1
0
mirror of https://github.com/fadden/6502bench.git synced 2024-11-12 06:06:12 +00:00
6502bench/SourceGen/AsmGen/AsmTass64.cs
Andy McFadden dae76d9b45 Rework string operand formatting
This generalizes the string pseudo-operand formatter, moving it into
the Asm65 library.  The assembly source generators have been updated
to use it.  This makes the individual generators simpler, and by
virtue of avoiding "test runs" should make them slightly faster.

This also introduces byte-to-character converters, though we're
currently still only supporting low/high ASCII.

Regression test output is unchanged.
2019-08-09 17:46:33 -07:00

751 lines
30 KiB
C#

/*
* Copyright 2019 faddenSoft
*
* Licensed under the Apache License, Version 2.0 (the "License");
* you may not use this file except in compliance with the License.
* You may obtain a copy of the License at
*
* http://www.apache.org/licenses/LICENSE-2.0
*
* Unless required by applicable law or agreed to in writing, software
* distributed under the License is distributed on an "AS IS" BASIS,
* WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
* See the License for the specific language governing permissions and
* limitations under the License.
*/
using System;
using System.Collections.Generic;
using System.ComponentModel;
using System.Diagnostics;
using System.IO;
using System.Text;
using Asm65;
using CommonUtil;
namespace SourceGen.AsmGen {
#region IGenerator
/// <summary>
/// Generate source code compatible with the 64tass assembler
/// (https://sourceforge.net/projects/tass64/).
///
/// The assembler is officially called "64tass", but it's sometimes written "tass64" because
/// in some cases you can't start an identifier with a number.
///
/// We need to deal with a couple of unusual aspects:
/// (1) The prefix for a local label is '_', which is generally a legal character. So
/// if somebody creates a label with a leading '_', and it's not actually local, we have
/// to "de-local" it somehow.
/// (2) By default, labels are handled in a case-insensitive fashion, which is extremely
/// rare for programming languages. Case sensitivity can be enabled with the "-C" flag.
/// Anybody who wants to assemble the generated code will need to be aware of this.
/// </summary>
public class GenTass64 : IGenerator {
private const string ASM_FILE_SUFFIX = "_64tass.S"; // must start with underscore
private const int MAX_OPERAND_LEN = 64;
// IGenerator
public DisasmProject Project { get; private set; }
// IGenerator
public Formatter SourceFormatter { get; private set; }
// IGenerator
public AppSettings Settings { get; private set; }
// IGenerator
public AssemblerQuirks Quirks { get; private set; }
// IGenerator
public LabelLocalizer Localizer { get { return mLocalizer; } }
/// <summary>
/// Working directory, i.e. where we write our output file(s).
/// </summary>
private string mWorkDirectory;
/// <summary>
/// If set, long labels get their own line.
/// </summary>
private bool mLongLabelNewLine;
/// <summary>
/// Output column widths.
/// </summary>
private int[] mColumnWidths;
/// <summary>
/// Base filename. Typically the project file name without the ".dis65" extension.
/// </summary>
private string mFileNameBase;
/// <summary>
/// StringBuilder to use when composing a line. Held here to reduce allocations.
/// </summary>
private StringBuilder mLineBuilder = new StringBuilder(100);
/// <summary>
/// Label localization helper.
/// </summary>
private LabelLocalizer mLocalizer;
/// <summary>
/// Stream to send the output to.
/// </summary>
private StreamWriter mOutStream;
/// <summary>
/// If we output a ".logical", we will need a ".here" eventually.
/// </summary>
private bool mNeedHereOp;
/// <summary>
/// Holds detected version of configured assembler.
/// </summary>
private CommonUtil.Version mAsmVersion = CommonUtil.Version.NO_VERSION;
// Version we're coded against.
private static CommonUtil.Version V1_53 = new CommonUtil.Version(1, 53, 1515);
// Pseudo-op string constants.
private static PseudoOp.PseudoOpNames sDataOpNames = new PseudoOp.PseudoOpNames() {
EquDirective = "=",
OrgDirective = ".logical",
//RegWidthDirective // .as, .al, .xs, .xl
DefineData1 = ".byte",
DefineData2 = ".word",
DefineData3 = ".long",
DefineData4 = ".dword",
//DefineBigData2
//DefineBigData3
//DefineBigData4
Fill = ".fill",
//Dense // no equivalent, use .byte with comma-separated args
StrGeneric = ".text",
//StrReverse
StrNullTerm = ".null",
StrLen8 = ".ptext",
//StrLen16
//StrDci
};
private const string HERE_PSEUDO_OP = ".here";
// IGenerator
public void GetDefaultDisplayFormat(out PseudoOp.PseudoOpNames pseudoOps,
out Formatter.FormatConfig formatConfig) {
pseudoOps = sDataOpNames.GetCopy();
formatConfig = new Formatter.FormatConfig();
SetFormatConfigValues(ref formatConfig);
}
// IGenerator
public void Configure(DisasmProject project, string workDirectory, string fileNameBase,
AssemblerVersion asmVersion, AppSettings settings) {
Debug.Assert(project != null);
Debug.Assert(!string.IsNullOrEmpty(workDirectory));
Debug.Assert(!string.IsNullOrEmpty(fileNameBase));
Project = project;
Quirks = new AssemblerQuirks();
mWorkDirectory = workDirectory;
mFileNameBase = fileNameBase;
Settings = settings;
mLongLabelNewLine = Settings.GetBool(AppSettings.SRCGEN_LONG_LABEL_NEW_LINE, false);
AssemblerConfig config = AssemblerConfig.GetConfig(settings,
AssemblerInfo.Id.Tass64);
mColumnWidths = (int[])config.ColumnWidths.Clone();
}
/// <summary>
/// Configures the assembler-specific format items.
/// </summary>
private void SetFormatConfigValues(ref Formatter.FormatConfig config) {
// Must be lower case when --case-sensitive is used.
config.mUpperOpcodes = false;
config.mUpperPseudoOpcodes = false;
config.mUpperOperandA = false;
config.mUpperOperandS = false;
config.mUpperOperandXY = false;
config.mBankSelectBackQuote = true;
config.mForceDirectOpcodeSuffix = string.Empty;
config.mForceAbsOpcodeSuffix = string.Empty;
config.mForceLongOpcodeSuffix = string.Empty;
config.mForceDirectOperandPrefix = string.Empty;
config.mForceAbsOperandPrefix = "@w"; // word
config.mForceLongOperandPrefix = "@l"; // long
config.mEndOfLineCommentDelimiter = ";";
config.mFullLineCommentDelimiterBase = ";";
config.mBoxLineCommentDelimiter = ";";
config.mAllowHighAsciiCharConst = false;
config.mExpressionMode = Formatter.FormatConfig.ExpressionMode.Common;
}
// IGenerator
public List<string> GenerateSource(BackgroundWorker worker) {
List<string> pathNames = new List<string>(1);
string fileName = mFileNameBase + ASM_FILE_SUFFIX;
string pathName = Path.Combine(mWorkDirectory, fileName);
pathNames.Add(pathName);
Formatter.FormatConfig config = new Formatter.FormatConfig();
GenCommon.ConfigureFormatterFromSettings(Settings, ref config);
SetFormatConfigValues(ref config);
SourceFormatter = new Formatter(config);
string msg = string.Format(Res.Strings.PROGRESS_GENERATING_FMT, pathName);
worker.ReportProgress(0, msg);
mLocalizer = new LabelLocalizer(Project);
if (!Settings.GetBool(AppSettings.SRCGEN_DISABLE_LABEL_LOCALIZATION, false)) {
mLocalizer.LocalPrefix = "_";
mLocalizer.Analyze();
}
mLocalizer.MaskLeadingUnderscores();
// Use UTF-8 encoding, without a byte-order mark.
using (StreamWriter sw = new StreamWriter(pathName, false, new UTF8Encoding(false))) {
mOutStream = sw;
if (Settings.GetBool(AppSettings.SRCGEN_ADD_IDENT_COMMENT, false)) {
OutputLine(SourceFormatter.FullLineCommentDelimiter +
string.Format(Res.Strings.GENERATED_FOR_VERSION_FMT,
"64tass", V1_53, AsmTass64.OPTIONS));
}
GenCommon.Generate(this, sw, worker);
if (mNeedHereOp) {
OutputLine(string.Empty, SourceFormatter.FormatPseudoOp(HERE_PSEUDO_OP),
string.Empty, string.Empty);
}
}
mOutStream = null;
return pathNames;
}
// IGenerator
public void OutputAsmConfig() {
CpuDef cpuDef = Project.CpuDef;
string cpuStr;
if (cpuDef.Type == CpuDef.CpuType.Cpu65816) {
cpuStr = "65816";
} else if (cpuDef.Type == CpuDef.CpuType.Cpu65C02) {
cpuStr = "65c02";
} else if (cpuDef.Type == CpuDef.CpuType.Cpu6502 && cpuDef.HasUndocumented) {
cpuStr = "6502i";
} else {
cpuStr = "6502";
}
OutputLine(string.Empty, SourceFormatter.FormatPseudoOp(".cpu"),
'\"' + cpuStr + '\"', string.Empty);
}
// IGenerator
public string ModifyOpcode(int offset, OpDef op) {
if (op.IsUndocumented) {
if (Project.CpuDef.Type == CpuDef.CpuType.Cpu65C02) {
// none of the "LDD" stuff is handled
return null;
}
if ((op.Mnemonic == OpName.ANC && op.Opcode != 0x0b) ||
(op.Mnemonic == OpName.JAM && op.Opcode != 0x02)) {
// There are multiple opcodes that match the mnemonic. Output the
// mnemonic for the first one and hex for the rest.
return null;
} else if (op.Mnemonic == OpName.NOP || op.Mnemonic == OpName.DOP ||
op.Mnemonic == OpName.TOP) {
// the various undocumented no-ops aren't handled
return null;
} else if (op.Mnemonic == OpName.SBC) {
// this is the alternate reference to SBC
return null;
} else if (op == OpDef.OpSHA_DPIndIndexY) {
// not recognized ($93)
return null;
}
}
if (op == OpDef.OpCOP_StackInt || op == OpDef.OpWDM_WDM) {
// 64tass doesn't like these to have an operand. Output as hex.
return null;
}
return string.Empty; // indicate original is fine
}
// IGenerator
public void GenerateShortSequence(int offset, int length, out string opcode,
out string operand) {
Debug.Assert(length >= 1 && length <= 4);
// Use a comma-separated list of individual hex bytes.
opcode = sDataOpNames.DefineData1;
StringBuilder sb = new StringBuilder(length * 4);
for (int i = 0; i < length; i++) {
if (i != 0) {
sb.Append(',');
}
sb.Append(SourceFormatter.FormatHexValue(Project.FileData[offset + i], 2));
}
operand = sb.ToString();
}
// IGenerator
public void OutputDataOp(int offset) {
Formatter formatter = SourceFormatter;
byte[] data = Project.FileData;
Anattrib attr = Project.GetAnattrib(offset);
string labelStr = string.Empty;
if (attr.Symbol != null) {
labelStr = mLocalizer.ConvLabel(attr.Symbol.Label);
}
string commentStr = SourceFormatter.FormatEolComment(Project.Comments[offset]);
string opcodeStr, operandStr;
FormatDescriptor dfd = attr.DataDescriptor;
Debug.Assert(dfd != null);
int length = dfd.Length;
Debug.Assert(length > 0);
bool multiLine = false;
switch (dfd.FormatType) {
case FormatDescriptor.Type.Default:
if (length != 1) {
Debug.Assert(false);
length = 1;
}
opcodeStr = sDataOpNames.DefineData1;
int operand = RawData.GetWord(data, offset, length, false);
operandStr = formatter.FormatHexValue(operand, length * 2);
break;
case FormatDescriptor.Type.NumericLE:
opcodeStr = sDataOpNames.GetDefineData(length);
operand = RawData.GetWord(data, offset, length, false);
operandStr = PseudoOp.FormatNumericOperand(formatter, Project.SymbolTable,
mLocalizer.LabelMap, dfd, operand, length,
PseudoOp.FormatNumericOpFlags.None);
break;
case FormatDescriptor.Type.NumericBE:
opcodeStr = sDataOpNames.GetDefineBigData(length);
if (opcodeStr == null) {
// Nothing defined, output as comma-separated single-byte values.
GenerateShortSequence(offset, length, out opcodeStr, out operandStr);
} else {
operand = RawData.GetWord(data, offset, length, true);
operandStr = PseudoOp.FormatNumericOperand(formatter, Project.SymbolTable,
mLocalizer.LabelMap, dfd, operand, length,
PseudoOp.FormatNumericOpFlags.None);
}
break;
case FormatDescriptor.Type.Fill:
opcodeStr = sDataOpNames.Fill;
operandStr = length + "," + formatter.FormatHexValue(data[offset], 2);
break;
case FormatDescriptor.Type.Dense:
multiLine = true;
opcodeStr = operandStr = null;
OutputDenseHex(offset, length, labelStr, commentStr);
break;
case FormatDescriptor.Type.StringGeneric:
case FormatDescriptor.Type.StringReverse:
case FormatDescriptor.Type.StringNullTerm:
case FormatDescriptor.Type.StringL8:
case FormatDescriptor.Type.StringL16:
case FormatDescriptor.Type.StringDci:
multiLine = true;
opcodeStr = operandStr = null;
OutputString(offset, labelStr, commentStr);
break;
default:
opcodeStr = "???";
operandStr = "***";
break;
}
if (!multiLine) {
opcodeStr = formatter.FormatPseudoOp(opcodeStr);
OutputLine(labelStr, opcodeStr, operandStr, commentStr);
}
}
private void OutputDenseHex(int offset, int length, string labelStr, string commentStr) {
Formatter formatter = SourceFormatter;
byte[] data = Project.FileData;
StringBuilder sb = new StringBuilder(MAX_OPERAND_LEN);
string opcodeStr = formatter.FormatPseudoOp(sDataOpNames.DefineData1);
int maxPerLine = MAX_OPERAND_LEN / 4;
int numChunks = (length + maxPerLine - 1) / maxPerLine;
for (int chunk = 0; chunk < numChunks; chunk++) {
int chunkStart = chunk * maxPerLine;
int chunkEnd = Math.Min((chunk + 1) * maxPerLine, length);
for (int i = chunkStart; i < chunkEnd; i++) {
if (i != chunkStart) {
sb.Append(',');
}
sb.Append(formatter.FormatHexValue(data[offset + i], 2));
}
OutputLine(labelStr, opcodeStr, sb.ToString(), commentStr);
labelStr = commentStr = string.Empty;
sb.Clear();
}
}
/// <summary>
/// Outputs formatted data in an unformatted way, because the code generator couldn't
/// figure out how to do something better.
/// </summary>
private void OutputNoJoy(int offset, int length, string labelStr, string commentStr) {
byte[] data = Project.FileData;
Debug.Assert(length > 0);
Debug.Assert(offset >= 0 && offset < data.Length);
bool singleValue = true;
byte val = data[offset];
for (int i = 1; i < length; i++) {
if (data[offset + i] != val) {
singleValue = false;
break;
}
}
if (singleValue) {
string opcodeStr = SourceFormatter.FormatPseudoOp(sDataOpNames.Fill);
string operandStr = length + "," + SourceFormatter.FormatHexValue(val, 2);
OutputLine(labelStr, opcodeStr, operandStr, commentStr);
} else {
OutputDenseHex(offset, length, labelStr, commentStr);
}
}
// IGenerator
public void OutputEquDirective(string name, string valueStr, string comment) {
OutputLine(name, SourceFormatter.FormatPseudoOp(sDataOpNames.EquDirective),
valueStr, SourceFormatter.FormatEolComment(comment));
}
// IGenerator
public void OutputOrgDirective(int offset, int address) {
// 64tass separates the "compile offset", which determines where the output fits
// into the generated binary, and "program counter", which determines the code
// the assembler generates. Since we need to explicitly specify every byte in
// the output file, the compile offset isn't very useful. We want to set it once
// before the first line of code, then leave it alone.
//
// Any subsequent ORG changes are made to the program counter, and take the form
// of a pair of ops (.logical <addr> to open, .here to end). Omitting the .here
// causes an error.
if (offset == 0) {
// Set the "compile offset" to the initial address.
OutputLine("*", "=", SourceFormatter.FormatHexValue(Project.AddrMap.Get(0), 4),
string.Empty);
} else {
if (mNeedHereOp) {
OutputLine(string.Empty, SourceFormatter.FormatPseudoOp(HERE_PSEUDO_OP),
string.Empty, string.Empty);
}
OutputLine(string.Empty, SourceFormatter.FormatPseudoOp(sDataOpNames.OrgDirective),
SourceFormatter.FormatHexValue(address, 4), string.Empty);
mNeedHereOp = true;
}
}
// IGenerator
public void OutputRegWidthDirective(int offset, int prevM, int prevX, int newM, int newX) {
if (prevM != newM) {
string mop = (newM == 0) ? ".al" : ".as";
OutputLine(string.Empty, SourceFormatter.FormatPseudoOp(mop),
string.Empty, string.Empty);
}
if (prevX != newX) {
string xop = (newX == 0) ? ".xl" : ".xs";
OutputLine(string.Empty, SourceFormatter.FormatPseudoOp(xop),
string.Empty, string.Empty);
}
}
// IGenerator
public void OutputLine(string fullLine) {
mOutStream.WriteLine(fullLine);
}
// IGenerator
public void OutputLine(string label, string opcode, string operand, string comment) {
// Break the line if the label is long and it's not a .EQ directive.
if (!string.IsNullOrEmpty(label) &&
!string.Equals(opcode, sDataOpNames.EquDirective,
StringComparison.InvariantCultureIgnoreCase)) {
if (mLongLabelNewLine && label.Length >= mColumnWidths[0]) {
mOutStream.WriteLine(label);
label = string.Empty;
}
}
mLineBuilder.Clear();
TextUtil.AppendPaddedString(mLineBuilder, label, mColumnWidths[0]);
TextUtil.AppendPaddedString(mLineBuilder, opcode, mColumnWidths[0] + mColumnWidths[1]);
TextUtil.AppendPaddedString(mLineBuilder, operand,
mColumnWidths[0] + mColumnWidths[1] + mColumnWidths[2]);
if (string.IsNullOrEmpty(comment)) {
// Trim trailing spaces off of opcode or operand. If they want trailing
// spaces at the end of a comment, that's fine.
CommonUtil.TextUtil.TrimEnd(mLineBuilder);
} else {
mLineBuilder.Append(comment);
}
mOutStream.WriteLine(mLineBuilder.ToString());
}
private void OutputString(int offset, string labelStr, string commentStr) {
// Normal ASCII strings are handled 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 could probably do something fancy with the character encoding options to
// make high-ASCII work nicely.
//
// We might be able to define a macro for DCI and Reverse.
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);
bool highAscii = false;
int hiddenLeadingBytes = 0;
int shownLeadingBytes = 0;
int trailingBytes = 0;
string opcodeStr;
switch (dfd.FormatType) {
case FormatDescriptor.Type.StringGeneric:
case FormatDescriptor.Type.StringReverse:
case FormatDescriptor.Type.StringDci:
opcodeStr = sDataOpNames.StrGeneric;
highAscii = (data[offset] & 0x80) != 0;
break;
case FormatDescriptor.Type.StringNullTerm:
opcodeStr = sDataOpNames.StrNullTerm;
highAscii = (data[offset] & 0x80) != 0;
trailingBytes = 1;
break;
case FormatDescriptor.Type.StringL8:
opcodeStr = sDataOpNames.StrLen8;
if (dfd.Length > 1) {
highAscii = (data[offset + 1] & 0x80) != 0;
}
hiddenLeadingBytes = 1;
break;
case FormatDescriptor.Type.StringL16:
opcodeStr = sDataOpNames.StrGeneric;
if (dfd.Length > 2) {
highAscii = (data[offset + 2] & 0x80) != 0;
}
shownLeadingBytes = 2;
break;
default:
Debug.Assert(false);
return;
}
if (highAscii) {
OutputNoJoy(offset, dfd.Length, labelStr, commentStr);
return;
}
StringOpFormatter stropf = new StringOpFormatter(SourceFormatter, '"',
StringOpFormatter.RawOutputStyle.CommaSep, MAX_OPERAND_LEN,
CharEncoding.ConvertLowAscii);
// Feed bytes in, skipping over hidden bytes (leading L8, trailing null).
stropf.FeedBytes(data, offset + hiddenLeadingBytes,
dfd.Length - hiddenLeadingBytes - trailingBytes, shownLeadingBytes, false);
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:
case FormatDescriptor.Type.StringDci:
// All good the first time.
break;
case FormatDescriptor.Type.StringNullTerm:
if (stropf.Lines.Count != 1 || stropf.HasEscapedText) {
// Must be single-line without quoted chars.
opcodeStr = sDataOpNames.StrGeneric;
redo = true;
}
break;
case FormatDescriptor.Type.StringL8:
if (stropf.Lines.Count != 1 || stropf.HasEscapedText) {
// Must be single-line without quoted chars.
opcodeStr = sDataOpNames.StrGeneric;
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, false);
}
opcodeStr = formatter.FormatPseudoOp(opcodeStr);
foreach (string str in stropf.Lines) {
OutputLine(labelStr, opcodeStr, str, commentStr);
labelStr = commentStr = string.Empty; // only show on first
}
}
}
#endregion IGenerator
#region IAssembler
/// <summary>
/// Cross-assembler execution interface.
/// </summary>
public class AsmTass64 : IAssembler {
public const string OPTIONS = "--case-sensitive --nostart --long-address -Wall";
// Paths from generator.
private List<string> mPathNames;
// Directory to make current before executing assembler.
private string mWorkDirectory;
// IAssembler
public void GetExeIdentifiers(out string humanName, out string exeName) {
humanName = "64tass Assembler";
exeName = "64tass";
}
// IAssembler
public AssemblerConfig GetDefaultConfig() {
return new AssemblerConfig(string.Empty, new int[] { 8, 8, 11, 73 });
}
// IAssembler
public AssemblerVersion QueryVersion() {
AssemblerConfig config =
AssemblerConfig.GetConfig(AppSettings.Global, AssemblerInfo.Id.Tass64);
if (config == null || string.IsNullOrEmpty(config.ExecutablePath)) {
return null;
}
ShellCommand cmd = new ShellCommand(config.ExecutablePath, "--version",
Directory.GetCurrentDirectory(), null);
cmd.Execute();
if (string.IsNullOrEmpty(cmd.Stdout)) {
return null;
}
// Windows - Stdout: "64tass Turbo Assembler Macro V1.53.1515\r\n"
// Linux - Stdout: "64tass Turbo Assembler Macro V1.53.1515?\n"
const string PREFIX = "Macro V";
string str = cmd.Stdout;
int start = str.IndexOf(PREFIX);
int end = (start < 0) ? -1 : str.IndexOfAny(new char[] { '?', '\r', '\n' }, start + 1);
if (start < 0 || end < 0 || start + PREFIX.Length >= end) {
Debug.WriteLine("Couldn't find version in " + str);
return null;
}
start += PREFIX.Length;
string versionStr = str.Substring(start, end - start);
CommonUtil.Version version = CommonUtil.Version.Parse(versionStr);
if (!version.IsValid) {
return null;
}
return new AssemblerVersion(versionStr, version);
}
// IAssembler
public void Configure(List<string> pathNames, string workDirectory) {
// Clone pathNames, in case the caller decides to modify the original.
mPathNames = new List<string>(pathNames.Count);
foreach (string str in pathNames) {
mPathNames.Add(str);
}
mWorkDirectory = workDirectory;
}
// IAssembler
public AssemblerResults RunAssembler(BackgroundWorker worker) {
// Reduce input file to a partial path if possible. This is really just to make
// what we display to the user a little easier to read.
string pathName = mPathNames[0];
if (pathName.StartsWith(mWorkDirectory)) {
pathName = pathName.Remove(0, mWorkDirectory.Length + 1);
} else {
// Unexpected, but shouldn't be a problem.
Debug.WriteLine("NOTE: source file is not in work directory");
}
AssemblerConfig config =
AssemblerConfig.GetConfig(AppSettings.Global, AssemblerInfo.Id.Tass64);
if (string.IsNullOrEmpty(config.ExecutablePath)) {
Debug.WriteLine("Assembler not configured");
return null;
}
worker.ReportProgress(0, Res.Strings.PROGRESS_ASSEMBLING);
string outFileName = pathName.Substring(0, pathName.Length - 2);
// Wrap pathname in quotes in case it has spaces.
// (Do we need to shell-escape quotes in the pathName?)
ShellCommand cmd = new ShellCommand(config.ExecutablePath,
OPTIONS + " \"" + pathName + "\"" + " -o \"" + outFileName + "\"",
mWorkDirectory, null);
cmd.Execute();
// Can't really do anything with a "cancel" request.
// Output filename is the input filename without the ".S". Since the filename
// was generated by us we can be confident in the format.
string outputFile = mPathNames[0].Substring(0, mPathNames[0].Length - 2);
return new AssemblerResults(cmd.FullCommandLine, cmd.ExitCode, cmd.Stdout,
cmd.Stderr, outputFile);
}
}
#endregion IAssembler
}