wxmedit/wxmedit

wxMEdit, a Cross-platform Text/Hex Editor, an improved version of MadEdit

wxmedit's Issues

Switch to wxSmith RAD tool and fix inappropriate dialog items bug of MadEdit/wxMEdit

Switch to wxSmith RAD tool.
And redesign the dialogs to fix the bug of MadEdit/wxMEdit that font size and controls layout of dialogs were inappropriate. e.g. :

• all fonts of all dialogs were too big under many Unix like OS.
• all combo boxes were not big enough under most Unix like OS.
• some controls of syntax highlighting dialog overlaped each other under most Unix like OS.
• some controls were clipped when resized the find/relace/find-in-files dialog to minimal size under Windows/Unix like OS.
• etc.

Most recent lists of wxMEdit/MadEdit were buggy

The recent lists:

• Recent Encodings
• Recent Fonts
• Recent Find Texts
• Recent Replace Texts
• Recent Find Directories
• Recent Find Filters
• Recent Find Exclude Filters

Problem of them:

1. All of them were case insensitive under Windows, but most of them should be case sensitive even under Windows.
2. The first(most recent) item shown in associated menu was truncated if it contained "/".

e.g.:

• Select view encoding "[CP437] Windows/DOS OEM (CP 437)", the most recent encoding in menu should be "[CP437] Windows/DOS OEM (CP 437)", but it shows "DOS OEM (CP 437)".

Crash when delete a character in a line which contains two or more characters with MadEdit/wxMEdit built with vc2010

Edit a file with MadEdit/wxMEdit built with vc2010.
Then delete a character from a line which contains two or more characters.
Then the editor crash.

CP936, CP949, CP950 should support converting '\xFF' to Unicode

CP936, CP949, CP950 should support '\xFF' converting to Unicode.

Test program on Windows7:

#include <windows.h>
#include <iostream>

int main()
{
    WCHAR dest[8] = {0,0,0,0,0,0,0,0};

    DWORD n = MultiByteToWideChar(936, 0, "\xFF", -1, NULL, 0);
    MultiByteToWideChar(936, 0, "\xFF", -1, dest, 8);
    std::cout << "936 " << n << " " << std::hex << dest[0] << std::endl;

    n = MultiByteToWideChar(949, 0, "\xFF", -1, NULL, 0);
    MultiByteToWideChar(949, 0, "\xFF", -1, dest, 8);
    std::cout << "949 " << n << " " << std::hex << dest[0] << std::endl;

    n = MultiByteToWideChar(950, 0, "\xFF", -1, NULL, 0);
    MultiByteToWideChar(950, 0, "\xFF", -1, dest, 8);
    std::cout << "950 " << n << " " << std::hex << dest[0] << std::endl;

    return 0;
}

the result:

936 2 f8f5
949 2 f8f7
950 2 f8f8

The Encoding "EUC-JP" of MadEdit/wxMEdit run under UNIX were quite different from under Windows

The Encoding "EUC-JP" of wxMEdit/MadEdit run under UNIX like OS were quite different from under Windows. Under windows it was DOUBLE-byte encoding "CP20932" with single-byte 0xA0 and 0xFF mapping changed, but under Linux it was TRIPLE-byte encoding.

e.g.:

characters: U+00FA(ú)      and U+00FC(ü)
CP20932:    "\xAB\x62"     and "\xAB\x64"
EUC-JP:     "\x8F\xAB\xE2" and "\x8F\xAB\xE4"

while MadEdit/wxMEdit treat them as DOUBLE-byte encoding even under UNIX like OS,and both the characters will be converted to "\x8F\xAB" under UNIX like OS.

Add translation for print preview

The title "wxMEdit Print Privew" and the status bar "Page of n / s" have no translation yet.

Chinese translation for "Words" of Word Count Dialog should updated

Chinese translation for "Words" of Word Count Dialog should be "词数"/"詞數".
Since wxMEdit count Chinese words rather than Chinese characters if the ICU library version >= 50.
e.g.:
Edit "汉字中文字符"/"漢字中文字符" with wxMEdit/ICU 52.
The count of words is 3, for there are 3 words(词/詞): "汉字"/"漢字", "中文", "字符".

While edit the same Chinese text with wxMEdit/ICU 44.
The count of words is 6, for there are 6 characters(字).
And we can still condisder it means 6 words(词/詞), but the word segmentation is not good enough :(.

Update packaging configuration files and make packages

• update packaging configuration files for RPM, debian and FreeBSD.
• make RPM, debian and FreeBSD packages for next release (2.9.6).

CP932(Windows-31J) missed converting some Single-bytes to Unicode

CP932(Windows-31J) missed converting some Single-bytes to Unicode.
Test program on Windows7:

#include <windows.h>
#include <iostream>

void win_mbtou(UINT cp, const char* src)
{
    WCHAR dest[8] = {0,0,0,0,0,0,0,0};

    DWORD n = MultiByteToWideChar(cp, 0, src, -1, NULL, 0);
    MultiByteToWideChar(cp, 0, src, -1, dest, 8);

    std::cout << cp << " " << n << " " << std::hex << dest[0] << std::endl;
    std::cout << std::dec;
}

int main()
{
    win_mbtou(932, "\xA0");
    win_mbtou(932, "\xFD");
    win_mbtou(932, "\xFE");
    win_mbtou(932, "\xFF");

    return 0;
}

the result:

932 2 f8f0
932 2 f8f1
932 2 f8f2
932 2 f8f3

Case-insensitive searching '\u00c0' matches '\u0020'(space) under Windows Chinese Edition

Case-insensitive searching '\u00c0'('À') matches '\u0020'(space) under Windows either Chinese-Simplified or Chinese-Traditional Edition.

These are many other characters meet the same problem.

Merge of the different branches

Hi, Thank you for your bug fix!

I have a suggestion about the project. Why not cooperate/combine with other similar projects?

There exist some user's patches to make it usable in recent compile environment in the original project web site and others in the internet. For example, here(https://code.google.com/p/madedit-pv), and yours(http://code.google.com/p/wxmedit). The user would be confused if there exist similar projects. It's also waste of time for a bug be solved multiple times (such as the ClientBitmap bug).

madedit-pv exist for a while and includes several useful features, such as hight-light of key words. It would be nice if both of the projects can sync to create a great editor.

Thanks!

Add new commands: delete previous/next word

Crash on restart if "Max line length before Line-Wrap" was set too large

reported by zaneto[at]users.sf.net.
https://sourceforge.net/p/madedit/discussion/421172/thread/033c9f33/

Version: 0.2.9.1

Repro steps:

1. In "Tools" - "Options" - "Edit", set the value of "Max line length before
   Line-Wrap" to "999999999".
2. Press "OK" button.
3. Close MadEdit.
4. Restart MadEdit.

Expected result:
MadEdit is restarted.

Actual result:
An error will pop up, signifying memory access failure (The memory could not be "written"). MadEdit cannot be started.

IsUTF8()/IsTextUTF8() of wxMEdit/MadEdit cannot distinguish invalid encoded bytes.

IsUTF8()/IsTextUTF8() of wxMEdit/MadEdit cannot distinguish invalid encoded bytes:

• first byte is '\xC0' - '\xC1'.
• first byte is '\xED' and sencod byte not less than '\xA0'.
• first byte is '\xE0' and second byte less than '\xA0'.
• first byte is '\xF0' and second byte less than '\x90'.
• first byte is '\xF4' and second byte greater than '\x8F'.

Add tags for each MadEdit release since 2006.3.5 subversion initialization

Can not convert case of some characters

Case conversion of some characters are incorrect.
e.g. U+A758 / U+A759 on linux, U+10414 / U+1043C and other non-bmp letters on windows.

Case-insensitive searching and sorting has the same problem.

Crash when hex-replacing a not-empty string with empty by "Replace All"

https://sourceforge.net/p/madedit/bugs/14/

Hex-replacing any not-empty string with empty, if found, "Replace All" will cause program crash.

Update Simplified Chinese translation

UTF-16 checker should be able to detect matched text encoding by surrogate pairs

Some characters were converted incorrectly from Unicode to Single-byte Encoding

Some characters were converted incorrectly from Unicode to Single-byte Encoding (U+00D7 should be 0xAA in ISO-8859-8).
It made test case "test_singlebyte_conv()" failed.
See also 4b30f50.

Some system encoding (e.g. Windows-1258) cannot be detected

Some system encoding (e.g. Windows-1258) can't be detected for wxFontEncoding doesn't support it.

Text less than 4 bytes and not match UTF-16LE/UTF16BE will be detected as UTF32LE after refactoring UTF-x checker

Some characters were converted incorrectly between Unicode and Single-byte Encoding

Some characters were converted incorrectly between Unicode and Single-byte Encoding (e.g. 0xA1 in ISO-8859-7 should be U+2018).
It made test case "test_singlebyte_conv()" failed.
See also db1ba03.

Correct invalid UTF-8 detection.

ICU Encoding Detector detect UTF-8 is not strict enough as UTF-8 checker.
If the UTF-8 checker detect a file is not UTF-8, the file should not be detected as UTF-8 text even if ICU Encoding Detector detect it as UTF-8 text.

Merge the patches of MadEdit 0.3.0 (gogo)

This is a sub-issue of #15 .
Merge the patches of MadEdit 0.3.0 by Gospodin Gurov (aka gogo) : https://sourceforge.net/p/madedit/discussion/421172/thread/48278865/ .

• bookmark support
• word wrap for non-ASCII words
• new commands: delete previous/next word
• "Search" dialog improvements
• maximized behavior improvements

Fonts garbled under Mac OS X

The line number font and status bar font of MadEdit/wxMEdit was garbled under Mac OS X.

Some characters were converted incorrectly from Unicode to Double-byte Encoding

Some characters were converted incorrectly from Unicode to Double-byte Encoding (e.g. U+2160 should be "\x87\x54" in Windows-31J, U+256D should be "\xA2\x7E" in CP950).
It made test case "test_doublebyte_conv()" failed.
See also 45d781a

Merge the patches in Arch AUR

This is a sub-issue of #15 .
Merge the patches of https://aur.archlinux.org/packages/madedit/ and https://aur.archlinux.org/packages/madedit-svn/ .

• madedit-add-file-changed-event.patch
• madedit-desktop-file.patch
• madedit-highlight-blink-fix.patch
• madedit-max-linelength-fix.patch
• madedit-onpaint-fix.patch
• madedit-remove-bundled.patch

Support more graph characters of ASCII-Art

Most of C0-characters should be mapped to graph characters if syntax is ASCII-Art.
e.g.
0x03, 0x04, 0x05, 0x06 should be mapped to the heart, club, diamond, spade characters:
♥♦♣♠

Improve Search Dialogs

Improve Search Dialog and Replace Dialog:

• add option "Wrap around"
• merge Search Dialog to Replace Dialog and remove Search Dialog
• change message notification from poping up message-box to showing in bottom of dialog (like Notepad++)
• change initial position of Dialog and font-size of input

make search dialog like in browsers

Search dialog would be nice to have in a main frame, like in browsers.

Crash on UNIX-like OS when create a second tab

https://sourceforge.net/p/madedit/bugs/15/
https://sourceforge.net/p/madedit/bugs/20/

MadEdit / wxMEdit f759f1e or before based on wxGTK on UNIX-like OS will crash when create a second tab, either by opening two / more files at the start time or opening a second new file / existed file.

Refactor encoding implementation to support more encodings easily

1. refactor encoding implementation.
2. fix encoding coversion bugs.
3. add more single byte encodings support.
4. add multibyte encodings(e.g. GB18030) support.

MadEdit/wxMEdit will crash under Unix like OS when open Find In Files Result panel if the caret is in top several lines of text editing area

1. Edit a file under Unix like OS with MadEdit/wxMEdit.
2. Move the caret to the top of editing area, e.g. the first line.
3. Select "Search" -> "Show/Hide the Results of Find/Replace in Files".

then the editor crashs.

Import the original MadEdit subversion repository and omit unused revisions

There are 38 revisions of svn://svn.code.sf.net/p/madedit/code/ should be omited when importing:

• 1-5
• 7,8
• 533-546
• 548,549
• 572,573
• 575,576
• 579-589

All these were unused.

New implemention of Fullwidth Counter missed to count Ambiguous Width Characters

e.g. "⊂" , "…" etc.

Switch to bakefile building system

Switch to bakefile building system.

Three modally dialogs should avoid the bug 4373 of wxMSW-2.8/wxMSW-2.6

Strictly speaking, this is the bug of wxMSW-2.8/wxMSW-2.6:
http://trac.wxwidgets.org/ticket/4373

To reproduce the bug:

1. Rebuild MadEdit / wxMEdit revision 8a26033 or before with official wxMSW-2.8.12 or lower.
2. Start the rebuilt MadEdit.exe / wxmedit.exe.
3. Create two new files NoName1 and NoName2, the focus is on NoName2 now.
4. Select the BUGGY-MODALLY-DIAGLOG in "Tools" menu.
5. Close the modally dialog.
6. Then switch to NoName1, and reopen the same BUGGY-MODALLY-DIAGLOG.
7. Close the modally dialog again, and the focus will jump to NoName2!!

The BUGGY-MODALLY-DIAGLOGs are:

• "Convert File Encoding..." dialog.
• "Syntax Highlighting Settings..." dialog.
• "Options..." dialog.

All of them were in "Tools" menu.

Add bookmark support

Crash when hex-searching a blank string

like #5

Hex-searching a string only contain space(s) and/or tab(s), whichever of "Find Next", "Find Previous", "Replace" or "Replace All" will cause program crash.

Add US-ASCII/ISO-646 text encoding detection support and edit US-ASCII/ISO-646 text file with default encoding

Open all files in one window if drags multi-files to wxmedit.desktop under Gnome/KDE

Refactor Encoding Detector and fix encoding detection problems of wxMEdit/MadEdit

• The source code of Universal Charset Detector shoud NOT be included.
• ISO-8859-1 was be detected as Windows-1252 most of the time.
• ISO-8859-x / Windows-125x could not be correctly detected some times.

And these detection problems made the test cases failure that were enabled in 33e7514 .

Desktop icon is fuzziness under Xfce4/KDE4

Desktop icon is fuzziness in Dolphin of KDE4 with 256 pixel icon view or desktop tips of Xfce4.

Cannot detect encoding of UTF-32BE/LE file by IsTextUTF32BE/LE() if the file is less than 16 bytes

Cannot detect encoding of UTF-32BE/LE file by IsTextUTF32BE/LE() of wxMEdit/MadEdit if the file is less than 16 bytes.

Replace Word Count related code with ICU implemention

Merge the patches in FreeBSD ports

This is a sub-issue of #15 .
Merge the patches in ftp://ftp.freebsd.org/pub/FreeBSD/ports/ports/ports.tar.gz .

• patch-charset-detector__src__entry__impl.cpp
• patch-charset-detector__src__prmem.h
• patch-src__MadEdit__MadEdit.cpp

Selection next word shoud select following spaces

Selection next word shoud select following spaces, as the behavior of SciTE/UltraEdit.
This also affect the behavior of deletion next word (#47).

Add GB18030 BOM detection support

port to wxWidgets 3.0