This is the documentation for SET's Editor.
Copyright © 1996, 1997 Salvador Eduardo Tropea
This documentation may be freely distributed, provided this copyright notice is left intact on all copies.
This documentation may be freely distributed with the editor or the RHIDE package or any part thereof, provided this copyright notice is left intact on all copies.
Permission is granted to copy and distribute modified versions of this manual under the conditions for verbatim copying, provided that the entire resulting derived work is distributed under the terms of a permission notice identical to this one.
People who helped me develop the editor:
This editor is included in the Robert Höhne's RHIDE as a replacement for the original TVision's editor class.
SET's editor is an editor designed to be used by programmers; the main target of the editor is C and C++ code but Pascal and Clipper are supported too.
The editor was designed to be very similar to the DOS standard editors for C, especially to Borland's IDE editor. The editor supports a lot of WordStar style commands plus some CUA commands, so if you have used any editor that uses these kinds of commands you'll find my editor very familiar. On the other hand if you have never used a DOS editor, especially if you use VI on UNIX machines you'll feel lost. You can configure the keyboard but this feature is relatively limited. How to configure the keyboard .
For now I only plan to distribute my editor inside of RHIDE. Requirements are the requirements of RHIDE.
E-Mail: Salvador Eduardo Tropea Telephone: (+541) 759-0013 Postal Address: Salvador E. Tropea Curapaligüe 2124 (1678) Caseros - 3 de Febrero Prov: Buenos Aires Argentina
Read the conventions topic first to understand my way of indicating keystrokes.
In the description of each command I'll include the internal name used by the editor because this name is needed to configure the keyboard.
I'll use some conventions when talking about the keystrokes needed to trigger some command. So here is what I use:
The key named Ctrl or Control is represented as ^; this key doesn't have any effect used alone inside of the editor so the ^ symbol will be used only in conjunction with the name of some key indicating that you must press the two keys at the same time. For example, ^A is Ctrl and A at the same time. When I say "at the same time" that means: press Ctrl, hold it, and press the other key; that's the reason to put A after Ctrl.
To indicate a sequence of keystrokes I'll use a dash to separate the keys. For example, ^K-B is Ctrl and K at the same time, and then press B, of course release ^K first.
To indicate keys pressed at the same time other than ^x I'll use a plus. For example, Shift+^Insert is the three keys at the same time!
I don't think that you are stupid; the editor is written for programmers, but I wanted to make that clear to avoid problems ;-).
Action | Command | Key | Alternate |
---|---|---|---|
Character left | cmbCharLeft | Left arrow | ^S |
Character right | cmbCharRight | Right arrow | ^D |
Word left | cmbWordLeft | ^Left arrow | ^A |
Word right | cmbWordRight | ^Right arrow | ^F |
Line up | cmbLineUp | Up arrow | ^E |
Line down | cmbLineDown | Down arrow | ^X |
Scroll the screen one line up | cmbScrollUp | ^W | |
Scroll the screen one line down | cmbScrollDown | ^Z | |
Page up | cmbPageUp | PgUp | ^R |
Page down | cmbPageDown | PgDn | ^C |
Beginning of line | cmbLineStart | Home | ^Q-S |
End of line | cmbLineEnd | End | ^Q-D |
Top of window | cmbFirstLineInScreen | ^Q-E | ^Home |
Bottom of window | cmbLastLineInScreen | ^Q-X | ^End |
Top of file | cmbTextStart | ^Q-R | ^PgUp |
Bottom of file | cmbTextEnd | ^Q-C | ^PgDn |
Action | Command | Key | Alternate |
---|---|---|---|
Delete the character under cursor | cmbDelChar | Del | ^G |
Delete character to left | cmbBackSpace | Backspace | ^H |
Delete line | cmbDelLine | ^Y | |
Delete to end of line | cmbDelEnd | ^Q-Y | Shift+^Y |
Delete to start of line | cmbDelStart | ^Q-H | |
Delete word at left | cmbDelWord | ^T | |
Delete word at right | cmbDelPrevWord | ^Backspace | |
Insert line | cmbNewLine | Enter | ^N |
Insert mode on/off | cmbInsMode | Ins | ^V |
A block is a selected portion of the text. You can copy, delete, etc. blocks of text. The associated commands are:
Action | Command | Key | Alternate |
---|---|---|---|
Move to beginning of block | cmbGoBeginBlock | ^Q-B | |
Move to end of block | cmbGoEndBlock | ^Q-K | |
Set beginning of block | cmbStartSelect | ^K-B | |
Set end of block | cmbEndSelect | ^K-K | |
Hide/Show block | cmbHideSelect | ^K-H | |
Mark line | cmbMarkLine | ^K-L | |
Mark word | cmbMarkWord | ^K-T | |
Delete block and copy it to the Clipboard | cmbCut | ^K-Y | Shift+Del |
Copy the selected block | cmbCopyBlock | ^K-C | |
Move block | cmbMoveBlock | ^K+V | |
Copy to Clipboard | cmbCopy | ^Ins | |
Delete block | cmbClear | ^Del | |
Paste from Clipboard | cmbPaste | Shift+Ins | |
Read block from disk | cmbReadBlock | ^K-R | Shift+^R |
Write block to disk | cmbWriteBlock | ^K-W | Shift+^W |
Replace the block by the Clipboard block | cmbReplaceSelect | Shift+^Ins | |
Convert to Uppercase | cmbToUpper | ^K-M | |
Convert to Lowercase | cmbToLower | ^K-O | |
Report the length of the block | cmbSelLength | ^Q-L |
There are two block modes. One is the mode that the old editor of RHIDE used. This mode is used in CUA programs. The other is called Persistent Blocks.
In the normal mode each time you select a block and then insert anything in it (with cmbPaste
or by typing anything) the selected block is deleted and is replaced by the new text.
In persistent blocks the selection is not replaced and is not lost when you move the cursor. From this comes the name "Persistent." In this mode you can use cmbMoveBlock
and
cmbCopyBlock
without using the Clipboard. In addition you can apply indentations to the block (
Indentation ), search only inside it, etc. That's what makes this mode much more powerful than the former. If you really like to replace the selected text by the
selection of the Clipboard, that's the default behaviour of the first mode. You can use the cmbReplaceSelect
command to achieve the same in the Persistent Blocks mode.
The described commands for selecting a block, cmbStartSelect
and cmbEndSelect
, are good but not so quick. There are other ways to do this.
Using the mouse you need only point to the start place, hold the left button pressed and move the mouse to the end point of your block.
To select a word with the mouse just double click on it.
Using the Shift key you only need to move the cursor to the start point, hold Shift pressed and move the cursor to the end point with any of the available cursor commands. Cursor movement .
You can indent or unindent a block of text using various commands, but you must keep in mind that for now the editor is limited in the following: Real Tabs If you are using tabs to indent your text, don't mix the tabs with spaces and, if you are using spaces to indent, don't mix them with real tabs.
Action | Command | Key | Alternate |
---|---|---|---|
Indent block one position adding a space | cmbIndentBlkOne | ^K-I | Shift+^I |
Unindent block one character - not an x position | cmbUnIndentBlkOne | ^K-U | Shift+^U |
Indent block | cmbIndentBlk | ^K-Tab | |
Unindent block | cmbUnIndentBlk | ^K-Shift+Tab | |
Smart Indent block | cmbSmartIndent | ^Tab | |
Smart Unindent block | cmbSmartUnIndent | Shift+^Tab |
cmbUnIndentBlkOne
unindents deleting one char at the start of the line so if the line is indented with tabs the line will retract one tab.
cmbIndentBlk
acts according to the mode. If you are using tabs, the editor will put one tab beginning each line. If you aren't using tabs the editor will operate the Tab command
on the first line and then will use this amount of indentation on the entire block.
Real Tabs .
cmbUnIndentBlk
acts according to the mode too.
Real Tabs mode. This is just like cmbUnIndentBlkOne
deleting one tab but if you don't use tabs the editor uses Backspace on the first used column of the
first line of the block and unindents by the resulting amount all the block.
cmbSmartIndent
and cmbSmartUnIndent
indents taking as reference the { } pair where the block is, for example:
{ line1 line2 line3 }After indenting a block that contains line1 to line3 is:
{ line1 line2 line3 }The indentation is made with spaces.
The editor includes a mode where you can select a rectangular portion of the text and copy, cut, clear, paste, move, etc. this region. This tool is very useful for modifications on columns.
Attention! The selected area is based on the X,Y coordinates. For this reason if you insert lines before the bottom of the rectangle the area won't be moved. I don't plan to move the area by now because that takes some CPU and I think that this selection is made just before using it. So don't report that like a bug. That is the way it works!
Action | Command | Key |
---|---|---|
Set beginning of block | cmbSelRectStart | ^K-Shift+B |
Set end of block | cmbSelRectEnd | ^K-Shift+K |
Hide/Show block | cmbSelRectHide | ^K-Shift+H |
Delete block and copy it to an special Clipboard | cmbSelRectCut | ^K-ShiftT |
Move block | cmbSelRectMove | ^K+Shift+V |
Copy to special Clipboard | cmbSelRectCopy | ^K-Shift+C |
Delete block | cmbSelRectDel | ^K-Shift+L |
Paste from special Clipboard | cmbSelRectPaste | ^K-Shift+P |
This feature is very useful to save some keystrokes. With this feature you can create a lot of shortcuts to make your life easiest.
So now, What's a Pseudo Macro? It's like a macro but is triggered by two things: one the text behind the cursor, and two the ^Space combination.
And what's the result? The result is customizable, but by default there are some predefined behaviours. For example, type in a new window the following two characters: #i and the press ^Space ...
Surprised? I bet! You got: #include <.h> and the cursor just in the right place to write the name of the header.
Now write the famous stdio word, press End, then Enter twice and now write the following two letters: ma and then ^Space ...
I bet this time you predicted better what will happen so you aren't so surprised ;-). Anyways, Surprise! Now type pr and the magic keys and ... printf("");
appears. To
end the happy history now type "Hello world!" That's all. You wrote the hello world program at a very good speed.
Now you know what I mean when I say pseudo-macros.
Is very easy. There is a file called pmacros.txt
in the same directory where you started the editor. This file contains the definitions for each pseudo macro. You can define up to 32
pseudo macros. (Please tell me if that isn't enough. I can sacrifice some time to make it dynamically expandable).
Important: If you are using RHIDE consult the RHIDE documentation to know where is stored the pmacros.txt
file.
The pmacros.txt
file is a very good example and is self-explained but I'll include here one example.
Trigger: "i(" Mode: 0,1,0,1,0 "if (@0)\n" " {@1\n" "}\n" "\belse\n" " {@2\n" "}"The Trigger keyword defines the two letter behind the cursor that will trigger the pseudo macro.
The Mode keyword indicates the mode that the editor will use when inserting the text. The modes are:
The rest is the code to insert surrounded by ". You can use \b to indicate backspace, \n for newline and \\ to indicate a simple \. If you want to insert a @ you have to type it twice @@, because this character has a special meaning (see below).
After the insertion the cursor is positioned in the place marked with @0. Don't forget to signal this point or the cursor will positioned at the beginning of the file. The places marked with @1, @2 and @3 are saved in the markers 7, 8 and 9.
Action | Command | Key |
---|---|---|
Autoindent mode on/off | cmbIndentMode | ^O |
Find place marker | cmbGotoMarkn | ^Q n* |
Set marker | cmbPutMarkn | ^K n* |
Search the open curly bracket where the cursor is | cmbSearchStart | ^[ |
Search the close curly bracket where the cursor is | cmbSearchEnd | ^] |
Search the ( where the cursor is | cmbSearchOpPar | Shift+^9 |
Search the ) where the cursor is | cmbSearchClPar | Shift+^0 |
Search the [ where the cursor is | cmbSearchOpCor | Shift+^[ |
Search the ] where the cursor is | cmbSearchClCor | Shift+^] |
Undo | cmbUndo | Alt+Backspace |
PMacro's Trigger | cmbExpandCode | ^Space |
Goto Line | cmbGotoEditorLine | ^J |
Set the options of the current window (Not in RHIDE) | cmbSetLocalOptions | Alt+L |
Set the default options (Not in RHIDE) | cmbSetGlobalOptions | Alt+G |
Convert all tabs in spaces | cmbExpandAllTabs | From menu |
Compact the text using tabs | cmbCompactBuffer | From menu |
Start recording a macro | cmbRecordMacro | Shift+F10 |
Stop recording a macro | cmbStopMacro | Alt+F10 |
Play a macro | cmbPlayMacro | ^F10 |
There are three associated files with the keyboard: keybind.cc
, keybind.h
and keybind.txt
.
The keybind.h
file contains the scancode number for each key and the number for each command. This file is included in keybind.cc
. If your keyboard has a different
assignment you can modify the scancodes in this file. For example, the German keyboards have the following sequence QWERTZ instead of the more common QWERTY, so each time you press ^Z you
won't get the function of ^Z, but of ^Y instead. Some can live with it, but it is fixable. To fix this problem you must correct the scancode of the Z key in
keybind.h.
But of course this affects ^Y too. That's only an example. The most common problem is with the symbols like {, }, [, etc. If you don't know how to find the
scancode of a key take this link.
A program to find the scancodes .
The keybind.cc
file is a C++ file but you don't need to compile it and link with the editor, you only need to preprocess it (I used this method to save the time needed to make a parser
;-).
To convert this file into a keybind.txt
file use:
cpp keybind.cc > keybind.txtWhere cpp is the C++ preprocessor. Any of the comment systems available in C++ may be used, as cpp does not pass comments to the file,
keybind.txt
. I guess that you have cpp, but if not
you can use another C++ preprocessor or even Clipper using the /P switch. There are seven sections in the file, each section is the keyboard translation under certain situation. The sections are:
Section | Contains |
---|---|
NORMAL | The single key press. |
SHIFT | Shift plus a key. |
CONTROL | Control plus a key. |
SHIFT_CONTROL | Shift plus control plus a key (all at the same time). |
QUICK | A key after ^Q or whatever you defined as cmbCtrlQ . |
BLOCK | A key after ^K or whatever you defined as cmbCtrlK . |
BLOCK_SHIFT | A key plus shift after ^K or whatever you defined as cmbCtrlK . |
Scan_Code = CommandNot all the scan codes are available. For example, you can't put:
RSHIFT = xxxxxLook at
keybind.h
to see the scan codes labels and the available commands.
Important Notes:
I)
The new assignments are mixed with the original assignments, so if you want to eliminate an old key combination you *must* assign a cmbNoCommand
to this key. For example, if you
wish eliminate the BACKSPACE key and use the ^O instead, you must:
BACKSPACE = cmbNoCommand
in the NORMAL section,
O = cmbBackSpace
in the CONTROL section.
II)
If some thing is wrong in the keybind.txt
file the program will inform the error. The line number is for the keybind.txt
and not for keybind.cc
(Sorry, some day
I'll fix this).
To continue the execution press any key and then enter.
III) You can repeat an assignment, the last wins:
BACKSPACE = cmbNoCommand BACKSPACE = cmbBackSpaceThe first is ignored.
The author of the following program is Robert Höhne:
#include#include #ifdef __GNUC__ #include #define REGS __dpmi_regs #define INTR(nr,r) __dpmi_int(nr,&r) #else #define REGS union REGS #define INTR(nr,r) int86(nr,&r,&r) #endif unsigned short getshiftstate(void) { REGS r; r.h.ah = 0x12; INTR(0x16,r); return r.x.ax; } unsigned short gkey(void) { REGS r; r.h.ah = 0x10; INTR(0x16,r); if (r.h.al == 0xE0 || r.h.al == 0) r.h.al = 0; return r.x.ax; } main() { unsigned short key; do { key = gkey(); fprintf(stdout,"%04X, scan code: %d\n",key,key>>8); } while (1); }
The settings are:
This setting controls if the typed characters are inserted in the buffer or if the typed characters replace the original ones. Insert and Delete for a detailed explanation.
This setting controls what happend when you press ENTER. If this setting is off the cursor goes to the column 1 of a new line. If the mode is on the editor will try to keep the indentation of the code inserting spaces.
This setting controls what happend when you press TAB. If this setting is on the editor will insert an ASCII 9 in this place.
An ASCII 9 is a TAB, that means that the width of this char is enough to move the cursor to the next tabulator column. In the editor the tabulator columns are equidistant and the positions are controled by the Tab Size value.
If this setting is off the editor won't put any ASCII 9 in your text. Instead the editor will insert enough spaces to move the cursor to the next tab-stop, but if the cursor is positioned on the first character of the line the editor will insert spaces to move this position to the next hole in the line over the line where you are. Confused? Sorry for my English, an example will clarify that:
This is a line over the line where you are
That's the line where the cursor is, at column one.
After pressing TAB you'll get:
This is a line over the line where you are
That's the line where the cursor is, at column one.
Now can ask, Why this? Is that useful? The answer is that's very useful to keep your code indented, experiment using that under a line with if, for, etc.
Another thing controlled by this setting is the behaviour of the Backspace key, when the setting is on the key is just 'delete the character at the left of the cursor' but when the setting is off this key is the 'Unindent' key.
The editor is much more coherent when you choose to use TABs or not use TABs. If you mix the two modes you'll get some unexpected things, specially in the indentation of the blocks.
Is better if you always work in a file without tabs and with this setting off. Then you'll get much from the editor.
Now you can say: But I really need tabs because I'm editing a make file! or because I'll send the file using an ultra slow link and I want the compression granted by the tabs. In these cases you can first expand all tabs, then work without real tabs and at the finish of your work compact all possible spaces and generate a file with all the tabs that you need. Miscellaneous section.
This setting controls the behaviour of the selected area.
Block modes chapter for a detailed explanation.
This mode was designed to be used in jointly with the Pseudo Macros ( Pseudo Macros ) and the Real Tabs mode in off to achieve an easy way to indent the code making a better work than the Autoindent mode.
In this mode the spaces inserted after pressing ENTER depend on the first word in the last line. For example, if you have:
if (a==b)_With the cursor in the '_' position and press ENTER you'll get:
if (a==b) _Now you can do either of two things, 1) press space and write the code that will be executed by the if, or 2) if this a multiline code press { and ENTER. In the last case you'll get:
if (a==b) { _Now type your first line of code. Press ENTER. Write your next line. Press ENTER again and then Backspace:
if (a==b) { 1st line; 2nd line; _Now type } and press ENTER one more time:
if (a==b) { 1st line; 2nd line; } _As you can see the code is perfectly aligned without a significant work on your part.
I tried to make this mode as smart as possible, but needs more work. If you have suggestions contact me.
Another important thing is that this indentation has a personal style, my style ;-), so maybe you don't like it. If that's your case you can do the following things:
pmacros.txt
file (
Pseudo Macros ), actually this file is coherent with this mode.
I'll try to describe the behaviour of the mode:
Each time you press ENTER the editor inserts a rn string in your text, after that the editor searchs one line located above the new line that has at least one character inside. This line is taken as reference. The editor analyzes this line searching for:
If the line contains } at the start the editor will go to the same column of the } and then will perform a Backspace. If the Real Tabs mode is off, that's an unindent. Real Tabs .
If the line starts with a C++ comment the effect is the same as in Autoindent. Autoindent .
If the line starts with a C comment the editor will try to skip the comment and analyze the rest of the line, but if the comment doesn't end on this line the editor will go to the column where the / is.
If the line starts with / the editor goes to this column.
If the line has more ( than ) the editor will go to the column of the first non-blank after the first (.
If the line has more ) than ( the editor will search the line where the number of parentheses is balanced, then will analyze this line. If the whole line still generates an unbalanced situation the editor will go to the first used column in the line that was found the first time. But if this line lets all balanced the editor will take the first word on the line and will use it as reference.
At last, and according to the word found, the editor will use this word as reference. The editor recognises the following keywords:
Keyword | Action |
---|---|
do | +2 but not if ... ; |
if | +2 but not if ... ; |
for | +3 but not if ... ; |
else | +2 |
case | +5 |
while | +2 but not if ... ; |
switch | +2 |
break | unindent |
return | unindent |
default | +5 |
Seasick? Do you have more examples? .
Note: Some of these features were added in v0.2.14 of the editor based on a suggestion of
Well here are some examples. I used a strange convention, like this: if I say type `a[ENTER]{', type the letter a, then press ENTER and finally press the `{' key.
Example 1: (Is similar to one explained before but is to show the convention)
Type: if (a==1)[ENTER]{[ENTER]a=2;[ENTER]b=3;[ENTER][BACKSPACE]}[ENTER] You'll get: if (a==1) { a=2; b=3; } [<--- cursor here]Example 2: A switch/case example
Type: switch(a)[ENTER]{[ENTER]case 1:[ENTER]a=2;[ENTER]b=3;[ENTER]break; [ENTER]case 2:[ENTER]b=5;[ENTER]break;[ENTER][BACKSPACE]}[ENTER] You'll get: switch(a) { case 1: a=2; b=3; break; case 2: b=5; break; } [<--- cursor here]Example 3: A call to a function that takes a lot of parameters
Type: printf([SPACE]"Num: %d",[ENTER]a[SPACE]);[ENTER] You'll get: printf( "Num: %d", a ); [<--- cursor here]Example 4: A lot of parentheses
Type: if[SPACE]([SPACE](a==1)[SPACE]||[ENTER](b==2)[SPACE]||[ENTER] c[SPACE])[ENTER] You'll get: if ( (a==1) || (b==2) || c ) [<--- cursor here]Example 5: Comment trying to interfere part 1
Type: /*-a-*/for[SPACE](x=1;x;--x)[ENTER] You'll get: /*-a-*/for (x=1;x;--x) [<--- cursor here]Example 6: Comment trying to interfere part 2
Type: for[SPACE](x=1;x;--x);[SPACE]//-b[ENTER] You'll get: for (x=1;x;--x); //-b [<--- cursor here]Note: Of course you can fool the editor but as you can see it is relatively smart ;-).
This setting enables an strange feature of the editor, when this mode is on the column where the cursor is is highlighted. This feature is very good to check if some part of your code is aligned.
If you like this mode but is very uncomfortable to use all the time contact me and if I get enough feedback I'll put this thing in a key to be turned on/off quickly.
This setting is similar to the column cursor but acts on the row where the cursor is. If you enable the two modes you'll get a cross on the screen showing where the cursor is.
This mode acts showing the pairs of (/), [/] and {/} on the fly. Each time you type one of these symbols the editor will search the matching pair, if the editor finds it and the match is on the screen both will be highlighted, if the match is outside the screen the editor will inform the position on the status line, and if there is no match the editor will inform the situation in the status line too.
That's very useful when you are typing complex parenthetical expressions or a very nested code. You can use it jointly with the cmbSearchStart
, cmbSearchEnd
,
cmbSearchOpPar
, cmbSearchClPar
, cmbSearchOpCor
and cmbSearchClCor
commands. (
Miscellaneous ).
The editor can highlight the syntax of your code. The available modes are:
The C/C++ highlight was designed for the GNU C compiler.
The Pascal highlight was designed for the GNU Pascal compiler.
The Clipper highlight was designed for the CA-Clipper 5 compiler.
The sources of the
calculator are in the parser.c
file. They are free and you can use it for any purpose.
Here is the documentation of the calculator written by Laszlo:
The purpose of this program, to provide a simple but powerful 'calculator' for programmers, to help with coding and debugging, where GDB's expression evaluator is not enough.
You may say "Hey, I can write a better one with flex and bison", and you may be right. I can make a better one too. But it'll be 4-5 times longer! This calculator is only 10 kbytes of C code. What I think? It's not that bad.
The parser algorithm I use is called 'Operator Precedence Parsing' (I translated this from Hungarian, so I may be wrong ;-). It works with 'operator precedence grammars' (a subset of LR(1) grammars), which means that there can't be two non-terminating tokens next to each other on the right side of your grammar rules. It's ideal for expression evaluation.
With this parser you can use numbers, operators, parentheses and functions like in C.
Here are the operators in decreasing precedence:
~
unary not -
unary minus
**
power
*
multiplication /
division %
modulo
+
plus -
binary minus
<<
shift left >>
shift right
&
logical and
^
logical xor
|
logical or
The calculator includes the following functions: sin, cos, tan, sinh, cosh, tanh, asin, acos, atan, log, log10, exp, abs, sqrt, ceil and floor. They work as you expected. Additionally the calculator provides some radix conversion routines: bin, oct, dec and hex.
The calculator uses doubles, but you can use numbers in the usual integer formats also: 0x... for base 16, 0b... for base 2 and 0... for base 8. The result of the calculation is displayed as a double for base 10, and converted to long long format for the other radixes.
Error codes:
Number | Meaning |
---|---|
-1 | yylex() failed e.g. 1+# |
-3 | yyparse() failed e.g. 1+2) |
-4 | floating point exception e.g. sqrt(-1) |
If you find a bug please contact me, the Undo thing is the most complex one and I know that needs some work on it.
When reporting a bug please don't tell me: Some times some strange thing happend ... Try to find a pattern to the problem. What situation triggers the problem? ... with which file(s)? Then send me the file and the description.
When sending to me a file UUEncode it to avoid problems related with the e-mail.