# Get the label for the id
$Label = $Labels{$OutName}{$Id};
# Print the label with a tag
$OutLine .= "$1$3";
# Is the name explicitly assigned a value?
if ($3 =~ /=$/) {
# Print all identifiers if there are any.
while ($Line =~ s/^([^_a-zA-Z]*?)(([\@?]?)[_a-zA-Z]\w*)//) {
# Add the non-label stuff.
$OutLine .= Cleanup ($1);
# Use the prefix if the label is local.
# Get the reference to that label if we find it.
$OutLine .= RefLabel ($OutName, ($3 ne "") ? "$CheapPrefix$2" : $2, $2);
}
# Add a remainder if there is one.
$OutLine .= Cleanup ($Line);
# The line is complete; print it.
next LINE;
}
}
# Print any leading whitespace and remove it, so we don't have to
# care about whitespace below.
if ($Line =~ s/^\s+//) {
$OutLine .= $&;
}
# Handle the import statements
if ($Line =~ s/^\.(?:(?:force)?import|importzp)\s+//i) {
# Print any fixed stuff from the line and remove it
$OutLine .= $&;
# Print all identifiers if there are any
while ($Line =~ s/^[_a-zA-Z]\w*//) {
# Remember the identifier
my $Id = $&;
# Variable to assemble HTML representation
my $Contents = "";
# Make this import a link target
if (exists ($Imports{$OutName}{$Id})) {
$Label = $Imports{$OutName}{$Id};
$Contents .= sprintf (" name=\"%s\"", $Label);
}
# If we have an export for this import, add a link to this
# export definition
if (exists ($Exports{$Id})) {
$Label = $Exports{$Id};
$Contents .= sprintf (" href=\"%s\"", $Label);
}
# Add the HTML stuff to the output line
if ($Contents ne "") {
$OutLine .= sprintf ("%s", $Contents, $Id);
} else {
$OutLine .= $Id;
}
# Check if another identifier follows
if ($Line =~ s/^\s*(?::\s*[A-Za-z]+\s*)?,\s*//) {
$OutLine .= $&;
} else {
last;
}
}
# Add an remainder if there is one
$OutLine .= Cleanup ($Line);
# Handle export statements
} elsif ($Line =~ s/^\.export(?:zp)?\s+//i) {
# Print the command and the whitespace.
$OutLine .= $&;
# Print all identifiers if there are any
while ($Line =~ s/^[_a-zA-Z]\w*//) {
# Remember the identifier
my $Id = $&;
# Variable to assemble HTML representation
my $Contents = "";
# If we have a definition for this export in this file, add
# a link to the definition.
if (exists ($Labels{$OutName}{$Id})) {
$Label = $Labels{$OutName}{$Id};
$Contents = sprintf (" href=\"#%s\"", $Label);
}
# If we have this identifier in the list of exports, add a
# jump target for the export.
if (exists ($Exports{$Id})) {
$Label = $Exports{$Id};
# Be sure to use only the label part
$Label =~ s/^.*#//;
$Contents .= sprintf (" name=\"%s\"", $Label);
}
# Add the HTML stuff to the output line
if ($Contents ne "") {
$OutLine .= sprintf ("%s", $Contents, $Id);
} else {
$OutLine .= $Id;
}
# Check if another identifier follows
if ($Line =~ s/^\s*(?::\s*[A-Za-z]+\s*)?,\s*//) {
$OutLine .= $&;
} else {
last;
}
}
# Add an remainder if there is one
$OutLine .= Cleanup ($Line);
# Handle actor statements.
} elsif ($Line =~ s/^(\.(?:(?:(?:con|de)struc|interrup)tor|condes)\s+)([_a-z]\w*)//i) {
# Print the command and the whitespace.
$OutLine .= $1;
# Remember the identifier.
$Id = $2;
# Variable to assemble HTML representation
my $Contents = "";
# If we have a definition for this actor, in this file,
# then add a link to that definition.
if (exists ($Labels{$OutName}{$Id})) {
$Contents = sprintf (" href=\"#%s\"", $Labels{$OutName}{$Id});
}
# Get the target, for linking from imports in other files.
$Label = $Exports{$Id};
# Be sure to use only the label part.
$Label =~ s/^.*#//;
# Add the HTML stuff and the remainder of the actor
# to the output line.
$OutLine .= sprintf ("%s%s", $Label,
$Contents, $Id, Cleanup ($Line));
# Check for .faraddr, .addr, .dword, .word, .dbyt, .byt, .byte, .res,
# .elseif, .if, .align, and .org.
} elsif ($Line =~ s/^\.(?:(?:far)?addr|d?word|d?byte?|res|(?:else)?if|align|org)\s+//i) {
# Print the command and the white space
$OutLine .= $&;
# Print all identifiers if there are any
while ($Line =~ s/^([^_a-zA-Z]*?)(([\@?]?)[_a-zA-Z]\w*)//) {
# Add the non label stuff
$OutLine .= Cleanup ($1);
# Use the prefix if the label is local.
# Get the reference to that label if we find it.
$OutLine .= RefLabel ($OutName, ($3 ne "") ? "$CheapPrefix$2" : $2, $2);
}
# Add an remainder if there is one
$OutLine .= Cleanup ($Line);
# Handle .proc
} elsif ($Line =~ /^(\.proc)(\s+)([_a-z]\w*)?(.*)$/i) {
# Do we have an identifier?
if ($3 ne "") {
# Remember the ID as the new cheap-local prefix.
$CheapPrefix = $3;
# Get the label for the id
$Label = $Labels{$OutName}{$3};
# Print the label with a tag
$OutLine .= "$1$2$3";
} else {
# Print a line that has invalid syntax (its operand isn't
# a correctly formed name).
$OutLine .= "$1$2";
}
# Add the remainder
$OutLine .= Cleanup ($4);
# Handle .include
} elsif ($Line =~ /^(\.include)(\s*)\"((?:[^\"]+?|\\\")+)(\".*)$/i) {
# Add the fixed stuff to the output line
$OutLine .= "$1$2"";
# Get the filename into a named variable
my $FileName = Cleanup ($3);
# Get the name without a path
my $Name = StripPath ($3);
# If the include file is among the list of our files, add a link,
# otherwise just add the name as is.
if (exists ($Files{$Name})) {
$OutLine .= sprintf ("%s", GetOutName ($Name), $FileName);
} else {
$OutLine .= $FileName;
}
# Add the remainder
$OutLine .= Cleanup ($4);
# Handle .dbg line
} elsif ($CRefs && $Line =~ s/^\.dbg\s+//) {
# Add the fixed stuff to the output line
$OutLine .= $&;
# Check for the type of the .dbg directive
if ($Line =~ /^(line,\s*)\"((?:[^\"]+?|\\\")+)\"(,\s*)(\d+)(.*)$/) {
# Add the fixed stuff to the output line
$OutLine .= "$1"";
# Get the filename and line number into named variables
my $DbgFile = $2;
my $DbgLine = $4;
# Remember the remainder
$Line = "\"$3$4$5";
# Get the name without a path
my $Name = StripPath ($DbgFile);
# We don't need FileName any longer as is, so clean it up
$DbgFile = Cleanup ($DbgFile);
# Add a link to the source file
$OutLine .= sprintf ("%s", $Name, $DbgLine, $DbgFile);
# Add the remainder
$OutLine .= Cleanup ($Line);
} elsif ($Line =~ /^(file,\s*)\"((?:[^\"]+?|\\\")+)\"(.*)$/) { #pf FIXME: doesn't handle \" correctly!
# Get the filename into a named variables
my $DbgFile = Cleanup ($2);
# Get the name without a path
my $Name = Cleanup (StripPath ($2));
# Add the fixed stuff to the output line
$OutLine .= sprintf ("%s\"%s\"%s",
$1, $Name, $DbgFile, $3);
} else {
# Add the remainder
$OutLine .= Cleanup ($Line);
}
} elsif ($CRefs && $Line =~ /^(\.dbg)(\s+line,\s*)\"((?:[^\"]+?|\\\")+)\"(,\s*)(\d+)(.*$)/) {
# Add the fixed stuff to the output line
$OutLine .= "$1$2"";
# Get the filename and line number into named variables
my $FileName = $3;
my $LineNo = $5;
# Remember the remainder
$Line = "\"$4$5$6";
# Get the name without a path
my $Name = StripPath ($FileName);
# We don't need FileName any longer as is, so clean it up
$FileName = Cleanup ($FileName);
# Add a link to the source file
$OutLine .= sprintf ("%s", $Name, $LineNo, $FileName);
# Add the remainder
$OutLine .= Cleanup ($Line);
# Check for .ifdef, .ifndef, .ifref, and .ifnref.
} elsif ($Line =~ s/^(\.ifn?[dr]ef\s+)(([\@?]?)[_a-z]\w*)?//i) {
# Print the command and the whitespace.
$OutLine .= $1;
if ($2 ne "") {
# Use the prefix if the label is local.
# Get the reference to that label if we find it.
$OutLine .= RefLabel ($OutName, ($3 ne "") ? "$CheapPrefix$2" : $2, $2);
}
# Add a remainder if there is one.
$OutLine .= Cleanup ($Line);
# Check for assertions.
} elsif ($Line =~ s/^(\.assert\s+)(.+?)(,\s*(?:error|warning)\s*(?:,.*)?)$/$2/i) {
# Print the command and the whitespace.
$OutLine .= $1;
$Comment = $3;
# Print all identifiers if there are any.
while ($Line =~ s/^([^_a-zA-Z]*?)(([\@?]?)[_a-zA-Z]\w*)//) {
# Add the non-label stuff.
$OutLine .= Cleanup ($1);
# Use the prefix if the label is local.
# Get the reference to that label if we find it.
$OutLine .= RefLabel ($OutName, ($3 ne "") ? "$CheapPrefix$2" : $2, $2);
}
# Add a remainder if there is one.
$OutLine .= Cleanup ($Line . $Comment);
# Check for instructions with labels
} elsif ($Line =~ s/^($LabelIns)\b(\s*)//io) {
# Print the instruction and white space
$OutLine .= ColorizeKeyword ($1) . $2;
# Print all identifiers if there are any.
while ($Line =~ s/^([^_a-zA-Z]*?)(([\@?]?)[_a-zA-Z]\w*)//) {
# Add the non-label stuff.
$OutLine .= Cleanup ($1);
# Is this a local label?
if ($3 ne "") {
# Use the prefix
$Id = "$CheapPrefix$2";
} else {
# Use as is
$Id = $2;
}
# Get the reference to this label if we find it
$OutLine .= RefLabel ($OutName, $Id, $2);
}
# Reassemble and print the line
$OutLine .= Cleanup ($Line);
# Check for all other instructions
} elsif ($Line =~ /^($OtherIns)\b(.*)$/io) {
# Colorize and print
$OutLine .= ColorizeKeyword ($1) . Cleanup ($2);
} else {
# Nothing known - print the line
$OutLine .= Cleanup ($Line);
}
} continue {
# Colorize all keywords
$OutLine =~ s/(?\n";
DocFooter (OUTPUT, $OutName);
# Close the files
close (INPUT);
close (OUTPUT);
}
# Pass2: Read all files the second time.
sub Pass2 () {
# Keep the user happy
Gabble ("Pass 2");
# Walk over the files
for my $InName (keys (%Files)) {
# Process one file
Process2 ($InName);
}
}
#-----------------------------------------------------------------------------#
# Create an index page #
# ----------------------------------------------------------------------------#
# Print a list of all files
sub FileIndex {
# File is argument
my $INDEX = $_[0];
# Print the file list in a table
print $INDEX "Files
\n";
print $INDEX "
| %s | \n", GetOutName ($File), $File; if (($Count % $IndexCols) == $IndexCols-1) { print $INDEX "
\n"; } # Print a list of all exports sub ExportIndex { # File is argument my $INDEX = $_[0]; # Print the file list in a table print $INDEX "
\n"; print $INDEX "
| %s | \n", $File, $Label, $Export; if (($Count % $IndexCols) == $IndexCols-1) { print $INDEX "
\n"; } sub CreateIndex { # Open the index page file open (INDEX, ">$HTMLDir$IndexName") or Abort ("Cannot open $IndexName: $!"); # Print the header DocHeader (INDEX, $IndexTitle, 0); # Print the file list in a table FileIndex (INDEX); ExportIndex (INDEX); # Print the document footer DocFooter (INDEX, $IndexName); # Close the index file close (INDEX); } #-----------------------------------------------------------------------------# # Print usage information # # ----------------------------------------------------------------------------# sub Usage { print "Usage: ca65html [options] file ...\n"; print "Options:\n"; print " --bgcolor c Use background color c instead of $BGColor\n"; print " --colorize Add color highlights to the output\n"; print " --commentcolor c Use color c for comments instead of $CommentColor\n"; print " --crefs Generate references to the C source file(s)\n"; print " --ctrlcolor c Use color c for directives instead of $CtrlColor\n"; print " --cvttabs Convert tabs to spaces in the output\n"; print " --help This text\n"; print " --htmldir dir Specify directory for HTML files\n"; print " --indexcols n Use n columns on index page (default $IndexCols)\n"; print " --indexname file Use file for the index file instead of $IndexName\n"; print " --indexpage Create an index page\n"; print " --indextitle title Use title as the index title instead of $IndexTitle\n"; print " --keywordcolor c Use color c for keywords instead of $KeywordColor\n"; print " --linelabels Generate a linexxx HTML label for each line\n"; print " --linenumbers Add line numbers to the output\n"; print " --linkstyle style Use the given link style\n"; print " --replaceext Replace source extension instead of appending .html\n"; print " --tabsize n Use n spaces when replacing tabs (default $TabSize)\n"; print " --textcolor c Use text color c instead of $TextColor\n"; print " --verbose Be more verbose\n"; } #-----------------------------------------------------------------------------# # Main # # ----------------------------------------------------------------------------# # Get program options GetOptions ("bgcolor=s" => \$BGColor, "colorize" => \$Colorize, "commentcolor=s" => \$CommentColor, "crefs" => \$CRefs, "ctrlcolor=s" => \$CtrlColor, "cvttabs" => \$CvtTabs, "debug!" => \$Debug, "help" => \$Help, "htmldir=s" => \$HTMLDir, "indexcols=i" => \$IndexCols, "indexname=s" => \$IndexName, "indexpage" => \$IndexPage, "indextitle=s" => \$IndexTitle, "keywordcolor=s" => \$KeywordColor, "linelabels" => \$LineLabels, "linenumbers" => \$LineNumbers, "linkstyle=i" => \$LinkStyle, "replaceext" => \$ReplaceExt, "tabsize=i" => \$TabSize, "textcolor=s" => \$TextColor, "verbose!" => \$Verbose, "<>" => \&AddFile); # Check some arguments if ($IndexCols <= 0 || $IndexCols >= 20) { Abort ("Invalid value for --indexcols option"); } if ($TabSize < 1 || $TabSize > 16) { Abort ("Invalid value for --tabsize option"); } if ($HTMLDir ne "" && $HTMLDir =~ /[^\/]$/) { # Add a trailing path separator $HTMLDir .= "/"; } # Print help if requested if ($Help) { Usage (); } # Check if we have input files given if ($FileCount == 0) { Abort ("No input files"); } # Convert the documents Pass1 (); Pass2 (); # Generate an index page if requested if ($IndexPage) { CreateIndex (); } # Done exit 0;