private void OutputString(int offset, string labelStr, string commentStr)
{
// Generic strings whose encoding matches the configured text encoding are output
// with a simple .text directive.
//
// CString and L8String have directives (.null, .ptext), but we can only use
// them if the string fits on one line and doesn't include delimiters.
//
// We might be able to define a macro for Reverse.
//
// We don't currently switch character encodings in the middle of a file. We could
// do so to flip between PETSCII, screen codes, low ASCII, and high ASCII, but it
// adds a lot of noise and it's unclear that this is generally useful.
Anattrib attr = Project.GetAnattrib(offset);
FormatDescriptor dfd = attr.DataDescriptor;
Debug.Assert(dfd != null);
Debug.Assert(dfd.IsString);
Debug.Assert(dfd.Length > 0);
CharEncoding.Convert charConv = null;
CharEncoding.Convert dciConv = null;
switch (dfd.FormatSubType)
{
case FormatDescriptor.SubType.Ascii:
charConv = CharEncoding.ConvertAscii;
dciConv = CharEncoding.ConvertLowAndHighAscii;
break;
case FormatDescriptor.SubType.HighAscii:
charConv = CharEncoding.ConvertHighAscii;
dciConv = CharEncoding.ConvertLowAndHighAscii;
break;
case FormatDescriptor.SubType.C64Petscii:
charConv = CharEncoding.ConvertC64Petscii;
dciConv = CharEncoding.ConvertLowAndHighC64Petscii;
break;
case FormatDescriptor.SubType.C64Screen:
charConv = CharEncoding.ConvertC64ScreenCode;
dciConv = CharEncoding.ConvertLowAndHighC64ScreenCode;
break;
default:
break;
}
if (charConv == null)
{
OutputNoJoy(offset, dfd.Length, labelStr, commentStr);
return;
}
// Issue a .enc, if needed.
UpdateCharacterEncoding(dfd);
Formatter formatter = SourceFormatter;
byte[] data = Project.FileData;
int hiddenLeadingBytes = 0;
int shownLeadingBytes = 0;
int trailingBytes = 0;
string opcodeStr;
switch (dfd.FormatType)
{
case FormatDescriptor.Type.StringGeneric:
case FormatDescriptor.Type.StringReverse:
opcodeStr = sDataOpNames.StrGeneric;
break;
case FormatDescriptor.Type.StringNullTerm:
opcodeStr = sDataOpNames.StrNullTerm;
trailingBytes = 1;
break;
case FormatDescriptor.Type.StringL8:
opcodeStr = sDataOpNames.StrLen8;
hiddenLeadingBytes = 1;
break;
case FormatDescriptor.Type.StringL16:
opcodeStr = sDataOpNames.StrGeneric;
shownLeadingBytes = 2;
break;
case FormatDescriptor.Type.StringDci:
opcodeStr = sDataOpNames.StrDci;
if ((Project.FileData[offset] & 0x80) != 0)
{
// ".shift" directive only works for strings where the low bit starts
// clear and ends high.
OutputNoJoy(offset, dfd.Length, labelStr, commentStr);
return;
}
break;
default:
Debug.Assert(false);
return;
}
StringOpFormatter stropf = new StringOpFormatter(SourceFormatter,
Formatter.DOUBLE_QUOTE_DELIM, StringOpFormatter.RawOutputStyle.CommaSep,
MAX_OPERAND_LEN, charConv);
if (dfd.FormatType == FormatDescriptor.Type.StringDci)
{
// DCI is awkward because the character encoding flips on the last byte. Rather
// than clutter up StringOpFormatter for this rare item, we just accept low/high
// throughout.
stropf.CharConv = dciConv;
}
// Feed bytes in, skipping over hidden bytes (leading L8, trailing null).
stropf.FeedBytes(data, offset + hiddenLeadingBytes,
dfd.Length - hiddenLeadingBytes - trailingBytes, shownLeadingBytes,
StringOpFormatter.ReverseMode.Forward);
Debug.Assert(stropf.Lines.Count > 0);
// See if we need to do this over.
bool redo = false;
switch (dfd.FormatType)
{
case FormatDescriptor.Type.StringGeneric:
case FormatDescriptor.Type.StringReverse:
case FormatDescriptor.Type.StringL16:
// All good the first time.
break;
case FormatDescriptor.Type.StringNullTerm:
case FormatDescriptor.Type.StringL8:
case FormatDescriptor.Type.StringDci:
if (stropf.Lines.Count != 1)
{
// Must be single-line.
opcodeStr = sDataOpNames.StrGeneric;
stropf.CharConv = charConv; // undo DCI hack
redo = true;
}
break;
default:
Debug.Assert(false);
return;
}
if (redo)
{
//Debug.WriteLine("REDO off=+" + offset.ToString("x6") + ": " + dfd.FormatType);
// This time, instead of skipping over leading length bytes, we include them
// explicitly.
stropf.Reset();
stropf.FeedBytes(data, offset, dfd.Length, hiddenLeadingBytes,
StringOpFormatter.ReverseMode.Forward);
}
opcodeStr = formatter.FormatPseudoOp(opcodeStr);
foreach (string str in stropf.Lines)
{
OutputLine(labelStr, opcodeStr, str, commentStr);
labelStr = commentStr = string.Empty; // only show on first
}
}
private void OutputString(int offset, string labelStr, string commentStr)
{
Formatter formatter = SourceFormatter;
byte[] data = Project.FileData;
Anattrib attr = Project.GetAnattrib(offset);
FormatDescriptor dfd = attr.DataDescriptor;
Debug.Assert(dfd != null);
Debug.Assert(dfd.IsString);
Debug.Assert(dfd.Length > 0);
string opcodeStr;
CharEncoding.Convert charConv;
switch (dfd.FormatSubType)
{
case FormatDescriptor.SubType.Ascii:
opcodeStr = sDataOpNames.StrGeneric;
charConv = CharEncoding.ConvertAscii;
break;
case FormatDescriptor.SubType.HighAscii:
opcodeStr = sDataOpNames.StrGeneric;
charConv = CharEncoding.ConvertHighAscii;
break;
case FormatDescriptor.SubType.C64Petscii:
opcodeStr = "!pet";
charConv = CharEncoding.ConvertC64Petscii;
break;
case FormatDescriptor.SubType.C64Screen:
opcodeStr = "!scr";
charConv = CharEncoding.ConvertC64ScreenCode;
break;
default:
Debug.Assert(false);
OutputNoJoy(offset, dfd.Length, labelStr, commentStr);
return;
}
int leadingBytes = 0;
switch (dfd.FormatType)
{
case FormatDescriptor.Type.StringGeneric:
case FormatDescriptor.Type.StringReverse:
case FormatDescriptor.Type.StringNullTerm:
case FormatDescriptor.Type.StringDci:
// Last byte may be output as hex.
break;
case FormatDescriptor.Type.StringL8:
// Length byte will be output as hex.
leadingBytes = 1;
break;
case FormatDescriptor.Type.StringL16:
// Length byte will be output as hex.
leadingBytes = 2;
break;
default:
Debug.Assert(false);
return;
}
StringOpFormatter stropf = new StringOpFormatter(SourceFormatter,
Formatter.DOUBLE_QUOTE_DELIM, StringOpFormatter.RawOutputStyle.CommaSep, charConv);
stropf.FeedBytes(data, offset, dfd.Length, leadingBytes,
StringOpFormatter.ReverseMode.Forward);
if (dfd.FormatSubType == FormatDescriptor.SubType.HighAscii && stropf.HasEscapedText)
{
// Can't !xor the output, because while it works for string data it
// also flips the high bits on the unprintable bytes we output as raw hex.
OutputNoJoy(offset, dfd.Length, labelStr, commentStr);
return;
}
if (dfd.FormatSubType == FormatDescriptor.SubType.HighAscii)
{
OutputLine(string.Empty, "!xor", "$80 {", string.Empty);
}
foreach (string str in stropf.Lines)
{
OutputLine(labelStr, opcodeStr, str, commentStr);
labelStr = commentStr = string.Empty; // only show on first
}
if (dfd.FormatSubType == FormatDescriptor.SubType.HighAscii)
{
OutputLine(string.Empty, "}", string.Empty, string.Empty);
}
}
private void OutputString(int offset, string labelStr, string commentStr)
{
// Normal ASCII strings are straightforward: they're just part of a .byte
// directive, and can mix with anything else in the .byte.
//
// For CString we can use .asciiz, but only if the string fits on one line
// and doesn't include delimiters. For L8String and L16String we can
// define simple macros, but their use has a similar restriction. High-ASCII
// strings also require a macro.
//
// We might be able to define a macro for DCI and Reverse as well.
//
// The limitation on strings with delimiters arises because (1) I don't see a
// way to escape them within a string, and (2) the simple macro workarounds
// only take a single argument, not a comma-separated list of stuff.
//
// Some ideas here:
// https://groups.google.com/forum/#!topic/comp.sys.apple2.programmer/5Wkw8mUPcU0
Anattrib attr = Project.GetAnattrib(offset);
FormatDescriptor dfd = attr.DataDescriptor;
Debug.Assert(dfd != null);
Debug.Assert(dfd.IsString);
Debug.Assert(dfd.Length > 0);
CharEncoding.Convert charConv;
bool isHighAscii = false;
switch (dfd.FormatSubType)
{
case FormatDescriptor.SubType.Ascii:
charConv = CharEncoding.ConvertAscii;
break;
case FormatDescriptor.SubType.HighAscii:
if (dfd.FormatType != FormatDescriptor.Type.StringGeneric)
{
OutputNoJoy(offset, dfd.Length, labelStr, commentStr);
return;
}
charConv = CharEncoding.ConvertHighAscii;
isHighAscii = true;
break;
case FormatDescriptor.SubType.C64Petscii:
case FormatDescriptor.SubType.C64Screen:
default:
OutputNoJoy(offset, dfd.Length, labelStr, commentStr);
return;
}
Formatter formatter = SourceFormatter;
byte[] data = Project.FileData;
int leadingBytes = 0;
int trailingBytes = 0;
switch (dfd.FormatType)
{
case FormatDescriptor.Type.StringGeneric:
case FormatDescriptor.Type.StringReverse:
case FormatDescriptor.Type.StringDci:
break;
case FormatDescriptor.Type.StringNullTerm:
trailingBytes = 1;
break;
case FormatDescriptor.Type.StringL8:
leadingBytes = 1;
break;
case FormatDescriptor.Type.StringL16:
leadingBytes = 2;
break;
default:
Debug.Assert(false);
return;
}
StringOpFormatter stropf = new StringOpFormatter(SourceFormatter,
Formatter.DOUBLE_QUOTE_DELIM, StringOpFormatter.RawOutputStyle.CommaSep, charConv);
stropf.FeedBytes(data, offset, dfd.Length - trailingBytes, leadingBytes,
StringOpFormatter.ReverseMode.Forward);
string opcodeStr = formatter.FormatPseudoOp(sDataOpNames.StrGeneric);
if (isHighAscii)
{
// Does this fit the narrow definition of what we can do with a macro?
Debug.Assert(dfd.FormatType == FormatDescriptor.Type.StringGeneric);
if (stropf.Lines.Count == 1 && !stropf.HasEscapedText)
{
if (!mHighAsciiMacroOutput)
{
mHighAsciiMacroOutput = true;
// Output a macro for high-ASCII strings.
//
// TODO(maybe): the preferred way to do this is apparently
// ".macpack apple2" to load some standard macros, then e.g.
// scrcode "My high-ASCII string". The macro takes 9 arguments and
// recognizes characters and numbers, so it should be possible to
// mix strings, string delimiters, and control chars so long as the
// argument count is not exceeded.
OutputLine(".macro", "HiAscii", "Arg", string.Empty);
OutputLine(string.Empty, ".repeat", ".strlen(Arg), I", string.Empty);
OutputLine(string.Empty, ".byte", ".strat(Arg, I) | $80", string.Empty);
OutputLine(string.Empty, ".endrep", string.Empty, string.Empty);
OutputLine(".endmacro", string.Empty, string.Empty, string.Empty);
}
opcodeStr = formatter.FormatPseudoOp("HiAscii");
}
else
{
// didn't work out, dump hex
OutputNoJoy(offset, dfd.Length, labelStr, commentStr);
return;
}
}
if (dfd.FormatType == FormatDescriptor.Type.StringNullTerm)
{
if (stropf.Lines.Count == 1 && !stropf.HasEscapedText)
{
// Keep it.
opcodeStr = sDataOpNames.StrNullTerm;
}
else
{
// Didn't fit, so re-emit it, this time with the terminating null byte.
stropf.Reset();
stropf.FeedBytes(data, offset, dfd.Length, leadingBytes,
StringOpFormatter.ReverseMode.Forward);
}
}
foreach (string str in stropf.Lines)
{
OutputLine(labelStr, opcodeStr, str, commentStr);
labelStr = commentStr = string.Empty; // only show on first
}
}
private void OutputString(int offset, string labelStr, string commentStr)
{
// This gets complicated.
//
// For Dci, L8String, and L16String, the entire string needs to fit in the
// operand of one line. If it can't, we need to separate the length byte/word
// or inverted character out, and just dump the rest as ASCII. Computing the
// line length requires factoring delimiter character escapes. (NOTE: contrary
// to the documentation, STR and STRL do include trailing hex characters in the
// length calculation, so it's possible to escape delimiters.)
//
// For Reverse, we can span lines, but only if we emit the lines in
// backward order. Also, Merlin doesn't allow hex to be embedded in a REV
// operation, so we can't use REV if the string contains a delimiter.
//
// For aesthetic purposes, zero-length CString, L8String, and L16String
// should be output as DFB/DW zeroes rather than an empty string -- makes
// it easier to read.
//
// NOTE: we generally assume that the input is in the correct format, e.g.
// the length byte in a StringL8 matches dfd.Length, and the high bits in DCI strings
// have the right pattern. If not, we will generate bad output. This would need
// to be scanned and corrected at a higher level.
Anattrib attr = Project.GetAnattrib(offset);
FormatDescriptor dfd = attr.DataDescriptor;
Debug.Assert(dfd != null);
Debug.Assert(dfd.IsString);
Debug.Assert(dfd.Length > 0);
// We can sort of do parts of C64 stuff, but it's probably more readable to just
// output a commented blob than something where only the capital letters are readable.
switch (dfd.FormatSubType)
{
case FormatDescriptor.SubType.Ascii:
case FormatDescriptor.SubType.HighAscii:
break;
case FormatDescriptor.SubType.C64Petscii:
case FormatDescriptor.SubType.C64Screen:
default:
OutputNoJoy(offset, dfd.Length, labelStr, commentStr);
return;
}
Formatter formatter = SourceFormatter;
byte[] data = Project.FileData;
StringOpFormatter.ReverseMode revMode = StringOpFormatter.ReverseMode.Forward;
int leadingBytes = 0;
string opcodeStr;
switch (dfd.FormatType)
{
case FormatDescriptor.Type.StringGeneric:
opcodeStr = sDataOpNames.StrGeneric;
break;
case FormatDescriptor.Type.StringReverse:
opcodeStr = sDataOpNames.StrReverse;
revMode = StringOpFormatter.ReverseMode.LineReverse;
break;
case FormatDescriptor.Type.StringNullTerm:
opcodeStr = sDataOpNames.StrGeneric; // no pseudo-op for this
if (dfd.Length == 1)
{
// Empty string. Just output the length byte(s) or null terminator.
GenerateShortSequence(offset, 1, out string opcode, out string operand);
OutputLine(labelStr, opcode, operand, commentStr);
return;
}
break;
case FormatDescriptor.Type.StringL8:
opcodeStr = sDataOpNames.StrLen8;
leadingBytes = 1;
break;
case FormatDescriptor.Type.StringL16:
opcodeStr = sDataOpNames.StrLen16;
leadingBytes = 2;
break;
case FormatDescriptor.Type.StringDci:
opcodeStr = sDataOpNames.StrDci;
break;
default:
Debug.Assert(false);
return;
}
// Merlin 32 uses single-quote for low ASCII, double-quote for high ASCII.
CharEncoding.Convert charConv;
char delim;
if (dfd.FormatSubType == FormatDescriptor.SubType.HighAscii)
{
charConv = CharEncoding.ConvertHighAscii;
delim = '"';
}
else
{
charConv = CharEncoding.ConvertAscii;
delim = '\'';
}
StringOpFormatter stropf = new StringOpFormatter(SourceFormatter,
new Formatter.DelimiterDef(delim),
StringOpFormatter.RawOutputStyle.DenseHex, charConv);
if (dfd.FormatType == FormatDescriptor.Type.StringDci)
{
// DCI is awkward because the character encoding flips on the last byte. Rather
// than clutter up StringOpFormatter for this rare item, we just accept low/high
// throughout.
stropf.CharConv = CharEncoding.ConvertLowAndHighAscii;
}
// Feed bytes in, skipping over the leading length bytes.
stropf.FeedBytes(data, offset + leadingBytes,
dfd.Length - leadingBytes, 0, revMode);
Debug.Assert(stropf.Lines.Count > 0);
// See if we need to do this over.
bool redo = false;
switch (dfd.FormatType)
{
case FormatDescriptor.Type.StringGeneric:
case FormatDescriptor.Type.StringNullTerm:
break;
case FormatDescriptor.Type.StringReverse:
if (stropf.HasEscapedText)
{
// can't include escaped characters in REV
opcodeStr = sDataOpNames.StrGeneric;
revMode = StringOpFormatter.ReverseMode.Forward;
redo = true;
}
break;
case FormatDescriptor.Type.StringL8:
if (stropf.Lines.Count != 1)
{
// single-line only
opcodeStr = sDataOpNames.StrGeneric;
leadingBytes = 1;
redo = true;
}
break;
case FormatDescriptor.Type.StringL16:
if (stropf.Lines.Count != 1)
{
// single-line only
opcodeStr = sDataOpNames.StrGeneric;
leadingBytes = 2;
redo = true;
}
break;
case FormatDescriptor.Type.StringDci:
if (stropf.Lines.Count != 1)
{
// single-line only
opcodeStr = sDataOpNames.StrGeneric;
stropf.CharConv = charConv;
redo = true;
}
break;
default:
Debug.Assert(false);
return;
}
if (redo)
{
//Debug.WriteLine("REDO off=+" + offset.ToString("x6") + ": " + dfd.FormatType);
// This time, instead of skipping over leading length bytes, we include them
// explicitly.
stropf.Reset();
stropf.FeedBytes(data, offset, dfd.Length, leadingBytes, revMode);
}
opcodeStr = formatter.FormatPseudoOp(opcodeStr);
foreach (string str in stropf.Lines)
{
OutputLine(labelStr, opcodeStr, str, commentStr);
labelStr = commentStr = string.Empty; // only show on first
}
}
/// <summary>
/// Converts a collection of bytes that represent a string into an array of formatted
/// string operands.
/// </summary>
/// <param name="formatter">Formatter object.</param>
/// <param name="opNames">Pseudo-opcode name table.</param>
/// <param name="dfd">Format descriptor.</param>
/// <param name="data">File data.</param>
/// <param name="offset">Offset, within data, of start of string.</param>
/// <param name="popcode">Pseudo-opcode string.</param>
/// <returns>Array of operand strings.</returns>
public static List <string> FormatStringOp(Formatter formatter, PseudoOpNames opNames,
FormatDescriptor dfd, byte[] data, int offset, out string popcode)
{
int hiddenLeadingBytes = 0;
int trailingBytes = 0;
StringOpFormatter.ReverseMode revMode = StringOpFormatter.ReverseMode.Forward;
Formatter.DelimiterSet delSet = formatter.Config.mStringDelimiters;
Formatter.DelimiterDef delDef;
CharEncoding.Convert charConv;
switch (dfd.FormatSubType)
{
case FormatDescriptor.SubType.Ascii:
if (dfd.FormatType == FormatDescriptor.Type.StringDci)
{
charConv = CharEncoding.ConvertLowAndHighAscii;
}
else
{
charConv = CharEncoding.ConvertAscii;
}
delDef = delSet.Get(CharEncoding.Encoding.Ascii);
break;
case FormatDescriptor.SubType.HighAscii:
if (dfd.FormatType == FormatDescriptor.Type.StringDci)
{
charConv = CharEncoding.ConvertLowAndHighAscii;
}
else
{
charConv = CharEncoding.ConvertHighAscii;
}
delDef = delSet.Get(CharEncoding.Encoding.HighAscii);
break;
case FormatDescriptor.SubType.C64Petscii:
if (dfd.FormatType == FormatDescriptor.Type.StringDci)
{
charConv = CharEncoding.ConvertLowAndHighC64Petscii;
}
else
{
charConv = CharEncoding.ConvertC64Petscii;
}
delDef = delSet.Get(CharEncoding.Encoding.C64Petscii);
break;
case FormatDescriptor.SubType.C64Screen:
if (dfd.FormatType == FormatDescriptor.Type.StringDci)
{
charConv = CharEncoding.ConvertLowAndHighC64ScreenCode;
}
else
{
charConv = CharEncoding.ConvertC64ScreenCode;
}
delDef = delSet.Get(CharEncoding.Encoding.C64ScreenCode);
break;
default:
Debug.Assert(false);
charConv = CharEncoding.ConvertAscii;
delDef = delSet.Get(CharEncoding.Encoding.Ascii);
break;
}
if (delDef == null)
{
delDef = Formatter.DOUBLE_QUOTE_DELIM;
}
switch (dfd.FormatType)
{
case FormatDescriptor.Type.StringGeneric:
// Generic character data.
popcode = opNames.StrGeneric;
break;
case FormatDescriptor.Type.StringReverse:
// Character data, full width specified by formatter. Show characters
// in reverse order.
popcode = opNames.StrReverse;
revMode = StringOpFormatter.ReverseMode.FullReverse;
break;
case FormatDescriptor.Type.StringNullTerm:
// Character data with a terminating null. Don't show the null byte.
popcode = opNames.StrNullTerm;
trailingBytes = 1;
//if (strLen == 0) {
// showHexZeroes = 1;
//}
break;
case FormatDescriptor.Type.StringL8:
// Character data with a leading length byte. Don't show the length.
hiddenLeadingBytes = 1;
//if (strLen == 0) {
// showHexZeroes = 1;
//}
popcode = opNames.StrLen8;
break;
case FormatDescriptor.Type.StringL16:
// Character data with a leading length word. Don't show the length.
Debug.Assert(dfd.Length > 1);
hiddenLeadingBytes = 2;
//if (strLen == 0) {
// showHexZeroes = 2;
//}
popcode = opNames.StrLen16;
break;
case FormatDescriptor.Type.StringDci:
// High bit on last byte is flipped.
popcode = opNames.StrDci;
break;
default:
Debug.Assert(false);
popcode = ".!!!";
break;
}
StringOpFormatter stropf = new StringOpFormatter(formatter, delDef,
StringOpFormatter.RawOutputStyle.CommaSep, MAX_OPERAND_LEN, charConv);
stropf.FeedBytes(data, offset + hiddenLeadingBytes,
dfd.Length - hiddenLeadingBytes - trailingBytes, 0, revMode);
return(stropf.Lines);
}