fix(ui): systemic escape normalization at the string source

The exit-world confirmation (ID_Client_EndCharacterSessionConfirm, table
0x23000001 key 0x0EB1C41D) rendered its literal two-character "\n" escapes
because escape decoding lived in individual consumers — Batch E centralized
it for authored captions only (DatWidgetFactory.ResolveAuthoredString), and
each new string surface had to remember its own copy. The installed DAT
carries the escape in 4,365 of 7,050 strings; per-consumer normalization
was structurally guaranteed to keep leaking.

Retail's placement is the SOURCE, not the widget: every public StringInfo
resolution ends in StringTableMetaLanguage::UnescapeString @ 0x0067BDC0
(StringInfo::InqString @ 0x0042E490, GetLiteralValue @ 0x0042CA50), the
write side escapes (SetLiteralValue @ 0x0042C980; AddVariable_String
@ 0x0042E6C0 for template variables), and widgets receive decoded text.
Ported exactly:

- NEW RetailStringEscapes: UnescapeString/EscapeString + the
  GetUnEscapedChar @ 0x0067B750 / GetEscapedChar @ 0x0067B6C0 tables
  (\n \t \r \q + the ten metalanguage self-escapes []!{}#\|^$,
  byte-verified against the PDB-paired 2013 binary at 0x3FE178;
  unrecognized pairs stay verbatim).
- DatStringResolver.Resolve/ResolveAll unescape at the source;
  ResolveTemplate escapes each variable on insert and unescapes the
  composed whole — retail's round trip, so variable content (player
  names) can never be corrupted by the final decode.
- RETIRED the consumer copies (double paths would corrupt an authored
  "\n" into a line break): DatWidgetFactory.NormalizeEscapes + BuildText's
  inline replace, RetailUiRuntime.NormalizeRetailNewlines + the
  OpenCaptureInstructions inline replace, DatRichText.Compose's replace,
  IndicatorDetailText.Shape's replace. ItemAppraisalTextLayout's replace
  stays — WIRE-domain (server strings never pass the DAT source; retail's
  ItemExamineUI::AddItemInfo @ 0x004AC050 appends wire text verbatim), now
  documented as such.
- Consumer CR-strips retired with them: the installed DATs contain ZERO
  real CR characters (sweep-measured) and UiText.WrapWords already drops
  strays.

Tests: RetailStringEscapes conformance (escape set, unknown pairs,
round trip), DatStringResolver source-decode pins (including the exact
user-reported exit-world text shape and a backslash-carrying variable),
the installed-DAT escape sweep (7,050 strings; every resolution must equal
the retail unescape of the raw entry; inventory printed), and the existing
caption/rich-text/live-DAT pins relocated to the source contract.

App 5550/3 (live-DAT), Runtime 1747/0, complete Release solution green
across all suites.

Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>
This commit is contained in:
Erik 2026-08-17 13:26:25 +02:00
parent fdc4fd496d
commit 967b9c57cf
15 changed files with 605 additions and 116 deletions

View file

@ -19,17 +19,16 @@ namespace AcDream.App.UI.Layout;
/// <para>
/// The description pages used to bypass this entirely: they assigned a raw
/// <c>LinesProvider</c> lambda returning ONE unwrapped <see cref="AcDream.App.UI.UiText.Line"/>
/// per composed string, with no escape-normalize and no word-wrap. Two
/// concrete symptoms this caused: literal two-character <c>"\n"</c>
/// escapes rendered as backslash-n instead of a real line break (the DAT
/// stores that literal escape — <c>DatWidgetFactory.BuildText</c>'s own
/// authored-string path already normalizes it for single-element authored
/// captions; this helper reproduces the SAME normalize for
/// runtime-composed multi-segment text), and — for the Town page
/// specifically — an unwrapped single line meant the town-specific SUFFIX
/// of the composed string rendered far outside the box's clipped viewport,
/// so switching towns looked like "the text never changes" even though the
/// underlying string genuinely did (only its INVISIBLE tail differed).
/// per composed string, with no word-wrap. Historical symptom (Batch C):
/// for the Town page an unwrapped single line meant the town-specific
/// SUFFIX of the composed string rendered far outside the box's clipped
/// viewport, so switching towns looked like "the text never changes" even
/// though the underlying string genuinely did (only its INVISIBLE tail
/// differed). Escape decoding (the DAT's literal two-character <c>"\n"</c>,
/// Batch C's other symptom) has since moved to the string source
/// (<see cref="DatStringResolver"/> → <see cref="RetailStringEscapes"/>,
/// the 2026-08-17 systemic round) — segments reach this composer with real
/// line breaks already in place.
/// </para>
/// </summary>
internal static class DatRichText
@ -41,7 +40,7 @@ internal static class DatRichText
public readonly record struct Segment(string? Text, Vector4 Color);
/// <summary>
/// Escape-normalizes and word-wraps every segment (independently, so
/// Word-wraps every segment (independently, so
/// each segment's wrapped lines keep ITS OWN color), then concatenates
/// the results in order. No separator is inserted between segments —
/// retail's own composition calls concatenate directly
@ -76,15 +75,12 @@ internal static class DatRichText
if (string.IsNullOrEmpty(segment.Text))
continue;
// The installed DAT stores the LITERAL two-character escape
// "\n" (0x5C 0x6E), not a real line break — same normalize
// DatWidgetFactory.BuildText's authored-string path already
// applies for single-element authored captions.
string normalized = segment.Text
.Replace("\\n", "\n")
.Replace("\r", string.Empty);
foreach (string wrapped in UiText.WrapWords(normalized, measure, maximumWidth))
// Escape decoding (the DAT's literal two-character "\n") happens
// at the string source (DatStringResolver → RetailStringEscapes,
// 2026-08-17 systemic round — retail's own placement), so
// segments arrive with real line breaks; WrapWords preserves
// them and drops any stray CR itself.
foreach (string wrapped in UiText.WrapWords(segment.Text, measure, maximumWidth))
lines.Add(new UiText.Line(wrapped, segment.Color));
}

View file

@ -9,9 +9,23 @@ namespace AcDream.App.UI.Layout;
/// The caller owns synchronization around <see cref="DatCollection"/> reads.
/// </summary>
/// <remarks>
/// <para>
/// Retail reference: <c>StringInfo::GetString</c> and
/// <c>compute_str_hash @ 0x00413110</c>. A StringInfo's token selects one
/// localized string variant; ordinary UI labels use token zero.
/// </para>
/// <para>
/// Every resolution decodes the DAT's two-character escapes
/// (<c>\n</c>, <c>\t</c>, <c>\r</c>, <c>\q</c>, and the metalanguage
/// self-escapes) HERE, at the source — retail's own placement: every public
/// <c>StringInfo</c> resolution ends in
/// <c>StringTableMetaLanguage::UnescapeString @ 0x0067BDC0</c>
/// (<c>StringInfo::InqString @ 0x0042E490</c>,
/// <c>StringInfo::GetLiteralValue @ 0x0042CA50</c>). Consumers receive
/// already-decoded text and must not re-decode — see
/// <see cref="RetailStringEscapes"/>' remarks for the double-decode hazard
/// (the 2026-08-17 systemic round that retired the per-consumer copies).
/// </para>
/// </remarks>
public sealed class DatStringResolver
{
@ -41,7 +55,9 @@ public sealed class DatStringResolver
return null;
int index = token >= 0 && token < entry.Strings.Count ? token : 0;
return entry.Strings[index].Value;
// StringInfo::InqString @ 0x0042E490's unconditional tail: the stored
// string is escaped; the resolved string is decoded.
return RetailStringEscapes.Unescape(entry.Strings[index].Value);
}
/// <summary>Returns every literal token for one retail StringInfo entry.</summary>
@ -57,7 +73,9 @@ public sealed class DatStringResolver
return table is not null
&& table.Strings.TryGetValue(stringId, out var entry)
&& entry.Strings.Count != 0
? entry.Strings.Select(value => value.Value).ToArray()
? entry.Strings
.Select(value => RetailStringEscapes.Unescape(value.Value))
.ToArray()
: null;
}
@ -105,14 +123,21 @@ public sealed class DatStringResolver
{
composed.Append(entry.Strings[i].Value);
// Variables are stored as the pre-computed name hashes (the same
// compute_str_hash space PlayerVariable lives in).
// compute_str_hash space PlayerVariable lives in). Each value is
// escaped on insert — retail's AddVariable_String @ 0x0042E6C0
// stores every variable through SetLiteralValue(escape=1)
// @ 0x0042C980 → EscapeString — so the final whole-string
// unescape below returns variable content verbatim while
// decoding the authored fragments' escapes.
if (i < entry.Variables.Count
&& variables.TryGetValue(entry.Variables[i], out string? value))
{
composed.Append(value);
composed.Append(RetailStringEscapes.Escape(value));
}
}
return composed.ToString();
// StringInfo::InqString @ 0x0042E490's unconditional tail, same as
// Resolve above: composed text decodes its escapes at the source.
return RetailStringEscapes.Unescape(composed.ToString());
}
/// <summary>

View file

@ -856,9 +856,11 @@ public static class DatWidgetFactory
{
// 2026-08-13 social gate: authored strings can carry embedded
// newlines (the fellowship empty-state is three sentences over
// '\n's). Gate round 2: the DAT stores the LITERAL two-character
// escape "\n" (0x5C 0x6E — probe-verified: the dump printed
// backslash-n, not a line break), so normalize the escape first.
// '\n's). The DAT stores those as the LITERAL two-character
// escape "\n" (0x5C 0x6E — probe-verified), decoded at the
// string SOURCE since the 2026-08-17 systemic round
// (DatStringResolver → RetailStringEscapes; retail's own
// placement) — `authored` arrives with REAL line breaks here.
// Gate round 3: retail additionally WORD-WRAPS each authored line
// within the element extent (its GlyphList draw — the same wrap
// the confirmation dialog view already uses), so a multiline
@ -868,10 +870,7 @@ public static class DatWidgetFactory
// re-wrapping them is a client-wide behavior change no gate has
// asked for). Providers re-read DefaultColor/width/font per call
// (NOT captured eagerly) so state-driven changes keep tracking.
string normalized = authored
.Replace("\\n", "\n")
.Replace("\r", string.Empty);
if (normalized.Contains('\n'))
if (authored.Contains('\n'))
{
float cachedWidth = float.NaN;
UiDatFont? cachedFont = null;
@ -897,7 +896,7 @@ public static class DatWidgetFactory
? font.MeasureWidth
: static value => value.Length * 8f;
cachedLines = [.. UiText
.WrapWords(normalized, measure, maximumWidth)
.WrapWords(authored, measure, maximumWidth)
.Select(line => new UiText.Line(line, t.DefaultColor))];
}
return cachedLines;
@ -906,7 +905,7 @@ public static class DatWidgetFactory
else
{
t.LinesProvider = () =>
[new UiText.Line(normalized, t.DefaultColor)];
[new UiText.Line(authored, t.DefaultColor)];
}
}
@ -921,7 +920,7 @@ public static class DatWidgetFactory
|| !state.Properties.Values.TryGetValue(0x17u, out var stateCaption)
|| stateCaption.Kind != UiPropertyKind.StringInfo)
continue;
if (NormalizeEscapes(stringResolve?.Invoke(stateCaption.StringInfoValue))
if (stringResolve?.Invoke(stateCaption.StringInfoValue)
is { Length: > 0 } text)
(stateStrings ??= new Dictionary<uint, string>())[stateId] = text;
}
@ -1220,6 +1219,20 @@ public static class DatWidgetFactory
.OrderBy(child => child.ReadOrder)
.ToArray();
/// <summary>
/// Resolves the effective authored caption (dat property <c>0x17</c>)
/// for a widget. Escape decoding is NOT done here: since the 2026-08-17
/// systemic round the string SOURCE (<see cref="DatStringResolver"/> →
/// <see cref="RetailStringEscapes"/>, retail's own placement — every
/// <c>StringInfo</c> resolution ends in
/// <c>StringTableMetaLanguage::UnescapeString @ 0x0067BDC0</c>) hands
/// every consumer already-decoded text. That supersedes R2-2 (Campaign
/// CC gate round 1 Batch E)'s consumer-level normalize, which covered
/// only the P0x17 resolutions in THIS file and missed sibling consumers
/// (the exit-world confirmation dialog, gate round 2) — the exact class
/// of bug source placement closes. Re-decoding here would corrupt an
/// authored <c>\\n</c> (escaped backslash then 'n') into a line break.
/// </summary>
private static string? ResolveAuthoredString(
ElementInfo info,
Func<UiStringInfoValue, string?>? stringResolve)
@ -1228,42 +1241,16 @@ public static class DatWidgetFactory
|| !info.TryGetEffectiveProperty(0x17u, out var property)
|| property.Kind != UiPropertyKind.StringInfo)
return null;
string? resolved = stringResolve(property.StringInfoValue);
// R2-2 (Campaign CC gate round 1 Batch E): the DAT stores the LITERAL
// two-character escape "\n" (0x5C 0x6E), not a real line break — same
// fact BuildText's own authored-string path already normalized for
// (see that call site's own comment). Centralizing the normalize
// HERE, at the single choke point every P0x17 caption resolution in
// this file goes through (BuildText, BuildButton's own caption AND
// its lifted-child caption, BuildButton's coexisting ValueLabel,
// BuildCheckbox), closes the exact class of bug R2-2 found: a caption
// like the Profession credits button's own "Attribute\n Credits"
// rendered the literal backslash-n because BuildButton never
// normalized while BuildText did. BuildText's own subsequent
// Replace("\\n","\n") is now a harmless no-op (idempotent) — left in
// place rather than removed, since it costs nothing and documents the
// same fact locally.
return NormalizeEscapes(resolved);
return stringResolve(property.StringInfoValue);
}
/// <summary>
/// R2-2 (Campaign CC gate round 1 Batch E): the shared escape-normalize
/// <see cref="ResolveAuthoredString"/> applies, pulled out so the
/// per-STATE authored-caption loop below (which resolves a state's own
/// <c>0x17</c> directly, bypassing the effective-property resolution
/// <see cref="ResolveAuthoredString"/> wraps) gets the SAME normalize
/// instead of a second, easily-forgotten copy.
/// </summary>
private static string? NormalizeEscapes(string? raw) =>
raw?.Replace("\\n", "\n").Replace("\r", string.Empty);
/// <summary>
/// #409 (client-wide retail tooltip system): resolves the already-
/// extracted <see cref="ElementInfo.TooltipText"/> (dat property
/// <c>0x49</c>) through <paramref name="stringResolve"/>, applying the
/// SAME escape normalization every other authored <c>StringInfo</c>
/// (captions, <c>0x17</c>) gets at this one choke point. Null when the
/// element authors no tooltip text or no resolver is available.
/// <c>0x49</c>) through <paramref name="stringResolve"/>. Arrives
/// escape-decoded from the string source, like every authored
/// <c>StringInfo</c> (see <see cref="ResolveAuthoredString"/>). Null
/// when the element authors no tooltip text or no resolver is available.
/// </summary>
internal static string? ResolveTooltipText(
ElementInfo info,
@ -1271,6 +1258,6 @@ public static class DatWidgetFactory
{
if (stringResolve is null || info.TooltipText is not { } tooltipText)
return null;
return NormalizeEscapes(stringResolve(tooltipText));
return stringResolve(tooltipText);
}
}

View file

@ -23,8 +23,13 @@ internal static class IndicatorDetailText
?? value.Length * 8f;
var lines = new List<UiText.Line>();
string normalized = text.Replace("\\n", "\n", StringComparison.Ordinal);
foreach (string paragraph in normalized.Split('\n'))
// DAT-resolved bodies (vitae, link status, effects) arrive with
// real line breaks — escapes decode at the string source
// (DatStringResolver → RetailStringEscapes, 2026-08-17 systemic
// round). Wire-sourced text (the appraisal inscription) renders
// verbatim, exactly like retail's ItemExamineUI::AddItemInfo
// @ 0x004AC050 → UIElement_Text::AppendTextWithFont direct append.
foreach (string paragraph in text.Split('\n'))
{
if (paragraph.Length == 0)
{

View file

@ -157,6 +157,14 @@ internal static class ItemAppraisalTextLayout
}
Vector4 color = ResolveColor(target, fragment.Style);
// WIRE-domain normalize — deliberately NOT the DAT source
// decode (DatStringResolver → RetailStringEscapes, 2026-08-17
// systemic round): appraisal fragments are server strings
// (long description, use text, inscription), which never pass
// the DAT string source, so this is not a duplicate path. It
// accommodates literal "\n" sequences in ACE's database
// strings; server strings with REAL line breaks flow through
// the Split below either way.
string normalized = fragment.Text.Replace(
"\\n",
"\n",

View file

@ -0,0 +1,140 @@
using System.Text;
namespace AcDream.App.UI.Layout;
/// <summary>
/// Exact port of retail's string-table escape codec
/// (<c>StringTableMetaLanguage::UnescapeString @ 0x0067BDC0</c> /
/// <c>EscapeString @ 0x0067BBC0</c> and their character tables
/// <c>GetUnEscapedChar @ 0x0067B750</c> / <c>GetEscapedChar @ 0x0067B6C0</c>).
/// </summary>
/// <remarks>
/// <para>
/// PLACEMENT (the systemic 2026-08-17 normalization round): retail decodes
/// escapes at the string SOURCE, not per-widget. Every public
/// <c>StringInfo</c> resolution runs the unescape unconditionally before any
/// consumer sees the text — <c>StringInfo::InqString @ 0x0042E490</c> tail
/// and <c>StringInfo::GetLiteralValue @ 0x0042CA50</c> both end in
/// <c>UnescapeString</c>. The write side is the inverse:
/// <c>StringInfo::SetLiteralValue @ 0x0042C980</c> runs <c>EscapeString</c>
/// when storing plain text (and <c>StringInfo::AddVariable_String
/// @ 0x0042E6C0</c> always stores variables that way), so stored text is
/// escaped, resolved text is decoded, and variable content round-trips
/// verbatim. acdream's equivalent source is <see cref="DatStringResolver"/>;
/// widgets and controllers receive already-decoded strings and must not
/// re-decode (a second pass corrupts an authored <c>\\n</c> — escaped
/// backslash then 'n' — into a line break).
/// </para>
/// <para>
/// The escape set (byte-verified against the PDB-paired 2013 binary; the
/// metalanguage character-set literal at file offset 0x3FE178 is the ten
/// characters <c>[]!{}#\|^$</c>):
/// <c>\n</c> → LF (0x0A), <c>\t</c> → TAB (0x09), <c>\r</c> → CR (0x0D),
/// <c>\q</c> → '"' (0x22), and a backslash before any of the ten
/// metalanguage characters yields that character itself. A backslash before
/// anything else is NOT an escape — retail copies it through verbatim
/// (<c>GetUnEscapedChar</c> returns 0 and <c>UnescapeString</c>'s
/// else-branch keeps the current character).
/// </para>
/// </remarks>
public static class RetailStringEscapes
{
/// <summary>The ten metalanguage-significant characters that escape to
/// themselves. Byte-decoded from the retail binary (see class remarks) —
/// the same literal both character tables test with <c>wcschr</c>.</summary>
private const string MetaCharacters = "[]!{}#\\|^$";
/// <summary>
/// <c>StringTableMetaLanguage::GetUnEscapedChar @ 0x0067B750</c>: the
/// character an escape pair <c>\</c>+<paramref name="value"/> decodes
/// to, or <c>'\0'</c> when the pair is not an escape.
/// </summary>
internal static char GetUnEscapedChar(char value) => value switch
{
'n' => '\n',
'q' => '"',
'r' => '\r',
't' => '\t',
not '\0' when MetaCharacters.Contains(value) => value,
_ => '\0',
};
/// <summary>
/// <c>StringTableMetaLanguage::GetEscapedChar @ 0x0067B6C0</c>: the
/// character that follows the backslash when <paramref name="value"/>
/// must be stored escaped, or <c>'\0'</c> when it is stored verbatim.
/// </summary>
internal static char GetEscapedChar(char value) => value switch
{
'\t' => 't',
'\n' => 'n',
'\r' => 'r',
'"' => 'q',
not '\0' when MetaCharacters.Contains(value) => value,
_ => '\0',
};
/// <summary>
/// <c>StringTableMetaLanguage::UnescapeString @ 0x0067BDC0</c>: decodes
/// every two-character escape pair; all other characters (including a
/// backslash that does not start a recognized pair, and a trailing
/// backslash) copy through verbatim.
/// </summary>
public static string Unescape(string value)
{
ArgumentNullException.ThrowIfNull(value);
// Fast path: a string with no backslash cannot contain an escape.
int first = value.IndexOf('\\');
if (first < 0)
return value;
var result = new StringBuilder(value.Length);
for (int i = 0; i < value.Length; i++)
{
char current = value[i];
// Retail reads the character AFTER the candidate backslash (the
// terminator — never an escape — when at the end of the buffer).
char next = i + 1 < value.Length ? value[i + 1] : '\0';
char unescaped = GetUnEscapedChar(next);
if (current == '\\' && unescaped != '\0')
{
result.Append(unescaped);
i++; // consume the pair
}
else if (current != '\0')
{
result.Append(current);
}
}
return result.ToString();
}
/// <summary>
/// <c>StringTableMetaLanguage::EscapeString @ 0x0067BBC0</c>: the exact
/// inverse — every character with a <see cref="GetEscapedChar"/> mapping
/// is stored as <c>\</c> + that mapping; everything else verbatim.
/// <c>Unescape(Escape(x)) == x</c> for every <paramref name="value"/> —
/// the round-trip retail relies on for template variables.
/// </summary>
public static string Escape(string value)
{
ArgumentNullException.ThrowIfNull(value);
StringBuilder? result = null;
for (int i = 0; i < value.Length; i++)
{
char current = value[i];
char escaped = GetEscapedChar(current);
if (escaped != '\0')
{
result ??= new StringBuilder(value.Length + 4)
.Append(value, 0, i);
result.Append('\\').Append(escaped);
}
else if (current != '\0')
{
result?.Append(current);
}
}
return result?.ToString() ?? value;
}
}

View file

@ -2973,10 +2973,10 @@ public sealed class RetailUiRuntime : IDisposable
});
if (text is null) return 0u; // no invented English
// The authored text stores its blank line as a literal
// "\n\n" two-character escape (live-probed) — same
// convention DatWidgetFactory/IndicatorDetailText already
// unescape for other DAT-authored strings.
text = text.Replace("\\n", "\n", StringComparison.Ordinal);
// "\n\n" two-character escape (live-probed), decoded at
// the string source (DatStringResolver →
// RetailStringEscapes, 2026-08-17 systemic round) —
// `text` arrives with real line breaks.
try
{
return DialogFactory.MakeWait(text, queueKey: 0x10000001u);
@ -4331,13 +4331,13 @@ public sealed class RetailUiRuntime : IDisposable
{
lock (_bindings.Assets.DatLock)
{
return NormalizeRetailNewlines(strings.ResolveTemplate(
stringTableId,
"ID_CharacterManagement_DeleteCharacterConfirmation",
new Dictionary<uint, string>
{
[DatStringResolver.PlayerVariable] = characterName,
})!);
return strings.ResolveTemplate(
stringTableId,
"ID_CharacterManagement_DeleteCharacterConfirmation",
new Dictionary<uint, string>
{
[DatStringResolver.PlayerVariable] = characterName,
})!;
}
}
@ -4353,16 +4353,16 @@ public sealed class RetailUiRuntime : IDisposable
confirmExit));
}
// Escape decoding (the DAT's literal two-character "\n" and friends)
// happens at the string source since the 2026-08-17 systemic round —
// DatStringResolver → RetailStringEscapes, retail's own placement — so
// this is a plain key-hash resolve. The former NormalizeRetailNewlines
// consumer copy is retired (double-decoding corrupts an authored "\\n").
private static string? ResolveCharacterManagementString(
DatStringResolver strings,
uint tableId,
string key) =>
strings.Resolve(tableId, DatStringResolver.ComputeHash(key)) is { } value
? NormalizeRetailNewlines(value)
: null;
private static string NormalizeRetailNewlines(string value) =>
value.Replace("\\n", "\n", StringComparison.Ordinal);
strings.Resolve(tableId, DatStringResolver.ComputeHash(key));
private void ConfigureCharacterCreation()
{

View file

@ -1904,18 +1904,21 @@ public sealed class CharacterCreationUiControllerTests
/// <summary>
/// GF-2: the composed description routes through the shared rich-text
/// helper — header segments (palette index 1) render in a DIFFERENT
/// color than body segments (index 0), and each segment's own escape
/// sequence is normalized. The fixture's description element carries
/// no authored <c>FontColorPalette</c>, so this also exercises
/// <see cref="DatRichText.PaletteColor"/>'s fallback (green header /
/// white body).
/// color than body segments (index 0), and an authored line break
/// splits into stacked lines. The harness resolver models
/// <c>DatStringResolver</c>'s post-decode output (the DAT's literal
/// "\n" escape decodes AT THE SOURCE since the 2026-08-17 systemic
/// round), so the fixture feeds a REAL '\n'. The fixture's description
/// element carries no authored <c>FontColorPalette</c>, so this also
/// exercises <see cref="DatRichText.PaletteColor"/>'s fallback (green
/// header / white body).
/// </summary>
[Fact]
public void HeritageDescription_ComposesGreenHeaderAndWhiteBodySegments()
{
using var environment = new EnvironmentHarness();
environment.Runtime.ResolvedStrings["ID_CharGen_Heritage_StartingSkills_Header"] = "Trained Starting Skills:";
environment.Runtime.ResolvedStrings["ID_CharGen_Heritage_StartingSkills"] = "Line one\\nLine two";
environment.Runtime.ResolvedStrings["ID_CharGen_Heritage_StartingSkills"] = "Line one\nLine two";
environment.Controller.Open();
environment.Runtime.SelectHeritageDirect(AluvianId);
BumpRevisionAndTick(environment);
@ -1924,7 +1927,7 @@ public sealed class CharacterCreationUiControllerTests
var lines = description.LinesProvider().ToList();
Assert.Contains(lines, l => l.Text == "Trained Starting Skills:" && l.Color == new Vector4(0f, 1f, 0f, 1f));
// The literal "\n" escape in the body segment must become TWO
// The source-decoded line break in the body segment must become TWO
// separate lines, not render as a literal backslash-n.
Assert.Contains(lines, l => l.Text == "Line one" && l.Color == Vector4.One);
Assert.Contains(lines, l => l.Text == "Line two" && l.Color == Vector4.One);

View file

@ -138,10 +138,12 @@ public sealed class CharacterManagementLiveDatTests
// Finding 1: MakeConfirmExitDialog@0x004ed250's text
// (compute_str_hash("ID_CharacterManagement_ConfirmExit"), table
// enum 0x10000002 -> 0x23000002). The raw DAT string carries a
// literal two-character "\n" escape (this test's Resolve() helper
// does not normalize it — RetailUiRuntime does, via
// NormalizeRetailNewlines, before handing it to the controller).
Assert.Equal("Are you sure you want to leave?\\n", Resolve(strings, table,
// literal two-character "\n" escape; since the 2026-08-17 systemic
// round DatStringResolver decodes it AT THE SOURCE (retail's own
// placement — StringInfo::InqString @ 0x0042E490's UnescapeString
// tail), so Resolve returns a REAL line break and no consumer
// normalizes again.
Assert.Equal("Are you sure you want to leave?\n", Resolve(strings, table,
"ID_CharacterManagement_ConfirmExit"));
string confirmation = Assert.IsType<string>(strings.ResolveTemplate(
table,

View file

@ -5,8 +5,10 @@ namespace AcDream.App.Tests.UI.Layout;
/// <summary>
/// Campaign CC gate round 1 Batch C: unit tests for the shared
/// escape-normalize + word-wrap + per-segment-color helper feeding
/// GF-2/GF-3/GF-11a and the Summary how-to text (Commit 3).
/// word-wrap + per-segment-color helper feeding GF-2/GF-3/GF-11a and the
/// Summary how-to text (Commit 3). Escape decoding moved to the string
/// source in the 2026-08-17 systemic round (DatStringResolver →
/// RetailStringEscapes) — segments reach Compose with real line breaks.
/// </summary>
public class DatRichTextTests
{
@ -16,17 +18,24 @@ public class DatRichTextTests
private static UiText MakeTarget(float width) =>
new() { Width = width, Height = 200f };
/// <summary>Segments arrive source-decoded (real '\n'); Compose keeps
/// the authored break as a line split. A literal backslash-n pair in a
/// segment must stay VERBATIM — re-decoding here is the double-decode
/// hazard the 2026-08-17 round retired.</summary>
[Fact]
public void Compose_NormalizesLiteralBackslashNEscape()
public void Compose_SplitsOnRealNewlines_AndKeepsLiteralPairsVerbatim()
{
UiText target = MakeTarget(1000f); // wide enough that nothing wraps
var segments = new[] { new DatRichText.Segment("line one\\nline two", White) };
var segments = new[]
{
new DatRichText.Segment("line one\nliteral \\n stays", White),
};
var lines = DatRichText.Compose(target, segments);
Assert.Equal(2, lines.Count);
Assert.Equal("line one", lines[0].Text);
Assert.Equal("line two", lines[1].Text);
Assert.Equal("literal \\n stays", lines[1].Text);
}
[Fact]

View file

@ -0,0 +1,136 @@
using System.IO;
using System.Text;
using AcDream.App.UI.Layout;
using AcDream.Content;
using DatReaderWriter.Options;
using StringTable = DatReaderWriter.DBObjs.StringTable;
namespace AcDream.App.Tests.UI.Layout;
/// <summary>
/// 2026-08-17 systemic escape round: installed-DAT sweep of EVERY string
/// table for literal escape content, plus the source-normalization contract
/// (<see cref="DatStringResolver"/> resolutions must equal
/// <see cref="RetailStringEscapes.Unescape"/> of the raw stored text —
/// retail's <c>StringInfo::InqString @ 0x0042E490</c> placement). This is
/// the measurement companion to the per-consumer normalize retirement: it
/// proves the escape class genuinely exists in shipping data and prints
/// which tables carry it.
/// </summary>
public sealed class DatStringEscapeSweepTests
{
[InstalledDatFact]
public void EveryInstalledStringResolvesSourceDecoded()
{
string datDirectory = Environment.GetEnvironmentVariable("ACDREAM_DAT_DIR")
?? Path.Combine(
Environment.GetFolderPath(Environment.SpecialFolder.UserProfile),
"Documents",
"Asheron's Call");
using var dats = new DatCollection(datDirectory, DatAccessType.Read);
var resolver = new DatStringResolver(dats);
int tables = 0;
int strings = 0;
int withNewlineEscape = 0;
int withTabEscape = 0;
int withCrEscape = 0;
int withQuoteEscape = 0;
int withMetaSelfEscape = 0;
int withUnknownPair = 0;
int withRealCr = 0;
var perTableNewlines = new SortedDictionary<uint, int>();
var examples = new List<string>();
string? userReportedExitText = null;
foreach (uint tableId in dats.GetAllIdsOfType<StringTable>().Order())
{
StringTable? table = dats.Get<StringTable>(tableId);
if (table is null)
continue;
tables++;
foreach ((uint stringId, var entry) in table.Strings)
{
for (int token = 0; token < entry.Strings.Count; token++)
{
string raw = entry.Strings[token].Value;
strings++;
bool newline = false, unknown = false, meta = false;
for (int i = 0; i < raw.Length - 1; i++)
{
if (raw[i] != '\\')
continue;
char next = raw[i + 1];
char decoded = RetailStringEscapes.GetUnEscapedChar(next);
switch (decoded)
{
case '\n': newline = true; break;
case '\t': withTabEscape++; break;
case '\r': withCrEscape++; break;
case '"': withQuoteEscape++; break;
case '\0': unknown = true; break;
default: meta = true; break;
}
i++; // the pair is consumed either way it decodes
}
if (newline)
{
withNewlineEscape++;
perTableNewlines[tableId] =
perTableNewlines.GetValueOrDefault(tableId) + 1;
if (examples.Count < 12)
examples.Add(
$"0x{tableId:X8}/0x{stringId:X8}: \"{Truncate(raw)}\"");
}
if (meta) withMetaSelfEscape++;
if (unknown) withUnknownPair++;
if (raw.Contains('\r')) withRealCr++;
// The source contract: what consumers receive from the
// resolver is EXACTLY the retail unescape of the stored
// text — nothing more (no consumer re-decode is owed),
// nothing less (no escape survives to render literally).
Assert.Equal(
RetailStringEscapes.Unescape(raw),
resolver.Resolve(tableId, stringId, token));
if (raw.Contains("exit your character", StringComparison.OrdinalIgnoreCase))
userReportedExitText =
$"0x{tableId:X8}/0x{stringId:X8}: \"{raw}\"";
}
}
}
var summary = new StringBuilder()
.AppendLine("[escape-sweep] installed-DAT string-table inventory:")
.AppendLine($" tables={tables} strings={strings}")
.AppendLine($" strings with literal \\n escape: {withNewlineEscape}")
.AppendLine($" \\t pairs: {withTabEscape}; \\r pairs: {withCrEscape}; \\q pairs: {withQuoteEscape}")
.AppendLine($" strings with metalanguage self-escapes: {withMetaSelfEscape}")
.AppendLine($" strings with unrecognized backslash pairs (kept verbatim): {withUnknownPair}")
.AppendLine($" strings containing a REAL CR character: {withRealCr}")
.AppendLine(" \\n-escape counts per table: "
+ string.Join(", ", perTableNewlines.Select(
static pair => $"0x{pair.Key:X8}={pair.Value}")))
.AppendLine(" examples:");
foreach (string example in examples)
summary.AppendLine($" {example}");
summary.AppendLine(userReportedExitText is null
? " user-reported exit-world text: NOT found by content scan"
: $" user-reported exit-world text: {userReportedExitText}");
Console.WriteLine(summary.ToString());
// The escape class must genuinely exist in shipping data — if this
// ever goes to zero the sweep (and the source decode) is measuring
// nothing and needs re-examination, not silent success.
Assert.True(
withNewlineEscape > 0,
"expected at least one installed string carrying the literal \\n escape");
}
private static string Truncate(string value) =>
(value.Length <= 90 ? value : value[..90] + "…")
.Replace("\r", "<CR>").Replace("\n", "<LF>");
}

View file

@ -85,6 +85,72 @@ public sealed class DatStringResolverTemplateTests
new Dictionary<uint, string>()));
}
/// <summary>
/// The 2026-08-17 systemic escape round: resolution decodes the DAT's
/// literal two-character escapes AT THE SOURCE — retail's own placement
/// (<c>StringInfo::InqString @ 0x0042E490</c>'s unconditional
/// <c>UnescapeString</c> tail). Consumers receive real line breaks; no
/// per-consumer normalize remains.
/// </summary>
[Fact]
public void ResolveDecodesEscapesAtTheSource()
{
var resolver = MakeResolver(
"ID_Confirm_Exit",
fragments: [
"This will exit your character from the game world.\\n\\nAre you sure?",
],
variables: []);
Assert.Equal(
"This will exit your character from the game world.\n\nAre you sure?",
resolver.Resolve(
TableId, DatStringResolver.ComputeHash("ID_Confirm_Exit")));
}
[Fact]
public void ResolveAllDecodesEveryVariant()
{
var resolver = MakeResolver(
"ID_Variants",
fragments: ["one\\nline", "two\\tcol"],
variables: []);
Assert.Equal(
["one\nline", "two\tcol"],
resolver.ResolveAll(
TableId, DatStringResolver.ComputeHash("ID_Variants")));
}
/// <summary>
/// Template composition decodes the authored fragments' escapes while
/// variable content round-trips VERBATIM — retail escapes each variable
/// on insert (<c>AddVariable_String @ 0x0042E6C0</c> →
/// <c>SetLiteralValue(escape=1) @ 0x0042C980</c>) and unescapes the
/// composed whole once, so a player name containing escape-significant
/// characters can never be corrupted by the final decode.
/// </summary>
[Fact]
public void ResolveTemplateDecodesFragmentsAndKeepsVariablesVerbatim()
{
var resolver = MakeResolver(
"ID_Delete_Confirmation",
fragments: ["Delete ", "?\\nType 'DELETE' to confirm."],
variables: [DatStringResolver.PlayerVariable]);
Assert.Equal(
"Delete Odd\\nName?\nType 'DELETE' to confirm.",
resolver.ResolveTemplate(
TableId,
"ID_Delete_Confirmation",
new Dictionary<uint, string>
{
// A pathological name carrying a REAL backslash before
// an 'n' — must come out verbatim, not as a line break.
[DatStringResolver.PlayerVariable] = "Odd\\nName",
}));
}
[Fact]
public void UnknownKeyResolvesNull()
{

View file

@ -1144,8 +1144,10 @@ public class DatWidgetFactoryTests
}
/// <summary>
/// 2026-08-13 social gate round 3: a MULTILINE authored string (literal
/// backslash-n escapes in the DAT) word-wraps each authored line to the
/// 2026-08-13 social gate round 3: a MULTILINE authored string (a real
/// '\n' — the DAT's literal backslash-n escape decodes at the string
/// source since the 2026-08-17 systemic round, so the resolver seam
/// hands this factory decoded text) word-wraps each authored line to the
/// widget's live width — retail's GlyphList draw, the same wrap the
/// confirmation dialog view uses. The fellowship empty-state was
/// rendering its three authored lines as three clipped runs.
@ -1165,7 +1167,7 @@ public class DatWidgetFactoryTests
// Width=100 fits 12 characters per wrapped line.
var text = Assert.IsType<UiText>(DatWidgetFactory.Create(
info, NoTex, null,
stringResolve: _ => "one two three four five\\nsix"));
stringResolve: _ => "one two three four five\nsix"));
var lines = text.LinesProvider!();
Assert.True(lines.Count >= 3);

View file

@ -0,0 +1,103 @@
using AcDream.App.UI.Layout;
namespace AcDream.App.Tests.UI.Layout;
/// <summary>
/// Conformance for <see cref="RetailStringEscapes"/> — the exact port of
/// retail's string-table escape codec
/// (<c>StringTableMetaLanguage::UnescapeString @ 0x0067BDC0</c> /
/// <c>EscapeString @ 0x0067BBC0</c>, character tables
/// <c>GetUnEscapedChar @ 0x0067B750</c> / <c>GetEscapedChar @ 0x0067B6C0</c>).
/// The metalanguage character set is byte-verified against the PDB-paired
/// 2013 binary (file offset 0x3FE178: <c>[]!{}#\|^$</c>).
/// </summary>
public sealed class RetailStringEscapesTests
{
[Theory]
[InlineData("line one\\nline two", "line one\nline two")]
[InlineData("a\\tb", "a\tb")]
[InlineData("a\\rb", "a\rb")]
[InlineData("say \\qhi\\q", "say \"hi\"")]
public void Unescape_DecodesTheFourCharacterEscapes(
string raw, string expected)
=> Assert.Equal(expected, RetailStringEscapes.Unescape(raw));
[Theory]
[InlineData("\\[", "[")]
[InlineData("\\]", "]")]
[InlineData("\\!", "!")]
[InlineData("\\{", "{")]
[InlineData("\\}", "}")]
[InlineData("\\#", "#")]
[InlineData("\\\\", "\\")]
[InlineData("\\|", "|")]
[InlineData("\\^", "^")]
[InlineData("\\$", "$")]
public void Unescape_DecodesEveryMetalanguageSelfEscape(
string raw, string expected)
=> Assert.Equal(expected, RetailStringEscapes.Unescape(raw));
/// <summary>
/// GetUnEscapedChar returns 0 for anything else — retail keeps the
/// backslash verbatim (UnescapeString's else-branch), including a
/// trailing backslash whose "next" character is the terminator.
/// </summary>
[Theory]
[InlineData("\\z", "\\z")]
[InlineData("C:\\path\\dir", "C:\\path\\dir")]
[InlineData("ends with \\", "ends with \\")]
[InlineData("\\N upper is not an escape", "\\N upper is not an escape")]
public void Unescape_KeepsUnrecognizedPairsVerbatim(
string raw, string expected)
=> Assert.Equal(expected, RetailStringEscapes.Unescape(raw));
/// <summary>
/// The double-decode hazard the 2026-08-17 systemic round exists to
/// close: an authored escaped backslash before an 'n' decodes ONCE to
/// the literal two characters backslash+n — a second decode pass (the
/// retired per-consumer copies) would corrupt it into a line break.
/// </summary>
[Fact]
public void Unescape_EscapedBackslashBeforeN_YieldsLiteralPair()
=> Assert.Equal("\\n", RetailStringEscapes.Unescape("\\\\n"));
[Fact]
public void Unescape_EmptyString_IsEmpty()
=> Assert.Equal(string.Empty, RetailStringEscapes.Unescape(string.Empty));
/// <summary>No backslash → no allocation: the same instance returns.</summary>
[Fact]
public void Unescape_NoEscapes_ReturnsTheSameInstance()
{
const string plain = "Please Wait";
Assert.Same(plain, RetailStringEscapes.Unescape(plain));
}
[Theory]
[InlineData("line one\nline two", "line one\\nline two")]
[InlineData("a\tb", "a\\tb")]
[InlineData("a\rb", "a\\rb")]
[InlineData("say \"hi\"", "say \\qhi\\q")]
[InlineData("[x]", "\\[x\\]")]
[InlineData("back\\slash", "back\\\\slash")]
[InlineData("plain", "plain")]
public void Escape_IsTheStorageInverse(string plain, string expected)
=> Assert.Equal(expected, RetailStringEscapes.Escape(plain));
/// <summary>
/// Retail's template-variable round trip
/// (<c>AddVariable_String @ 0x0042E6C0</c> escapes on insert;
/// <c>InqString @ 0x0042E490</c> unescapes the composed whole): variable
/// content must come out verbatim.
/// </summary>
[Theory]
[InlineData("plain name")]
[InlineData("Odd\\Name")]
[InlineData("multi\nline")]
[InlineData("tabs\tand \"quotes\"")]
[InlineData("[]!{}#\\|^$")]
[InlineData("")]
public void UnescapeOfEscape_RoundTripsVerbatim(string value)
=> Assert.Equal(value, RetailStringEscapes.Unescape(
RetailStringEscapes.Escape(value)));
}

View file

@ -553,16 +553,20 @@ public class UiButtonTests
Assert.True(confinedWidth < button.Width, "the confined width must be narrower than the full button");
}
// ── R2-2 escape-normalize ────────────────────────────────────────────
// ── R2-2 authored caption (source-decoded) ───────────────────────────
/// <summary>
/// R2-2: BuildButton's own P0x17 caption escape-normalizes the same way
/// BuildText's authored-string path always has — the DAT stores the
/// LITERAL two-character escape "\n" (0x5C 0x6E), and the Profession
/// credits button's own authored caption is exactly this shape.
/// R2-2's successor contract (2026-08-17 systemic round): the DAT's
/// LITERAL two-character escape "\n" (0x5C 0x6E — the Profession
/// credits button's own authored caption is exactly this shape) decodes
/// at the string SOURCE (DatStringResolver → RetailStringEscapes,
/// retail's own placement), so the resolver seam hands BuildButton a
/// caption with a REAL line break — and the factory passes it through
/// verbatim, with no second decode that would corrupt an authored
/// backslash pair.
/// </summary>
[Fact]
public void BuildButton_OwnCaption_NormalizesLiteralBackslashNEscape()
public void BuildButton_OwnCaption_PassesSourceDecodedTextThrough()
{
uint stringId = 444u;
var info = new ElementInfo { Type = 1, Width = 150, Height = 50 };
@ -575,11 +579,14 @@ public class UiButtonTests
var button = Assert.IsType<UiButton>(DatWidgetFactory.Create(
info, NoTex, null,
// The raw resolved string carries the LITERAL two characters
// '\' and 'n', matching what the installed DAT actually stores.
stringResolve: value => value.StringId == stringId ? "Attribute\\n Credits" : null));
// The resolver seam models DatStringResolver's post-decode
// output: a REAL '\n', plus a literal backslash pair that a
// stray second decode would corrupt into a line break.
stringResolve: value => value.StringId == stringId
? "Attribute\n Credits \\not-an-escape"
: null));
Assert.Equal("Attribute\n Credits", button.Label);
Assert.Equal("Attribute\n Credits \\not-an-escape", button.Label);
}
private static UiButton ButtonWithStates(params string[] states)