1
0
mirror of https://github.com/fadden/6502bench.git synced 2024-06-09 20:29:32 +00:00
6502bench/SourceGen/FormatDescriptor.cs
Andy McFadden 4e5c34f457 Binary includes
This adds a new data format option, "binary include", that takes a
filename operand.  When assembly sources are generated, the section
of file is replaced with an appropriate pseudo-op, and binary files
are generated that hold the file contents.  This is a convenient way
to remove large binary blobs, such as music or sound samples, that
aren't useful to have in text form in the sources.

Partial pathnames are allowed, so you can output a sound blob to
"sounds/blather.bin".  For safety reasons, we don't allow the files
to be created above the project directory, and existing files will
only be overwritten if they have a matching length (so you don't
accidentally stomp on your project file).

The files are not currently shown in the GenAsm dialog, which lets
you see a preview of the generated sources.  The hex dump tool
can do this for the (presumably rare) situations where it's useful.

A new regression test, 20300-binary-include, has been added.  The
pseudo-op name can be overridden on-screen in the settings.

We don't currently do anything new for text/HTML exports.  It might
be useful to generate an optional appendix with a hex dump of the
excised sections.

(issue #144)
2024-05-31 14:22:39 -07:00

675 lines
27 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.Diagnostics;
namespace SourceGen {
/// <summary>
/// Format descriptor for data items and instruction operands. Instances are immutable.
///
/// A list of these is saved as part of the project definition. Code and data that
/// doesn't have one of these will be formatted with default behavior. For data that
/// means a single hexadecimal byte.
///
/// These are referenced from the project and the Anattribs array. Entries in the
/// latter may come from the project (as specified by the user), or be auto-generated
/// by the data analysis pass.
///
/// There may be a large number of these, so try to keep the size down. These are usually
/// stored in lists, not arrays, so declaring as a struct wouldn't help with that.
///
/// IMPORTANT: The stringified names of the enum values are currently serialized into
/// the project file. DO NOT rename members of the enumerations without creating an
/// upgrade path. If new values are added, the project file version number should
/// be incremented.
/// </summary>
public class FormatDescriptor {
/// <summary>
/// General data type. Generally corresponds to the pseudo-opcode.
///
/// The UI only allows big-endian values in certain situations. Internally we want
/// to be orthogonal in case the policy changes.
/// </summary>
public enum Type : byte {
Unknown = 0,
REMOVE, // special type, only used by operand editor
Default, // means "unformatted", same as not having a FormatDescriptor
NumericLE, // 1-4 byte number, little-endian
NumericBE, // 1-4 byte number, big-endian
StringGeneric, // character string
StringReverse, // character string, in reverse order
StringNullTerm, // C-style null-terminated string
StringL8, // string with 8-bit length prefix
StringL16, // string with 16-bit length prefix
StringDci, // string terminated by flipped high bit (Dextral Char Inverted)
Dense, // raw data, represented as compactly as possible
Fill, // fill memory with a value
Uninit, // uninitialized data storage area
Junk, // contents of memory are not interesting
BinaryInclude // file contents will be loaded from external file during asm
}
/// <summary>
/// Additional data type detail. Generally affects the operand.
///
/// Some things are extracted from the data itself, e.g. we don't need to specify
/// what value to use for Fill.
/// </summary>
public enum SubType : byte {
None = 0,
ASCII_GENERIC, // internal place-holder, used when loading older projects
// NumericLE/BE; default is "raw", which can have a context-specific display format
Hex,
Decimal,
Binary,
Address, // wants to be an address, but no symbol defined
Symbol, // symbolic ref; replace with Expression, someday?
// Strings and NumericLE/BE (single character)
Ascii, // ASCII (high bit clear)
HighAscii, // ASCII (high bit set)
C64Petscii, // C64 PETSCII (lower case $41-5a, upper case $c1-da)
C64Screen, // C64 screen code
// Dense; no sub-types
// Fill; no sub-types
// Uninit; no sub-types
// Junk; data may exist for alignment purposes. Sub-type indicates boundary.
// (SubType=None indicates no alignment)
Align2, // must be consecutive ascending powers of 2
Align4,
Align8,
Align16,
Align32,
Align64,
Align128,
Align256,
Align512,
Align1024,
Align2048,
Align4096,
Align8192,
Align16384,
Align32768,
Align65536
}
// Maximum length of a NumericLE/BE item (32-bit value or 4-byte instruction).
public const int MAX_NUMERIC_LEN = 4;
// Create some "stock" descriptors. For simple cases we return one of these
// instead of allocating a new object.
private static FormatDescriptor ONE_DEFAULT = new FormatDescriptor(1,
Type.Default, SubType.None);
private static FormatDescriptor ONE_NONE = new FormatDescriptor(1,
Type.NumericLE, SubType.None);
private static FormatDescriptor ONE_HEX = new FormatDescriptor(1,
Type.NumericLE, SubType.Hex);
private static FormatDescriptor ONE_DECIMAL = new FormatDescriptor(1,
Type.NumericLE, SubType.Decimal);
private static FormatDescriptor ONE_BINARY = new FormatDescriptor(1,
Type.NumericLE, SubType.Binary);
private static FormatDescriptor ONE_LOW_ASCII = new FormatDescriptor(1,
Type.NumericLE, SubType.Ascii);
/// <summary>
/// Length, in bytes, of the data to be formatted.
///
/// For an instruction, this must match what the code analyzer found as the length
/// of the entire instruction, or the descriptor will be ignored.
///
/// For data items, this determines the length of the formatted region.
/// </summary>
public int Length { get; private set; }
/// <summary>
/// Primary format. The actual data must match the format:
/// - Numeric values must be 1-4 bytes.
/// - String values must be ASCII characters with a common high bit (although
/// the start or end may diverge from this based on the sub-type).
/// - Fill areas must contain identical bytes.
/// </summary>
public Type FormatType { get; private set; }
/// <summary>
/// Sub-format specifier. Each primary format has specific sub-formats, but we
/// lump them all together for convenience.
/// </summary>
public SubType FormatSubType { get; private set; }
/// <summary>
/// Symbol reference for Type=Numeric SubType=Symbol. null otherwise.
///
/// Numeric values, such as addresses and constants, can be generated with an
/// expression. Currently we only support using a single symbol, but the goal
/// is to allow free-form expressions like "(sym1+sym2+$80)/3".
///
/// If the symbol exists, the symbol's name will be shown, possibly with an adjustment
/// to make the symbol value match the operand or data item.
///
/// Note this reference has a "part" modifier, so we can use it for e.g. "#>label".
/// </summary>
public WeakSymbolRef SymbolRef { get; private set; }
/// <summary>
/// Optional extra data, used for special cases like BinaryInclude. May be null.
/// </summary>
/// <remarks>
/// It's unfortunate that we have this field for every object, even though very few
/// will actually make use of it. The SymbolRef field has a very specific purpose
/// and shouldn't be used to hold it (asserts and other logic gets upset). Storing
/// the filenames in a separate table has some advantages, but requires integrating
/// changes with the undo/redo mechanism, and the space savings doesn't justify the
/// complexity cost.
/// </remarks>
public string Extra { get; private set; }
// Crude attempt to see how effective the prefab object creation is. Note we create
// these for DefSymbols, so there will be one prefab for every platform symbol entry.
public static int DebugCreateCount { get; private set; }
public static int DebugPrefabCount { get; private set; }
public static void DebugPrefabBump(int adj=1) {
DebugCreateCount += adj;
DebugPrefabCount += adj;
}
/// <summary>
/// Constructor for base type data item.
/// </summary>
/// <param name="length">Length, in bytes.</param>
/// <param name="fmt">Format type.</param>
/// <param name="subFmt">Format sub-type.</param>
private FormatDescriptor(int length, Type fmt, SubType subFmt) {
Debug.Assert(length > 0);
Debug.Assert(length <= MAX_NUMERIC_LEN || !IsNumeric);
Debug.Assert(fmt != Type.Default || length == 1);
Debug.Assert(fmt != Type.BinaryInclude);
Debug.Assert(subFmt == SubType.None || (fmt != Type.Junk) ^ IsJunkSubType(subFmt));
Length = length;
FormatType = fmt;
FormatSubType = subFmt;
}
/// <summary>
/// Constructor for symbol item.
/// </summary>
/// <param name="length">Length, in bytes.</param>
/// <param name="sym">Weak symbol reference.</param>
/// <param name="isBigEndian">Set to true for big-endian data.</param>
private FormatDescriptor(int length, WeakSymbolRef sym, bool isBigEndian) {
Debug.Assert(sym != null);
Debug.Assert(length > 0 && length <= MAX_NUMERIC_LEN);
Length = length;
FormatType = isBigEndian ? Type.NumericBE : Type.NumericLE;
FormatSubType = SubType.Symbol;
SymbolRef = sym;
}
/// <summary>
/// Constructor for item with arbitrary string data.
/// </summary>
/// <param name="length">Length, in bytes.</param>
/// <param name="fmt">Format type.</param>
/// <param name="str">String data.</param>
private FormatDescriptor(int length, Type fmt, string stringData) {
Debug.Assert(length > 0);
Debug.Assert(fmt == Type.BinaryInclude);
Debug.Assert(!string.IsNullOrEmpty(stringData));
Length = length;
FormatType = fmt;
FormatSubType = SubType.None;
Extra = stringData;
}
/// <summary>
/// Returns a descriptor with the requested characteristics. For common cases this
/// returns a pre-allocated object, for less-common cases this allocates a new object.
///
/// Objects are immutable and do not specify a file offset, so they may be re-used
/// by the caller.
/// </summary>
/// <param name="length">Length, in bytes.</param>
/// <param name="fmt">Format type.</param>
/// <param name="subFmt">Format sub-type.</param>
/// <returns>New or pre-allocated descriptor, or null if the arguments are
/// invalid.</returns>
public static FormatDescriptor Create(int length, Type fmt, SubType subFmt) {
if (subFmt != SubType.None && !((fmt != Type.Junk) ^ IsJunkSubType(subFmt))) {
return null;
}
DebugCreateCount++;
DebugPrefabCount++;
if (length == 1) {
if (fmt == Type.Default) {
Debug.Assert(subFmt == SubType.None);
return ONE_DEFAULT;
} else if (fmt == Type.NumericLE) {
switch (subFmt) {
case SubType.None:
return ONE_NONE;
case SubType.Hex:
return ONE_HEX;
case SubType.Decimal:
return ONE_DECIMAL;
case SubType.Binary:
return ONE_BINARY;
case SubType.Ascii:
return ONE_LOW_ASCII;
}
}
}
// For a new file, this will be mostly strings and Fill.
DebugPrefabCount--;
return new FormatDescriptor(length, fmt, subFmt);
}
/// <summary>
/// Returns a descriptor with a symbol.
/// </summary>
/// <param name="length">Length, in bytes.</param>
/// <param name="sym">Weak symbol reference.</param>
/// <param name="isBigEndian">Set to true for big-endian data.</param>
/// <returns>New or pre-allocated descriptor.</returns>
public static FormatDescriptor Create(int length, WeakSymbolRef sym, bool isBigEndian) {
DebugCreateCount++;
return new FormatDescriptor(length, sym, isBigEndian);
}
/// <summary>
/// Returns a descriptor with arbitrary string data.
/// </summary>
/// <param name="length">Length, in bytes.</param>
/// <param name="fmt">Format type.</param>
/// <param name="str">String data.</param>
/// <returns>New or pre-allocated descriptor.</returns>
public static FormatDescriptor Create(int length, Type fmt, string str) {
DebugCreateCount++;
return new FormatDescriptor(length, fmt, str);
}
/// <summary>
/// True if the descriptor is okay to use on an instruction operand. The CPU only
/// understands little-endian numeric values, so that's all we allow.
/// </summary>
public bool IsValidForInstruction {
get {
switch (FormatType) {
case Type.Default:
case Type.NumericLE:
return true;
//case Type.NumericBE:
default:
return false;
}
}
}
/// <summary>
/// True if the FormatDescriptor has a symbol.
/// </summary>
public bool HasSymbol {
get {
Debug.Assert(SymbolRef == null || (IsNumeric && FormatSubType == SubType.Symbol));
return SymbolRef != null;
}
}
/// <summary>
/// True if the FormatDescriptor is a numeric type (NumericLE or NumericBE).
/// </summary>
public bool IsNumeric {
get {
return FormatType == Type.NumericLE || FormatType == Type.NumericBE;
}
}
/// <summary>
/// True if the FormatDescriptor is a string type.
/// </summary>
public bool IsString {
get {
switch (FormatType) {
case Type.StringGeneric:
case Type.StringReverse:
case Type.StringNullTerm:
case Type.StringL8:
case Type.StringL16:
case Type.StringDci:
return true;
default:
return false;
}
}
}
/// <summary>
/// True if the FormatDescriptor is a string or character.
/// </summary>
public bool IsStringOrCharacter {
get {
switch (FormatSubType) {
case SubType.ASCII_GENERIC:
case SubType.Ascii:
case SubType.HighAscii:
case SubType.C64Petscii:
case SubType.C64Screen:
return true;
default:
return false;
}
}
}
/// <summary>
/// True if the FormatDescriptor has a symbol or is Numeric/Address.
/// </summary>
public bool HasSymbolOrAddress {
// Derived from other fields, so you can ignore this in equality tests. This is
// of interest to undo/redo, since changing a symbol reference can affect data scan.
get {
return HasSymbol || FormatSubType == SubType.Address;
}
}
/// <summary>
/// Numeric base specific by format/sub-format. Returns 16 when uncertain.
/// </summary>
public int NumBase {
get {
if (FormatType != Type.NumericLE && FormatType != Type.NumericBE) {
Debug.Assert(false);
return 16;
}
switch (FormatSubType) {
case SubType.None:
case SubType.Hex:
return 16;
case SubType.Decimal:
return 10;
case SubType.Binary:
return 2;
default:
Debug.Assert(false);
return 16;
}
}
}
/// <summary>
/// Returns the FormatSubType enum constant for the specified numeric base.
/// </summary>
/// <param name="numBase">Base (2, 10, or 16).</param>
/// <returns>Enum value.</returns>
public static SubType GetSubTypeForBase(int numBase) {
switch (numBase) {
case 2: return SubType.Binary;
case 10: return SubType.Decimal;
case 16: return SubType.Hex;
default:
Debug.Assert(false);
return SubType.Hex;
}
}
/// <summary>
/// Returns true if the sub-type is exclusively for use with the Junk type. Notably,
/// returns false for SubType.None.
/// </summary>
public bool IsAlignedJunk {
get { return IsJunkSubType(FormatSubType); }
}
private static bool IsJunkSubType(SubType subType) {
return ((int)subType >= (int)SubType.Align2 &&
(int)subType <= (int)SubType.Align65536);
}
/// <summary>
/// Converts a power of 2 value to the corresponding alignment sub-type.
/// </summary>
/// <param name="pwr">Power of 2.</param>
/// <returns>The matching sub-type, or None if nothing matches.</returns>
public static SubType PowerToAlignment(int pwr) {
if (pwr < 1 || pwr > 16) {
return SubType.None;
}
// pwr==1 --> 2^1 --> Align2
return (SubType)((int)SubType.Align2 - 1 + pwr);
}
/// <summary>
/// Converts an alignment sub-type to the corresponding power of 2.
/// </summary>
/// <param name="align">Alignment value.</param>
/// <returns>The matching power of 2, or -1 if the sub-type isn't valid.</returns>
public static int AlignmentToPower(SubType align) {
Debug.Assert(IsJunkSubType(align));
if ((int)align < (int)SubType.Align2 || (int)align > (int)SubType.Align65536) {
return -1;
}
return (int)align - (int)SubType.Align2 + 1;
}
/// <summary>
/// Generates a string describing the format, suitable for use in the UI.
/// </summary>
public string ToUiString(bool showBytes = true) {
// NOTE: this should be made easier to localize
string retstr = string.Empty;
if (showBytes) {
retstr = Length + "-byte ";
}
if (IsString) {
switch (FormatSubType) {
case SubType.Ascii:
retstr += "ASCII";
break;
case SubType.HighAscii:
retstr += "ASCII (high)";
break;
case SubType.C64Petscii:
retstr += "C64 PETSCII";
break;
case SubType.C64Screen:
retstr += "C64 Screen";
break;
default:
retstr += "???";
break;
}
switch (FormatType) {
case Type.StringGeneric:
retstr += " string";
break;
case Type.StringReverse:
retstr += " string (reverse)";
break;
case Type.StringNullTerm:
retstr += " string (null term)";
break;
case Type.StringL8:
retstr += " string (1-byte len)";
break;
case Type.StringL16:
retstr += " string (2-byte len)";
break;
case Type.StringDci:
retstr += " string (DCI)";
break;
default:
retstr += " ???";
break;
}
return retstr;
}
switch (FormatSubType) {
case SubType.None:
switch (FormatType) {
case Type.Default:
case Type.NumericLE:
retstr += "numeric (little-endian)";
break;
case Type.NumericBE:
retstr += "numeric (big-endian)";
break;
case Type.Dense:
retstr += "dense";
break;
case Type.Fill:
retstr += "fill";
break;
case Type.Uninit:
retstr += "uninitialized data";
break;
case Type.Junk:
retstr += "unaligned junk";
break;
case Type.BinaryInclude:
retstr += "binary include";
break;
default:
// strings handled earlier
retstr += "???";
break;
}
break;
case SubType.Hex:
retstr += "numeric, Hex";
break;
case SubType.Decimal:
retstr += "numeric, Decimal";
break;
case SubType.Binary:
retstr += "numeric, Binary";
break;
case SubType.Address:
retstr += "address";
break;
case SubType.Symbol:
if (SymbolRef.IsVariable) {
retstr += "local var \"" + SymbolRef.Label + "\"";
} else {
retstr += "symbol \"" + SymbolRef.Label + "\"";
}
break;
case SubType.Ascii:
retstr += "numeric, ASCII";
break;
case SubType.HighAscii:
retstr += "numeric, ASCII (high)";
break;
case SubType.C64Petscii:
retstr += "numeric, C64 PETSCII";
break;
case SubType.C64Screen:
retstr += "numeric, C64 Screen";
break;
case SubType.Align2:
case SubType.Align4:
case SubType.Align8:
case SubType.Align16:
case SubType.Align32:
case SubType.Align64:
case SubType.Align128:
case SubType.Align256:
case SubType.Align512:
case SubType.Align1024:
case SubType.Align2048:
case SubType.Align4096:
case SubType.Align8192:
case SubType.Align16384:
case SubType.Align32768:
case SubType.Align65536:
retstr += "alignment to " + (1 << AlignmentToPower(FormatSubType));
break;
default:
retstr += "???";
break;
}
return retstr;
}
public override string ToString() {
return "[FmtDesc: len=" + Length + " fmt=" + FormatType + " sub=" + FormatSubType +
" sym=" + SymbolRef + " xtra=" + Extra + "]";
}
public static bool operator ==(FormatDescriptor a, FormatDescriptor b) {
if (ReferenceEquals(a, b)) {
return true; // same object, or both null
}
if (ReferenceEquals(a, null) || ReferenceEquals(b, null)) {
return false; // one is null
}
return a.Length == b.Length && a.FormatType == b.FormatType &&
a.FormatSubType == b.FormatSubType && a.SymbolRef == b.SymbolRef &&
a.Extra == b.Extra;
}
public static bool operator !=(FormatDescriptor a, FormatDescriptor b) {
return !(a == b);
}
public override bool Equals(object obj) {
return obj is FormatDescriptor && this == (FormatDescriptor)obj;
}
public override int GetHashCode() {
int hashCode = 0;
if (SymbolRef != null) {
hashCode = SymbolRef.GetHashCode();
}
hashCode ^= Length;
hashCode ^= (int)FormatType;
hashCode ^= (int)FormatSubType;
if (Extra != null) {
hashCode ^= Extra.GetHashCode();
}
return hashCode;
}
/// <summary>
/// Debugging utility function to dump a sorted list of objects.
/// </summary>
public static void DebugDumpSortedList(SortedList<int, FormatDescriptor> list) {
if (list == null) {
Debug.WriteLine("FormatDescriptor list is empty");
return;
}
Debug.WriteLine("FormatDescriptor list (" + list.Count + " entries)");
foreach (KeyValuePair<int, FormatDescriptor> kvp in list) {
int offset = kvp.Key;
FormatDescriptor dfd = kvp.Value;
Debug.WriteLine(" +" + offset.ToString("x6") + ",+" +
(offset + dfd.Length - 1).ToString("x6") + ": " + dfd.FormatType +
"(" + dfd.FormatSubType + ")");
}
}
}
}