fix(app): Reorder among older items, export and text import that hold up

- Reorder: rows from before an order existed all share one position, and
  one moved among them landed at their end; the run it goes into now
  gets positions of its own (placeAmong), for notes, files, paths,
  steps and sub-studysets. Rows whose deletion can still be undone count
  for nothing, as the list leaves them out, and a dropped row takes its
  place at once instead of snapping back until the order is read again.
- Export: two studysets inside each other (moved so on two devices
  offline) no longer make it run forever; links to files whose names
  hold #, % or brackets work; PDFs, images and sounds go into the zip
  without being compressed again; any error says the export failed.
- Text import: a real CSV reader (quoted fields over several lines, a
  back over several quoted fields, a stray quote spoiling only its own
  line), the separator guessed from the text (semicolons from German
  spreadsheets), and files in UTF-16 or Windows-1252 read as meant.
- A tab whose notes, paths or files couldn't be read says so rather than
  "No … yet"; an Anki import from the Library that fails leaves no empty
  studyset behind.

Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_01MMm7usAo6q9ZZRoPkWjoxi
Signed-off-by: Tizian 「ティツィアーン」 <raisondetredev@pm.me>

halkyon_learn_flutter/lib/l10n/app_de.arb

@@ -262,11 +262,11 @@
   "studysetImportCards": "Karteikarten",
   "studysetImportCardsHint": "Eingefügt oder aus einer CSV- oder TSV-Datei",
   "cardTextImportTitle": "Karteikarten importieren",
-  "cardTextImportDescription": "Eine Karte pro Zeile: ihre Vorderseite, ein Trennzeichen, dann ihre Rückseite. Jede Zeile wird an ihrem ersten Trennzeichen geteilt; Zeilen ohne eines werden ausgelassen.",
+  "cardTextImportDescription": "Eine Karte pro Zeile: ihre Vorderseite, ein Trennzeichen, dann ihre Rückseite. Jede Zeile wird an ihrem ersten Trennzeichen geteilt; Zeilen ohne eines werden ausgelassen. Ein Feld in doppelten Anführungszeichen, wie Tabellenprogramme sie schreiben, darf Trennzeichen und Zeilenumbrüche enthalten.",
   "cardTextImportHint": "Füge die Karten hier ein",
   "cardTextImportLoadFile": "CSV- oder TSV-Datei laden",
   "cardTextImportFileType": "CSV- oder TSV-Dateien",
-  "cardTextImportUnreadable": "Diese Datei ist kein Text in UTF-8.",
+  "cardTextImportUnreadable": "Diese Datei enthält keinen Text.",
   "cardTextImportSeparator": "Trennzeichen",
   "cardTextSeparatorTab": "Tab",
   "cardTextSeparatorComma": "Komma",
@@ -438,6 +438,9 @@
   "syncMismatchMore": "Mehr",
 
   "tagsReadFailed": "Die Tags konnten nicht gelesen werden.",
+  "notesReadFailed": "Die Notizen konnten nicht gelesen werden.",
+  "pathsReadFailed": "Die Lernpfade konnten nicht gelesen werden.",
+  "filesReadFailed": "Die Dateien konnten nicht gelesen werden.",
   "tagPickerTitle": "Tags",
   "tagPickerManage": "Tags verwalten",
   "tagPickerSearchLabel": "Tag suchen oder erstellen",

halkyon_learn_flutter/lib/l10n/app_en.arb

@@ -290,11 +290,11 @@
   "studysetImportCards": "Flashcards",
   "studysetImportCardsHint": "Pasted, or from a CSV or TSV file",
   "cardTextImportTitle": "Import flashcards",
-  "cardTextImportDescription": "One card per line: its front, a separator, then its back. Each line is split at its first separator; lines without one are left out.",
+  "cardTextImportDescription": "One card per line: its front, a separator, then its back. Each line is split at its first separator; lines without one are left out. A field in double quotes, as spreadsheets write them, may hold separators and line breaks.",
   "cardTextImportHint": "Paste the cards here",
   "cardTextImportLoadFile": "Load a CSV or TSV file",
   "cardTextImportFileType": "CSV or TSV files",
-  "cardTextImportUnreadable": "This file isn't text in UTF-8.",
+  "cardTextImportUnreadable": "This file doesn't hold text.",
   "cardTextImportSeparator": "Separator",
   "cardTextSeparatorTab": "Tab",
   "cardTextSeparatorComma": "Comma",
@@ -475,6 +475,9 @@
   "syncMismatchMore": "More",
 
   "tagsReadFailed": "Couldn't read the tags.",
+  "notesReadFailed": "Couldn't read the notes.",
+  "pathsReadFailed": "Couldn't read the learning paths.",
+  "filesReadFailed": "Couldn't read the files.",
   "tagPickerTitle": "Tags",
   "tagPickerManage": "Manage tags",
   "tagPickerSearchLabel": "Search or create a tag",

halkyon_learn_flutter/lib/l10n/app_localizations.dart

@@ -1633,7 +1633,7 @@ abstract class AppLocalizations {
   /// No description provided for @cardTextImportDescription.
   ///
   /// In en, this message translates to:
-  /// **'One card per line: its front, a separator, then its back. Each line is split at its first separator; lines without one are left out.'**
+  /// **'One card per line: its front, a separator, then its back. Each line is split at its first separator; lines without one are left out. A field in double quotes, as spreadsheets write them, may hold separators and line breaks.'**
   String get cardTextImportDescription;
 
   /// No description provided for @cardTextImportHint.
@@ -1657,7 +1657,7 @@ abstract class AppLocalizations {
   /// No description provided for @cardTextImportUnreadable.
   ///
   /// In en, this message translates to:
-  /// **'This file isn\'t text in UTF-8.'**
+  /// **'This file doesn\'t hold text.'**
   String get cardTextImportUnreadable;
 
   /// No description provided for @cardTextImportSeparator.
@@ -2650,6 +2650,24 @@ abstract class AppLocalizations {
   /// **'Couldn\'t read the tags.'**
   String get tagsReadFailed;
 
+  /// No description provided for @notesReadFailed.
+  ///
+  /// In en, this message translates to:
+  /// **'Couldn\'t read the notes.'**
+  String get notesReadFailed;
+
+  /// No description provided for @pathsReadFailed.
+  ///
+  /// In en, this message translates to:
+  /// **'Couldn\'t read the learning paths.'**
+  String get pathsReadFailed;
+
+  /// No description provided for @filesReadFailed.
+  ///
+  /// In en, this message translates to:
+  /// **'Couldn\'t read the files.'**
+  String get filesReadFailed;
+
   /// No description provided for @tagPickerTitle.
   ///
   /// In en, this message translates to:

halkyon_learn_flutter/lib/l10n/app_localizations_de.dart

@@ -970,7 +970,7 @@ class AppLocalizationsDe extends AppLocalizations {
 
   @override
   String get cardTextImportDescription =>
-      'Eine Karte pro Zeile: ihre Vorderseite, ein Trennzeichen, dann ihre Rückseite. Jede Zeile wird an ihrem ersten Trennzeichen geteilt; Zeilen ohne eines werden ausgelassen.';
+      'Eine Karte pro Zeile: ihre Vorderseite, ein Trennzeichen, dann ihre Rückseite. Jede Zeile wird an ihrem ersten Trennzeichen geteilt; Zeilen ohne eines werden ausgelassen. Ein Feld in doppelten Anführungszeichen, wie Tabellenprogramme sie schreiben, darf Trennzeichen und Zeilenumbrüche enthalten.';
 
   @override
   String get cardTextImportHint => 'Füge die Karten hier ein';
@@ -982,7 +982,7 @@ class AppLocalizationsDe extends AppLocalizations {
   String get cardTextImportFileType => 'CSV- oder TSV-Dateien';
 
   @override
-  String get cardTextImportUnreadable => 'Diese Datei ist kein Text in UTF-8.';
+  String get cardTextImportUnreadable => 'Diese Datei enthält keinen Text.';
 
   @override
   String get cardTextImportSeparator => 'Trennzeichen';
@@ -1552,6 +1552,15 @@ class AppLocalizationsDe extends AppLocalizations {
   @override
   String get tagsReadFailed => 'Die Tags konnten nicht gelesen werden.';
 
+  @override
+  String get notesReadFailed => 'Die Notizen konnten nicht gelesen werden.';
+
+  @override
+  String get pathsReadFailed => 'Die Lernpfade konnten nicht gelesen werden.';
+
+  @override
+  String get filesReadFailed => 'Die Dateien konnten nicht gelesen werden.';
+
   @override
   String get tagPickerTitle => 'Tags';
 

halkyon_learn_flutter/lib/l10n/app_localizations_en.dart

@@ -966,7 +966,7 @@ class AppLocalizationsEn extends AppLocalizations {
 
   @override
   String get cardTextImportDescription =>
-      'One card per line: its front, a separator, then its back. Each line is split at its first separator; lines without one are left out.';
+      'One card per line: its front, a separator, then its back. Each line is split at its first separator; lines without one are left out. A field in double quotes, as spreadsheets write them, may hold separators and line breaks.';
 
   @override
   String get cardTextImportHint => 'Paste the cards here';
@@ -978,7 +978,7 @@ class AppLocalizationsEn extends AppLocalizations {
   String get cardTextImportFileType => 'CSV or TSV files';
 
   @override
-  String get cardTextImportUnreadable => 'This file isn\'t text in UTF-8.';
+  String get cardTextImportUnreadable => 'This file doesn\'t hold text.';
 
   @override
   String get cardTextImportSeparator => 'Separator';
@@ -1542,6 +1542,15 @@ class AppLocalizationsEn extends AppLocalizations {
   @override
   String get tagsReadFailed => 'Couldn\'t read the tags.';
 
+  @override
+  String get notesReadFailed => 'Couldn\'t read the notes.';
+
+  @override
+  String get pathsReadFailed => 'Couldn\'t read the learning paths.';
+
+  @override
+  String get filesReadFailed => 'Couldn\'t read the files.';
+
   @override
   String get tagPickerTitle => 'Tags';
 

halkyon_learn_flutter/lib/src/core/sync/fractional_position.dart

@@ -28,3 +28,55 @@ String positionAmong(List<String> others, int index) {
   }
   return FractionalIndexing.generateKeyBetween(lower, upper);
 }
+
+/// Where an item moved to [index] among [others] (the other items in their
+/// order, with their ids) goes: its new position under [movedId], and new
+/// positions for others where it needs them. Items that share a position
+/// (all those from before an order existed stand at `a0`) leave no room
+/// between them, so a run of them around the place gets positions of its
+/// own first, in the order shown, the moved item among them.
+Map<String, String> placeAmong(
+  List<({String id, String position})> others,
+  int index,
+  String movedId,
+) {
+  final at = index.clamp(0, others.length);
+  final lower = at > 0 ? others[at - 1].position : null;
+  final next = at < others.length ? others[at].position : null;
+  if (lower == null || next == null || lower != next) {
+    return {
+      movedId: positionAmong([for (final other in others) other.position], at),
+    };
+  }
+  var start = at - 1;
+  while (start > 0 && others[start - 1].position == lower) {
+    start--;
+  }
+  var end = at;
+  while (end < others.length && others[end].position == lower) {
+    end++;
+  }
+  final keys = FractionalIndexing.generateNKeysBetween(
+    start > 0 ? others[start - 1].position : null,
+    end < others.length ? others[end].position : null,
+    end - start + 1,
+  );
+  return {
+    for (var i = start; i < at; i++) others[i].id: keys[i - start],
+    movedId: keys[at - start],
+    for (var i = at; i < end; i++) others[i].id: keys[i - start + 1],
+  };
+}
+
+/// [items] with the one [isMoved] picks at [toIndex] among the others, as
+/// a list shows a row dropped there before its new position is stored;
+/// as they are when none is picked.
+List<T> movedTo<T>(List<T> items, bool Function(T item) isMoved, int toIndex) {
+  final from = items.indexWhere(isMoved);
+  if (from < 0) {
+    return items;
+  }
+  final moved = [...items];
+  final item = moved.removeAt(from);
+  return moved..insert(toIndex.clamp(0, moved.length), item);
+}

halkyon_learn_flutter/lib/src/feature/anki/bloc/anki_import_bloc.dart

@@ -87,15 +87,16 @@ class AnkiImportBloc(
     if (state case final AnkiImportStateReady ready when package != null) {
       final options = ready.options;
       emit(AnkiImportStateRunning(ready.summary, options, 0));
+      String? made;
       try {
         final result = await _importer.import(
           package,
           studysetId:
               _studysetId ??
-              await _importer.newStudysetFor(
+              (made = await _importer.newStudysetFor(
                 package,
                 fallbackTitle: _fileTitle,
-              ),
+              )),
           mode: options.mode,
           keepProgress: options.keepProgress && ready.summary.hasReviews,
           includeMedia: options.includeMedia,
@@ -106,6 +107,13 @@ class AnkiImportBloc(
         emit(AnkiImportStateDone(result, options.mode));
       } on Exception catch (error) {
         log('Importing an Anki package failed', name: 'Anki', error: error);
+        if (made != null) {
+          try {
+            await _importer.discardIfEmpty(made);
+          } on Object catch (error) {
+            log('Discarding its studyset failed', name: 'Anki', error: error);
+          }
+        }
         emit(const AnkiImportStateFailed(null));
       }
     }

halkyon_learn_flutter/lib/src/feature/anki/data/anki_importer.dart

@@ -69,6 +69,15 @@ class AnkiImporter(
     );
   }
 
+  /// Takes away the studyset [studysetId] made for an import that failed,
+  /// if nothing came into it, so a failure leaves no empty studyset behind.
+  /// One that holds cards stays: what was imported is there to see.
+  Future<void> discardIfEmpty(String studysetId) async {
+    if ((await _cards.cardsOf(studysetId)).isEmpty) {
+      await _studysets.delete(studysetId);
+    }
+  }
+
   /// Imports [package] into the studyset [studysetId] as [mode] says, its
   /// review history if [keepProgress], its files if [includeMedia];
   /// [onProgress] hears how far it got, from 0 to 1.

halkyon_learn_flutter/lib/src/feature/card/model/card_text_import.dart

@@ -16,62 +16,160 @@ enum CardSeparator(final String char) {
 /// A card's front and back, read from a line of text.
 typedef CardLine = ({String front, String back});
 
-/// The cards in [text]: one per line, split at the line's first [separator],
+/// The cards in [text]: one per record, as spreadsheets write them: a line,
-/// so the back may hold the separator too. A field in double quotes, as
+/// unless a field in double quotes runs over several (a doubled quote
-/// spreadsheets write one holding the separator, is read without them (a
+/// inside stands for one). A record's fields are split at [separator]
-/// doubled quote inside standing for one). Blank lines, lines without the
+/// outside quotes; the first is the front, the rest the back, joined by the
-/// separator and lines with an empty side (a header without a back, say)
+/// separator again, so a back may hold one too. Blank records, those
-/// are left out.
+/// without a second field and those with an empty side (a header without a
+/// back, say) are left out, and so is a line whose quote never closes.
 List<CardLine> parseCardLines(String text, CardSeparator separator) => [
-  for (final line in const LineSplitter().convert(
+  for (final fields in _records(_withoutBom(text), separator.char))
-    // Spreadsheets put a byte order mark before UTF-8 text.
+    if (fields.length >= 2)
-    text.startsWith('') ? text.substring(1) : text,
+      if ((
-  ))
+            front: fields.first.trim(),
-    if (_split(line, separator.char) case final card?
+            back: fields.skip(1).join(separator.char).trim(),
-        when card.front.isNotEmpty && card.back.isNotEmpty)
+          )
-      card,
+          case final card when card.front.isNotEmpty && card.back.isNotEmpty)
+        card,
 ];
 
-CardLine? _split(String line, String separator) {
+/// The separator [text] uses, by what most of its first lines hold outside
-  final start = line.length - line.trimLeft().length;
+/// quotes: tabs, semicolons (what spreadsheets write where a comma is the
-  if (!line.startsWith('"', start)) {
+/// decimal mark, in German say) or commas; by [fileName] when they hold
-    final at = line.indexOf(separator);
+/// none ([CardSeparator.forFile]).
-    return at < 0
+CardSeparator guessSeparator(String text, {required String fileName}) {
-        ? null
+  final lines = [
-        : (
+    for (final line in const LineSplitter().convert(_withoutBom(text)))
-            front: line.substring(0, at).trim(),
+      if (line.trim().isNotEmpty) line.replaceAll(RegExp('"[^"]*"'), ''),
-            back: _unquoted(line.substring(at + separator.length).trim()),
+  ].take(20);
-          );
+  var best = CardSeparator.forFile(fileName);
-  }
+  var bestCount = 0;
-  // A quoted front: up to its closing quote, then the separator.
+  for (final separator in CardSeparator.values) {
-  final front = StringBuffer();
+    final count = lines.where((line) => line.contains(separator.char)).length;
-  var i = start + 1;
+    if (count > bestCount) {
-  while (true) {
+      best = separator;
-    if (i >= line.length) {
+      bestCount = count;
-      return null;
-    }
-    if (line[i] != '"') {
-      front.write(line[i++]);
-    } else if (line.startsWith('""', i)) {
-      front.write('"');
-      i += 2;
-    } else {
-      i++;
-      break;
     }
   }
-  final at = line.indexOf(separator, i);
+  return best;
-  if (at < 0 || line.substring(i, at).trim().isNotEmpty) {
-    return null;
-  }
-  return (
-    front: front.toString().trim(),
-    back: _unquoted(line.substring(at + separator.length).trim()),
-  );
 }
 
-/// [field] without the double quotes around it, if it has them.
+/// The text in [bytes] as a spreadsheet or text editor saved it: UTF-8, or
-String _unquoted(String field) =>
+/// UTF-16 with its byte order mark (what Excel saves as "Unicode text"), or
-    field.length >= 2 && field.startsWith('"') && field.endsWith('"')
+/// else Windows-1252 (what older Excel saves as CSV in western Europe).
-    ? field.substring(1, field.length - 1).replaceAll('""', '"')
+String decodeCardText(List<int> bytes) {
-    : field;
+  if (bytes.length >= 2) {
+    final (first, second) = (bytes[0], bytes[1]);
+    final littleEndian = first == 0xFF && second == 0xFE;
+    if (littleEndian || (first == 0xFE && second == 0xFF)) {
+      return String.fromCharCodes([
+        for (var i = 2; i + 1 < bytes.length; i += 2)
+          if (littleEndian)
+            bytes[i] | bytes[i + 1] << 8
+          else
+            bytes[i] << 8 | bytes[i + 1],
+      ]);
+    }
+  }
+  try {
+    return utf8.decode(bytes);
+  } on FormatException {
+    return String.fromCharCodes([
+      for (final byte in bytes) _windows1252[byte] ?? byte,
+    ]);
+  }
+}
+
+/// Where Windows-1252 differs from Latin-1: 0x80 to 0x9F hold signs (€,
+/// curly quotes, dashes) rather than control characters.
+const _windows1252 = {
+  0x80: 0x20AC,
+  0x82: 0x201A,
+  0x83: 0x0192,
+  0x84: 0x201E,
+  0x85: 0x2026,
+  0x86: 0x2020,
+  0x87: 0x2021,
+  0x88: 0x02C6,
+  0x89: 0x2030,
+  0x8A: 0x0160,
+  0x8B: 0x2039,
+  0x8C: 0x0152,
+  0x8E: 0x017D,
+  0x91: 0x2018,
+  0x92: 0x2019,
+  0x93: 0x201C,
+  0x94: 0x201D,
+  0x95: 0x2022,
+  0x96: 0x2013,
+  0x97: 0x2014,
+  0x98: 0x02DC,
+  0x99: 0x2122,
+  0x9A: 0x0161,
+  0x9B: 0x203A,
+  0x9C: 0x0153,
+  0x9E: 0x017E,
+  0x9F: 0x0178,
+};
+
+/// [text] without the byte order mark spreadsheets put before UTF-8 text.
+String _withoutBom(String text) =>
+    text.startsWith('\uFEFF') ? text.substring(1) : text;
+
+/// The records of [text], each its fields: split at [separator] outside
+/// double quotes, a record ending at a line break outside them. A quote
+/// that never closes spoils only its own line: reading goes on after it.
+List<List<String>> _records(String text, String separator) {
+  final records = <List<String>>[];
+  var i = 0;
+  while (i < text.length) {
+    final start = i;
+    final fields = <String>[];
+    var field = StringBuffer();
+    var quoted = false;
+    var unclosed = false;
+    while (true) {
+      if (i >= text.length) {
+        unclosed = quoted;
+        break;
+      }
+      final char = text[i];
+      if (quoted) {
+        if (char == '"' && text.startsWith('""', i)) {
+          field.write('"');
+          i += 2;
+        } else if (char == '"') {
+          quoted = false;
+          i++;
+        } else {
+          field.write(char);
+          i++;
+        }
+      } else if (char == '"' && field.toString().trim().isEmpty) {
+        // A quoted field, spaces before it left out.
+        field = StringBuffer();
+        quoted = true;
+        i++;
+      } else if (text.startsWith(separator, i)) {
+        fields.add(field.toString());
+        field = StringBuffer();
+        i += separator.length;
+      } else if (char == '\n' || char == '\r') {
+        i += char == '\r' && text.startsWith('\r\n', i) ? 2 : 1;
+        break;
+      } else {
+        field.write(char);
+        i++;
+      }
+    }
+    if (unclosed) {
+      final end = text.indexOf('\n', start);
+      i = end < 0 ? text.length : end + 1;
+      continue;
+    }
+    fields.add(field.toString());
+    records.add(fields);
+  }
+  return records;
+}

halkyon_learn_flutter/lib/src/feature/card/widget/card_text_import_sheet.dart

@@ -1,5 +1,4 @@
 import 'dart:async';
-import 'dart:convert';
 
 import 'package:halkyon_learn_flutter/l10n/app_localizations.dart';
 import 'package:halkyon_learn_flutter/src/core/content/rich_content.dart';
@@ -53,8 +52,9 @@ class _CardTextSheetState() extends State<_CardTextSheet> {
     super.dispose();
   }
 
-  /// Reads a CSV or TSV file into the field, its separator chosen by its
+  /// Reads a CSV or TSV file into the field, in whichever encoding it was
-  /// name (the person may still change it).
+  /// saved, its separator guessed from what it holds (the person may still
+  /// change it).
   Future<void> _load() async {
     final l10n = AppLocalizations.of(context);
     final file = await DependenciesScope.of(context).files.pick(
@@ -64,16 +64,12 @@ class _CardTextSheetState() extends State<_CardTextSheet> {
     if (file == null || !mounted) {
       return;
     }
-    try {
+    final text = decodeCardText(file.bytes);
-      final text = utf8.decode(file.bytes);
+    setState(() {
-      setState(() {
+      _text.text = text;
-        _text.text = text;
+      _separator = guessSeparator(text, fileName: file.name);
-        _separator = CardSeparator.forFile(file.name);
+      _unreadable = text.contains('\u0000');
-        _unreadable = false;
+    });
-      });
-    } on FormatException {
-      setState(() => _unreadable = true);
-    }
   }
 
   @override

halkyon_learn_flutter/lib/src/feature/export/data/studyset_markdown.dart

@@ -83,7 +83,7 @@ class StudysetMarkdownExporter(
       throw ArgumentError.value(studysetId, 'studysetId', 'is no studyset');
     }
     final export = _Export(texts, style, _fileOf);
-    await _studyset(export, root, all, depth: 1);
+    await _studyset(export, root, all, depth: 1, seen: {});
     return (
       markdown: '${export.blocks.join('\n\n')}\n',
       files: export.files,
@@ -91,12 +91,19 @@ class StudysetMarkdownExporter(
     );
   }
 
+  /// [studyset] and everything inside it, once each: two studysets inside
+  /// each other (moved so on two devices offline) end the walk, as [seen]
+  /// keeps those written.
   Future<void> _studyset(
     _Export out,
     Studyset studyset,
     List<Studyset> all, {
     required int depth,
+    required Set<String> seen,
   }) async {
+    if (!seen.add(studyset.id)) {
+      return;
+    }
     final content = studyset.content;
     out.heading(depth, content.title);
     if (content.secondaryTitle case final secondary?
@@ -184,7 +191,7 @@ class StudysetMarkdownExporter(
     }
 
     for (final child in childrenOf(studyset.id, all)) {
-      await _studyset(out, child, all, depth: depth + 1);
+      await _studyset(out, child, all, depth: depth + 1, seen: seen);
     }
   }
 
@@ -431,7 +438,7 @@ class _Export(
               '.${_extensionOf(mimeType)}',
     );
     files.add((path: path, bytes: file.bytes));
-    return _links[assetId] = path;
+    return _links[assetId] = _linkTo(path);
   }
 
   /// [name] in [folder], numbered (`-2`, `-3`…) where a file has it already.
@@ -448,6 +455,15 @@ class _Export(
   }
 }
 
+/// [path] as a Markdown link names it: the characters that would end the
+/// link or start a fragment there (`#`, `%`, brackets, …) percent-encoded,
+/// the rest (Japanese, say) as it is.
+String _linkTo(String path) => path.replaceAllMapped(
+  RegExp(r'[%#?()\[\]<> ]'),
+  (match) =>
+      '%${match[0]!.codeUnitAt(0).toRadixString(16).toUpperCase().padLeft(2, '0')}',
+);
+
 /// [name] without what file systems refuse in a file's name, and spaces as
 /// underscores; null for none left.
 String? _safe(String? name) {

halkyon_learn_flutter/lib/src/feature/export/widget/export_studyset.dart

@@ -41,7 +41,7 @@ Future<void> exportStudyset(BuildContext context, Studyset studyset) async {
       texts: exportTexts(l10n),
       style: style,
     );
-  } on Exception catch (_) {
+  } on Object catch (_) {
     say(l10n.exportFailed);
     return;
   }
@@ -58,7 +58,14 @@ Future<void> exportStudyset(BuildContext context, Studyset studyset) async {
     } else {
       final archive = Archive()..add(ArchiveFile.bytes('$name.md', markdown));
       for (final file in exported.files) {
-        archive.add(ArchiveFile.bytes(file.path, file.bytes));
+        archive.add(
+          ArchiveFile.bytes(file.path, file.bytes)
+            // PDFs, images and sounds are compressed already: deflating
+            // them again takes long and saves nothing.
+            ..compression = _compressed.hasMatch(file.path)
+                ? CompressionType.none
+                : null,
+        );
       }
       saved = await saver.save(
         name: '$name.zip',
@@ -66,7 +73,7 @@ Future<void> exportStudyset(BuildContext context, Studyset studyset) async {
         mimeType: 'application/zip',
       );
     }
-  } on Exception catch (_) {
+  } on Object catch (_) {
     say(l10n.exportFailed);
     return;
   }
@@ -121,3 +128,9 @@ String _fileStem(String title) {
       .replaceAll(RegExp(r'\s+'), ' ');
   return stem.isEmpty || stem.startsWith('.') ? 'studyset' : stem;
 }
+
+/// Files whose content is compressed already, by their extension.
+final _compressed = RegExp(
+  r'\.(pdf|png|jpe?g|gif|webp|mp3|m4a|aac|ogg|opus|mp4|webm|zip)$',
+  caseSensitive: false,
+);

halkyon_learn_flutter/lib/src/feature/learning_path/bloc/path_editor_bloc.dart

@@ -4,6 +4,7 @@ import 'dart:developer';
 import 'package:equatable/equatable.dart';
 import 'package:flutter_bloc/flutter_bloc.dart';
 import 'package:halkyon_learn_flutter/src/core/content/rich_content.dart';
+import 'package:halkyon_learn_flutter/src/core/sync/fractional_position.dart';
 import 'package:halkyon_learn_flutter/src/feature/learning_path/data/learning_path_deletions.dart';
 import 'package:halkyon_learn_flutter/src/feature/learning_path/data/learning_path_repository.dart';
 import 'package:halkyon_learn_flutter/src/feature/learning_path/model/exercise.dart';
@@ -52,10 +53,24 @@ class PathEditorBloc(
       );
       emit(state.copyWith(added: created.stepId));
     });
-    on<PathEditorEventStepMoved>(
+    on<PathEditorEventStepMoved>((event, emit) {
-      (event, emit) =>
+      // The list takes the new order at once; the stored one follows.
-          _paths.moveStep(_pathId, event.stepId, toIndex: event.toIndex),
+      emit(
-    );
+        state.copyWith(
+          steps: movedTo(
+            state.steps,
+            (step) => step.id == event.stepId,
+            event.toIndex,
+          ),
+        ),
+      );
+      return _paths.moveStep(
+        _pathId,
+        event.stepId,
+        toIndex: event.toIndex,
+        hidden: _pending,
+      );
+    });
   }
 
   List<LearningPathStep>? _steps;

halkyon_learn_flutter/lib/src/feature/learning_path/bloc/path_list_bloc.dart

@@ -3,6 +3,7 @@ import 'dart:developer';
 
 import 'package:equatable/equatable.dart';
 import 'package:flutter_bloc/flutter_bloc.dart';
+import 'package:halkyon_learn_flutter/src/core/sync/fractional_position.dart';
 import 'package:halkyon_learn_flutter/src/feature/learning_path/data/learning_path_deletions.dart';
 import 'package:halkyon_learn_flutter/src/feature/learning_path/data/learning_path_repository.dart';
 import 'package:halkyon_learn_flutter/src/feature/learning_path/data/path_import.dart';
@@ -57,10 +58,24 @@ class PathListBloc(
         );
       }
     });
-    on<PathListEventMoveRequested>(
+    on<PathListEventMoveRequested>((event, emit) {
-      (event, emit) =>
+      // The list takes the new order at once; the stored one follows.
-          _paths.movePath(_studysetId, event.id, toIndex: event.toIndex),
+      emit(
-    );
+        state.copyWith(
+          paths: movedTo(
+            state.paths,
+            (path) => path.id == event.id,
+            event.toIndex,
+          ),
+        ),
+      );
+      return _paths.movePath(
+        _studysetId,
+        event.id,
+        toIndex: event.toIndex,
+        hidden: _pending,
+      );
+    });
     on<PathListEventResetRequested>((event, emit) async {
       final stepIds = {
         for (final step in await _paths.stepsOf(event.id)) step.id,

halkyon_learn_flutter/lib/src/feature/learning_path/data/learning_path_repository.dart

@@ -77,20 +77,27 @@ class LearningPathRepository(final LearningPathGateway _gateway) {
     _changes.add(null);
   }
 
-  /// Moves the path [id] to [toIndex] among the paths of [studysetId].
+  /// Moves the path [id] to [toIndex] among the paths of [studysetId] the
+  /// list shows: those [hidden] (their deletion can still be undone) left
+  /// out, as the list leaves them out.
   Future<void> movePath(
     String studysetId,
     String id, {
     required int toIndex,
+    Set<String> hidden = const {},
   }) async {
     final others = [
       for (final path in await pathsOf(studysetId))
-        if (path.id != id) (id: path.id, position: path.position),
+        if (path.id != id && !hidden.contains(path.id))
+          (id: path.id, position: path.position),
     ];
-    await _gateway.updatePathPosition(
+    for (final MapEntry(key: moved, value: position) in placeAmong(
+      others,
+      toIndex,
       id,
-      positionAmong([for (final other in others) other.position], toIndex),
+    ).entries) {
-    );
+      await _gateway.updatePathPosition(moved, position);
+    }
     _changes.add(null);
   }
 
@@ -181,20 +188,27 @@ class LearningPathRepository(final LearningPathGateway _gateway) {
     return id;
   }
 
-  /// Moves the step [id] to [toIndex] among the steps of [pathId].
+  /// Moves the step [id] to [toIndex] among the steps of [pathId] the
+  /// editor shows: those [hidden] (their deletion can still be undone) left
+  /// out, as it leaves them out.
   Future<void> moveStep(
     String pathId,
     String id, {
     required int toIndex,
+    Set<String> hidden = const {},
   }) async {
     final others = [
       for (final row in await _gateway.steps({pathId}))
-        if (row.id != id) (id: row.id, position: row.position),
+        if (row.id != id && !hidden.contains(row.id))
+          (id: row.id, position: row.position),
     ]..sort((a, b) => _byPosition(a.position, a.id, b.position, b.id));
-    await _gateway.updateStepPosition(
+    for (final MapEntry(key: moved, value: position) in placeAmong(
+      others,
+      toIndex,
       id,
-      positionAmong([for (final other in others) other.position], toIndex),
+    ).entries) {
-    );
+      await _gateway.updateStepPosition(moved, position);
+    }
     _changes.add(null);
   }
 

halkyon_learn_flutter/lib/src/feature/learning_path/widget/path_list.dart

@@ -44,6 +44,13 @@ class const SliverPathList({
               child: const AppLoadingIndicator(),
             );
           }
+          if (state.failed && state.paths.isEmpty) {
+            // Read once, they show; nothing to add until then.
+            return SliverCentered(
+              padding: padding,
+              child: Text(l10n.pathsReadFailed),
+            );
+          }
           if (state.paths.isEmpty) {
             return SliverCentered(
               padding: padding,

halkyon_learn_flutter/lib/src/feature/note/bloc/note_list_bloc.dart

@@ -3,6 +3,7 @@ import 'dart:developer';
 
 import 'package:equatable/equatable.dart';
 import 'package:flutter_bloc/flutter_bloc.dart';
+import 'package:halkyon_learn_flutter/src/core/sync/fractional_position.dart';
 import 'package:halkyon_learn_flutter/src/feature/note/data/note_deletions.dart';
 import 'package:halkyon_learn_flutter/src/feature/note/data/note_repository.dart';
 import 'package:halkyon_learn_flutter/src/feature/note/model/note.dart';
@@ -16,10 +17,24 @@ class NoteListBloc(
 ) extends Bloc<NoteListEvent, NoteListState> {
   this : super(const NoteListState()) {
     on<NoteListEventStarted>(_onStarted);
-    on<NoteListEventMoveRequested>(
+    on<NoteListEventMoveRequested>((event, emit) {
-      (event, emit) =>
+      // The list takes the new order at once; the stored one follows.
-          _notes.move(_studysetId, event.id, toIndex: event.toIndex),
+      emit(
-    );
+        state.copyWith(
+          notes: movedTo(
+            state.notes,
+            (note) => note.id == event.id,
+            event.toIndex,
+          ),
+        ),
+      );
+      return _notes.move(
+        _studysetId,
+        event.id,
+        toIndex: event.toIndex,
+        hidden: _pending,
+      );
+    });
   }
 
   /// The notes as last read, and whether they were read at all.

halkyon_learn_flutter/lib/src/feature/note/data/note_repository.dart

@@ -118,17 +118,27 @@ class NoteRepository(final NoteGateway _gateway) {
     return ids;
   }
 
-  /// Moves the note [noteId] to [toIndex] among the notes of [studysetId].
+  /// Moves the note [noteId] to [toIndex] among the notes of [studysetId]
+  /// the list shows: those [hidden] (their deletion can still be undone)
+  /// left out, as the list leaves them out.
   Future<void> move(
     String studysetId,
     String noteId, {
     required int toIndex,
+    Set<String> hidden = const {},
   }) async {
     final others = [
       for (final note in await notesOf(studysetId))
-        if (note.id != noteId) note.position,
+        if (note.id != noteId && !hidden.contains(note.id))
+          (id: note.id, position: note.position),
     ];
-    await _gateway.updatePosition(noteId, positionAmong(others, toIndex));
+    for (final MapEntry(key: id, value: position) in placeAmong(
+      others,
+      toIndex,
+      noteId,
+    ).entries) {
+      await _gateway.updatePosition(id, position);
+    }
     _changes.add(null);
   }
 

halkyon_learn_flutter/lib/src/feature/note/widget/note_list.dart

@@ -44,6 +44,13 @@ class const SliverNoteList({
               child: const AppLoadingIndicator(),
             );
           }
+          if (state.failed && state.notes.isEmpty) {
+            // Read once, they show; nothing to add until then.
+            return SliverCentered(
+              padding: padding,
+              child: Text(l10n.notesReadFailed),
+            );
+          }
           if (state.notes.isEmpty) {
             return SliverCentered(
               padding: padding,

halkyon_learn_flutter/lib/src/feature/pdf_source/bloc/pdf_list_bloc.dart

@@ -5,6 +5,7 @@ import 'dart:typed_data';
 import 'package:equatable/equatable.dart';
 import 'package:flutter_bloc/flutter_bloc.dart';
 import 'package:halkyon_learn_flutter/src/core/asset/data/asset_repository.dart';
+import 'package:halkyon_learn_flutter/src/core/sync/fractional_position.dart';
 import 'package:halkyon_learn_flutter/src/feature/card/data/card_repository.dart';
 import 'package:halkyon_learn_flutter/src/feature/card/model/card.dart';
 import 'package:halkyon_learn_flutter/src/feature/note/data/note_repository.dart';
@@ -69,10 +70,24 @@ class PdfListBloc(
         await _import(emit, bytes, fileName, true);
       }
     });
-    on<PdfListEventMoveRequested>(
+    on<PdfListEventMoveRequested>((event, emit) {
-      (event, emit) =>
+      // The list takes the new order at once; the stored one follows.
-          _pdfs.move(_studysetId, event.id, toIndex: event.toIndex),
+      emit(
-    );
+        state.copyWith(
+          rows: movedTo(
+            state.rows,
+            (row) => row.source.id == event.id,
+            event.toIndex,
+          ),
+        ),
+      );
+      return _pdfs.move(
+        _studysetId,
+        event.id,
+        toIndex: event.toIndex,
+        hidden: _pending,
+      );
+    });
     on<PdfListEventRenamed>(
       (event, emit) => _pdfs.rename(event.pdfId, event.title),
     );