CodeCow/Notepad3
1
0
Fork 0
mirror of https://github.com/rizonesoft/Notepad3.git synced 2026-06-11 21:03:05 +08:00
Code Issues Actions Packages Projects Releases Wiki Activity

fix: doc & typos

Browse Source
This commit is contained in:
Rainer Kottenhoff 2026-05-02 18:15:29 +02:00
parent 77cce78a40
commit 82aaa3066d
6 changed files with 105 additions and 25 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

2
plans/schema_enhancements.md
View File

@ -27,7 +27,7 @@ low-to-moderate effort; several are trivial single-call-site fixes.
|------|------------------|
| `src/Styles.c` | `g_pLexArray[]`, load/save, mini-DSL parsers, Customize / Select / Themes dialogs, Import / Export |
| `src/StyleLexers/EditLexer.h` | `EDITLEXER`, `EDITSTYLE`, `KEYWORDLIST`, `BUFSIZE_STYLE_VALUE = 256` |
| `src/TypeDefs.h` | `STYLE_EXTENTIONS_BUFFER = 512`, `EXTENTIONS_FILTER_BUFFER = 1024` |
| `src/TypeDefs.h` | `STYLE_EXTENSIONS_BUFFER = 512`, `EXTENSIONS_FILTER_BUFFER = 1024` |
| `src/StyleLexers/*.c` | 60+ per-language lexer definitions |
| `src/Config/Config.cpp` | Settings → theme name persistence (`Settings.CurrentThemeName`) |
| `language/common_res.h` | Menu IDs: `IDM_VIEW_SCHEME=41001`, `IDM_VIEW_USE2NDDEFAULT=41002`, `IDM_VIEW_SCHEMECONFIG=41003`, `IDM_THEMES_FACTORY_RESET=37001`, `IDM_THEMES_STD_CFG=37002` |

75
readme/schema/CustomSchema.md
View File

@ -78,7 +78,7 @@ Every `EDITSTYLE::szValue` is a **semicolon-delimited list of `attribute:value`
font:Consolas; size:11; fore:#D7BA7D; back:#1E1E1E; bold
```
Buffer limit: **255 characters** per style string. Longer strings are silently truncated — see [§11 Limitations](#11-known-limitations--quirks).
Buffer limit: **255 characters** per style string (`BUFSIZE_STYLE_VALUE` in `src/StyleLexers/EditLexer.h`). Longer strings are silently truncated on load.
### 3.1 Attributes
@ -165,7 +165,7 @@ When you open a file, Notepad3 runs a pipeline to pick the right schema. It stop
1. **File variables** (`vim:` / `emacs:` modelines inside the file). If the file declares `mode: python;` or `-*- mode: cpp -*-`, the declared mode name is matched against schema names (case-insensitive prefix) and then against extension lists. Disabled by `Settings2.NoFileVariables=1`.
2. **Shebang detection** for `.cgi` / `.fcgi` files or files flagged as CGI by mode. Recognised interpreters: `python`, `ruby`, `bash`/`sh`, `perl`, `tcl`, `node`/`js`, `php`. Disabled by `Settings2.NoCGIGuess=1`.
3. **Regex match on the file name.** Any entry in a schema's extension list that starts with a backslash is treated as a regex — e.g. `\^CMakeLists$` matches the file name `CMakeLists.txt` once combined with the bare `txt` entry below.
3. **Regex match on the file name.** Any entry in a schema's extension list that starts with a backslash is treated as a PCRE2 regex applied to the bare filename — e.g. `\^CMakeLists$` matches the extensionless file `CMakeLists`. See the [Extension lists](#extension-lists) subsection below for full syntax, escaping rules, and worked examples.
4. **Plain extension match.** First lexer whose extension list contains the file's extension wins. Extension comparison is case-insensitive; separator is `;` (semicolon or space both work in practice).
5. **HTML/XML sniff** — if the first bytes start with `<`, classify as HTML or XML. Disabled by `Settings2.NoHTMLGuess=1`.
6. **Shebang fallback** for extension-less files (same recognisers as step 2).
@ -181,16 +181,79 @@ When you open a file, Notepad3 runs a pipeline to pick the right schema. It stop
### Extension lists
Each schema has a hard-coded default extension list (compiled in), and a user-editable override stored in the schema's INI section as:
Each schema carries a hard-coded default extension list (compiled into `Notepad3.exe`) plus an optional user override stored in the schema's INI section:
```ini
[Python Script]
FileNameExtensions=py;pyw;pyi;\^setup\.py$
FileNameExtensions=py;pyw;pyi;\^setup\.py$;\^test_.*\.py$
```
The override replaces the default completely (not merged). Clearing the field in *Customize Schemes* restores the default. Buffer limit: **~512 characters per schema**.
The override **replaces** the default; the two are not merged. The field accepts both plain file extensions and full-filename regex patterns in a single semicolon-separated list.
> If two schemas both claim the same extension, the one that appears first in the internal schema array wins. There is currently no UI to reorder schemas.
#### Syntax
| Aspect | Rule |
|---|---|
| Separator | `;` (semicolon) is canonical. A single space also works as a separator and surrounding whitespace around `;` is tolerated. |
| Case | Comparison is **case-insensitive** for both plain and regex entries. |
| Plain entry | A bare extension token like `py` or `cpp`. Matched as a whole token against the file's extension; substring matches are excluded. The leading `.` of the file's extension is stripped before comparison, so `py` matches `foo.py`, `foo.PY`, and `foo.Py`. |
| Regex entry | An entry whose first character is a **backslash** `\`. Everything after the leading `\` is the regex pattern. The `\` is the marker, not part of the pattern. |
| Buffer limit | **512 characters** per schema (counting all entries combined). Longer values are silently truncated when read from the INI — no warning is shown. |
#### Regex extension patterns
- **Engine:** PCRE2 (the same engine the editor's Find/Replace dialog uses), so the full PCRE2 syntax is available — character classes, alternation, lookarounds, named groups, inline modifiers.
- **Match target:** the **bare filename** including extension (e.g. `setup.py`, `CMakeLists.txt`). The directory portion of the path is stripped before matching, so patterns cannot anchor on parent directories.
- **Anchors are NOT implicit.** A pattern like `\setup\.py` is treated as `setup\.py` and matches anywhere in the filename — it accepts `setup.py`, `mysetup.python`, and `setup.python.bak`. Add `^` and `$` explicitly when you mean the whole filename.
- **`.` is a metachar.** Always escape literal dots as `\.`. Otherwise `\foo.py` will also match `fooXpy`, `foo-py`, etc.
- **Multiple regex entries are allowed** in the same field, each prefixed with its own `\` and separated by `;`.
- **Plain and regex entries can be freely mixed** in the same `FileNameExtensions` value.
- **No per-pattern case-sensitivity flag** is exposed by the dialog. If you need a case-sensitive regex for one pattern, use the PCRE2 inline modifier `(?-i)` at the start of that pattern.
#### Worked examples
```ini
; CMakeLists has no extension — only a regex can catch it
[CMake]
FileNameExtensions=cmake;ctest;\^CMakeLists$;\^CMakeLists\.txt$
; Dockerfile family — match Dockerfile, Dockerfile.dev, Dockerfile.prod, …
[Docker]
FileNameExtensions=dockerfile;\^Dockerfile(\..+)?$
; Hidden shell config files starting with a dot
[Bash Script]
FileNameExtensions=sh;bash;\^\.bashrc$;\^\.bash_profile$;\^\.profile$
; Route Python test files to a separate schema
[Python Test]
FileNameExtensions=\^test_.*\.py$;\^.*_test\.py$
; Nginx config files — fixed filenames plus the conf.d/*.conf convention
[Nginx Config]
FileNameExtensions=conf;\^nginx\.conf$;\^mime\.types$
```
Common pitfalls illustrated above:
- `\^setup\.py$` matches `setup.py` but **not** `setup.py.bak` (because of the trailing `$`). It does match `Setup.py`, but only thanks to the global case-insensitive flag.
- `\setup\.py` (no anchors) also matches `mysetup.py.bak` and `setup.python` — almost never what you want.
- `\^foo.py$` (missing `\.`) matches `foo.py` *and* `fooXpy` — always escape the dot.
#### Precedence
Within `Style_SetLexerFromFile()` the auto-detect pipeline tries the **regex match across all schemas first**, and only then falls back to the plain-extension match across all schemas. Within a single schema the two are independent: a regex hit and a plain-extension hit can both be present, but only the regex pass gets a chance to fire before the plain pass runs.
When two schemas claim the same plain extension, or both match a file with their regex entries, the schema that appears first in the internal schema array wins. There is currently no UI to reorder schemas.
#### Editing in *Customize Schemes*
Select a schema node in the left-hand tree and edit the *FileNameExtensions* field. On save:
- **Empty field** (or whitespace only) → the compiled-in default is restored.
- **Field equals the default** → the INI key is removed entirely so `Notepad3.ini` stays clean.
- **Field differs from default** → the full value is written back to the schema's section.
Persistence still requires *Save Settings On Exit* or *Settings → Save Settings Now* (**F7**); see [§2](#2-layered-override-model).
---

6
src/Dialogs.c
View File

@ -1483,7 +1483,7 @@ CASE_WM_CTLCOLOR_SET:
LPWSTR const args2_buf = StrgWriteAccessBuf(hargs2_str, StrgGetAllocLength(hargs_str));
HSTRINGW hflt_str = StrgCreate(NULL);
LPWSTR const flt_buf = StrgWriteAccessBuf(hflt_str, EXTENTIONS_FILTER_BUFFER);
LPWSTR const flt_buf = StrgWriteAccessBuf(hflt_str, EXTENSIONS_FILTER_BUFFER);
GetDlgItemText(hwnd, IDC_COMMANDLINE, args_buf, (int)StrgGetAllocLength(hargs_str));
StrgSanitize(hargs_str);
@ -7335,7 +7335,7 @@ bool OpenFileDlg(HWND hwnd, HPATHL hfile_pth_io, const HPATHL hinidir_pth)
Path_Empty(hfile_pth_io, false);
}
WCHAR szDefExt[64] = { L'\0' };
WCHAR szFilter[EXTENTIONS_FILTER_BUFFER];
WCHAR szFilter[EXTENSIONS_FILTER_BUFFER];
Style_GetFileFilterStr(szFilter, COUNTOF(szFilter), szDefExt, COUNTOF(szDefExt), false);
HPATHL hpth_dir = Path_Copy(hinidir_pth);
@ -7369,7 +7369,7 @@ bool SaveFileDlg(HWND hwnd, HPATHL hfile_pth_io, const HPATHL hinidir_pth)
}
WCHAR szDefExt[64] = { L'\0' };
WCHAR szFilter[EXTENTIONS_FILTER_BUFFER];
WCHAR szFilter[EXTENSIONS_FILTER_BUFFER];
Style_GetFileFilterStr(szFilter, COUNTOF(szFilter), szDefExt, COUNTOF(szDefExt), true);
HPATHL hpth_dir = Path_Copy(hinidir_pth);

2
src/StyleLexers/EditLexer.h
View File

@ -40,7 +40,7 @@ typedef struct _editlexer
int resID; // language resource
LPCWSTR pszName; // config/settings section
LPCWSTR pszDefExt; // default file name ext (4 reset)
WCHAR szExtensions[STYLE_EXTENTIONS_BUFFER];
WCHAR szExtensions[STYLE_EXTENSIONS_BUFFER];
PKEYWORDLIST pKeyWords;
EDITSTYLE Styles[]; // must be last

41
src/Styles.c
View File

@ -716,7 +716,7 @@ bool Style_Import(HWND hwnd)
HPATHL hfile_pth = Path_Allocate(NULL);
HSTRINGW hflt_str = StrgCreate(NULL);
wchar_t* const flt_buf = StrgWriteAccessBuf(hflt_str, EXTENTIONS_FILTER_BUFFER);
wchar_t* const flt_buf = StrgWriteAccessBuf(hflt_str, EXTENSIONS_FILTER_BUFFER);
GetLngString(IDS_MUI_FILTER_INI, flt_buf, (int)StrgGetAllocLength(hflt_str));
StrgSanitize(hflt_str);
@ -754,8 +754,22 @@ static void _LoadLexerFileExtensions()
Lexer_Section = (iLexer == 0) ? L"Default Text" : L"2nd Default Text";
}
IniSectionGetString(Lexer_Section, L"FileNameExtensions", g_pLexArray[iLexer]->pszDefExt,
g_pLexArray[iLexer]->szExtensions, COUNTOF(g_pLexArray[iLexer]->szExtensions));
// Read into an oversized buffer so we can detect when the INI value
// would exceed the per-schema STYLE_EXTENSIONS_BUFFER limit and warn,
// instead of silently dropping the tail.
WCHAR tmpExt[STYLE_EXTENSIONS_BUFFER * 2] = { L'\0' };
size_t const cchRead = IniSectionGetString(Lexer_Section, L"FileNameExtensions",
g_pLexArray[iLexer]->pszDefExt, tmpExt, COUNTOF(tmpExt));
if (cchRead >= (size_t)STYLE_EXTENSIONS_BUFFER) {
WCHAR warnMsg[256];
StringCchPrintf(warnMsg, COUNTOF(warnMsg),
L"Notepad3: FileNameExtensions for [%s] is %zu chars, exceeds %d-char limit; truncating.\n",
Lexer_Section, cchRead, STYLE_EXTENSIONS_BUFFER - 1);
OutputDebugStringW(warnMsg);
}
StringCchCopy(g_pLexArray[iLexer]->szExtensions,
COUNTOF(g_pLexArray[iLexer]->szExtensions), tmpExt);
// don't allow empty extensions settings => use default ext
if (StrIsEmpty(g_pLexArray[iLexer]->szExtensions)) {
@ -1001,7 +1015,7 @@ bool Style_Export(HWND hwnd)
HPATHL hfile_pth = Path_Allocate(NULL);
HSTRINGW hflt_str = StrgCreate(NULL);
wchar_t* const flt_buf = StrgWriteAccessBuf(hflt_str, EXTENTIONS_FILTER_BUFFER);
wchar_t* const flt_buf = StrgWriteAccessBuf(hflt_str, EXTENSIONS_FILTER_BUFFER);
GetLngString(IDS_MUI_FILTER_INI, flt_buf, (int)StrgGetAllocLength(hflt_str));
StrgSanitize(hflt_str);
@ -2640,6 +2654,9 @@ PEDITLEXER Style_RegExMatchLexer(LPCWSTR lpszFileName)
++f; // exclude '\'
char regexpat[HUGE_BUFFER] = { '\0' };
WideCharToMultiByte(CP_UTF8, 0, f, (int)(e-f), regexpat, (int)COUNTOF(regexpat), NULL, NULL);
// Strip incidental whitespace around the pattern so entries like
// "py; \^setup\.py$ ;txt" don't carry a leading/trailing space into PCRE2.
StrTrimA(regexpat, " \t");
if (RegExFind(regexpat, chFilePath, false, NULL) >= 0) {
return g_pLexArray[iLex];
@ -3050,7 +3067,7 @@ bool Style_GetFileFilterStr(LPWSTR lpszFilter, int cchFilter, LPWSTR lpszDefExt,
WCHAR filterAll[80] = { L'\0' };
GetLngString(IDS_MUI_FILTER_ALL, filterAll, COUNTOF(filterAll));
WCHAR filterDef[EXTENTIONS_FILTER_BUFFER] = { L'\0' };
WCHAR filterDef[EXTENSIONS_FILTER_BUFFER] = { L'\0' };
WCHAR ext[64] = { L'\0' };
WCHAR append[80] = { L'\0' };
bool bCurExtIncl = false;
@ -4457,7 +4474,7 @@ void Style_GetStyleDisplayName(PEDITSTYLE pStyle, LPWSTR lpszName, int cchName)
//
int Style_GetLexerIconId(PEDITLEXER plex)
{
WCHAR pszFile[STYLE_EXTENTIONS_BUFFER << 1];
WCHAR pszFile[STYLE_EXTENSIONS_BUFFER << 1];
LPCWSTR pszExtensions = StrIsNotEmpty(plex->szExtensions) ? plex->szExtensions : plex->pszDefExt;
StringCchCopy(pszFile, COUNTOF(pszFile), L"*.");
@ -4564,10 +4581,10 @@ static bool _ApplyDialogItemText(HWND hwnd, PEDITLEXER pDlgLexer, PEDITSTYLE pD
bool bChgNfy = false;
bool bForce = false;
WCHAR szBuf[max(BUFSIZE_STYLE_VALUE, STYLE_EXTENTIONS_BUFFER)] = { L'\0' };
WCHAR szBuf[max(BUFSIZE_STYLE_VALUE, STYLE_EXTENSIONS_BUFFER)] = { L'\0' };
GetDlgItemText(hwnd, IDC_STYLEEDIT, szBuf, COUNTOF(szBuf));
// normalize
WCHAR szBufNorm[max(BUFSIZE_STYLE_VALUE, STYLE_EXTENTIONS_BUFFER)] = { L'\0' };
WCHAR szBufNorm[max(BUFSIZE_STYLE_VALUE, STYLE_EXTENSIONS_BUFFER)] = { L'\0' };
Style_CopyStyles_IfNotDefined(szBuf, szBufNorm, COUNTOF(szBufNorm));
if (StringCchCompareXI(szBufNorm, pDlgStyle->szValue) != 0) {
@ -4652,7 +4669,7 @@ INT_PTR CALLBACK Style_CustomizeSchemesDlgProc(HWND hwnd, UINT umsg, WPARAM wPar
static bool bWarnedNoIniFile = false;
static int iDMHighliteContrast = 75;
static WCHAR tchTmpBuffer[max(BUFSIZE_STYLE_VALUE, STYLE_EXTENTIONS_BUFFER)] = {L'\0'};
static WCHAR tchTmpBuffer[max(BUFSIZE_STYLE_VALUE, STYLE_EXTENSIONS_BUFFER)] = {L'\0'};
static UT_array *pStylesBackup = NULL;
switch (umsg) {
@ -4753,7 +4770,7 @@ INT_PTR CALLBACK Style_CustomizeSchemesDlgProc(HWND hwnd, UINT umsg, WPARAM wPar
pCurrentStyle = &(pCurrentLexer->Styles[STY_DEFAULT]);
iCurStyleIdx = STY_DEFAULT;
SendDlgItemMessage(hwnd, IDC_STYLEEDIT, EM_LIMITTEXT, max(BUFSIZE_STYLE_VALUE, STYLE_EXTENTIONS_BUFFER) - 1, 0);
SendDlgItemMessage(hwnd, IDC_STYLEEDIT, EM_LIMITTEXT, max(BUFSIZE_STYLE_VALUE, STYLE_EXTENSIONS_BUFFER) - 1, 0);
MakeBitmapButton(hwnd, IDC_PREVSTYLE, IDB_PREV, -1, -1);
MakeBitmapButton(hwnd, IDC_NEXTSTYLE, IDB_NEXT, -1, -1);
@ -5098,7 +5115,7 @@ CASE_WM_CTLCOLOR_SET:
//ImageList_EndDrag();
HTREEITEM htiTarget = TreeView_GetDropHilight(hwndTV);
if (htiTarget) {
WCHAR tchCopy[max(BUFSIZE_STYLE_VALUE, STYLE_EXTENTIONS_BUFFER)] = {L'\0'};
WCHAR tchCopy[max(BUFSIZE_STYLE_VALUE, STYLE_EXTENSIONS_BUFFER)] = {L'\0'};
TreeView_SelectDropTarget(hwndTV, NULL);
GetDlgItemText(hwnd, IDC_STYLEEDIT, tchCopy, COUNTOF(tchCopy));
TreeView_Select(hwndTV, htiTarget, TVGN_CARET);
@ -5216,7 +5233,7 @@ CASE_WM_CTLCOLOR_SET:
case IDC_STYLEEDIT: {
if (HIWORD(wParam) == EN_CHANGE) {
WCHAR tch[max(BUFSIZE_STYLE_VALUE, STYLE_EXTENTIONS_BUFFER)] = {L'\0'};
WCHAR tch[max(BUFSIZE_STYLE_VALUE, STYLE_EXTENSIONS_BUFFER)] = {L'\0'};
GetDlgItemText(hwnd, IDC_STYLEEDIT, tch, COUNTOF(tch));

4
src/TypeDefs.h
View File

@ -218,8 +218,8 @@ typedef enum BUFFER_SIZES {
EDGELINE_NUM_LIMIT = 256,
ANSI_CHAR_BUFFER = 258,
STYLE_EXTENTIONS_BUFFER = 512,
EXTENTIONS_FILTER_BUFFER = (STYLE_EXTENTIONS_BUFFER << 1),
STYLE_EXTENSIONS_BUFFER = 512,
EXTENSIONS_FILTER_BUFFER = (STYLE_EXTENSIONS_BUFFER << 1),
FNDRPL_BUFFER = 4096, // TODO: eliminate limit
LONG_LINES_MARKER_LIMIT = 8192,
CMDLN_LENGTH_LIMIT = 8192