A minor issue with the TRichMemo operation.

[jianwt]

I want to manipulate text strings in RichMemo1 and have two small questions to ask everyone:

When richmemo1.scrollbars =ssBoth, how can the total number of paragraphs in RichMemo1 be obtained instead of richmemo1.lines.count? In the following routine, there are 4 paragraphs instead of 6.

2. There are 4 paragraphs in the routine. How can the text of odd-numbered paragraphs be displayed in red and that of even-numbered paragraphs in green?

I have no idea how to handle the above-mentioned issues.

Thank you all for your help.

[rvk]

1) For the number of paragraphs...
You can put the content of the rtf in a string, which eliminates the linewrap, and then count the linebreaks.

Code: Pascal  [Select][+][-]

1. var
2.   ParaCount: Integer;
3.   TextContent: string;
4. begin
5.   TextContent := RichMemo1.Text;
6.   ParaCount := 0;
7.   if TextContent <> '' then
8.     ParaCount := Length(SplitString(TextContent, sLineBreak));

Keep in mind that this can count the last linebreak as a paragraph.
Also... if you have blanc lines you would need to change sLineBreak + sLineBreak to a single sLineBreak first.

2) For coloring the odd numbered paragraphs you would really need to parse through the code and apply coloring to the paragraph yourself. There is no standard way of doing this in RichMemo/RichEdit.

[paule32]

isn't TRICHMemo not bound to win32API ?

there could be give a API Function for it ...

[Thaddy]

No. It is cross-platform
https://wiki.lazarus.freepascal.org/RichMemo#:~:text=The%20easiest%20way%20to%20install%20the%20stable%20version,by%20this%20GitHub%20link.%20Download%20the%20package%20set.

[jianwt]

@rvk
Thank you for your answer. However, it seems that the code does not take the total number of paragraphs and returns 1

Code: Pascal  [Select][+][-]

1. var
2.   ParaCount: Integer;
3.   TextContent: string;
4. begin
5.   TextContent := RichMemo1.Text;
6.   ParaCount := 0;
7.   if TextContent <> '' then
8.     ParaCount := Length(SplitString(TextContent, sLineBreak));
9.  

[rvk]

Thank you for your answer. However, it seems that the code does not take the total number of paragraphs and returns 1

On what platform?
Maybe change the sLineBreak in #13.
Could be that on Windows RichEd20 does #13#10 as linebreak while Linux only does #13.
In both cases will counting just the #13 work.

Code: Pascal  [Select][+][-]

1. ParaCount := Length(SplitString(TextContent, #13));

[CM630]

Yes, it is only #13 for tRichMemo even on Windows.
But also ParaCount := Length(SplitString(TextContent, #13)) -1; because TRichMemo also has a parasite end line.

[rvk]

Yes, it is only #13 for tRichMemo even on Windows.
But also ParaCount := Length(SplitString(TextContent, #13)) -1; because TRichMemo also has a parasite end line.

Grr  Yes. SplitString splits on ALL delimiters/characters. The parameter is not a splitstring. So passing sLineBreak (#13#10) will split on both #13#10. I missed that  (that's why it worked here when I tested it)

[CM630]

As for colouring, I use:

Code: Pascal  [Select][+][-]

1. procedure AppendMemo(RichMemo: TRichMemo; aText: String; Colour: TColor = clDefault);
2. var
3.   prms : TFontParams;
4. begin
5.   prms := GetFontParams(RichMemo.Font);
6.   if (Colour <> clDefault) then prms.Color := Colour;
7.   InsertFontText(RichMemo,aText, prms);
8.   //Next line is optional
9.   //SendMessage(RichMemo.handle, WM_VSCROLL, SB_BOTTOM, 0);
10. end;

EDIT: Actually, this seems to return the number of the paragraphs. I wonder why no -1 is required.

Code: Pascal  [Select][+][-]

1. ParaCount := string(RichMemo1.Text).CountChar(#13);

[jianwt]

Thank you to all the friends above for your help. I have basically achieved my requirements. Thank you again

[paule32]

on Windows the Line-End-Mark is #13#10

[rvk]

on Windows the Line-End-Mark is #13#10

Not for riched20 and thus TRichMemo on Windows.
That's because the .Text property gets it's string from TRichEditManager.GetTextW().
In Windows this returns default #13, not #13#10.

I have looked into this problem and found the cause.
Although the TRichMemo.Lines is a TStrings it gets its .Text property from another class.
It uses TRichEditManager.GetTextW() to retrieve the complete .Text.

There a EM_GETTEXTEX is issued to get the text from the RichEdit (Windows) component.
There is a flag setting possible for getting the text with #13#10 instead of the default #13 (GTL_USECRLF).
But it isn't used (and that's the reason .Text has #13 instead of #13#10).

I find this very confusing and is definitely not like it is in the TRichEdit in Delphi.

If TStringList has only #13 on Linux systems (I don't know) then this could also be for TRichMemo.Lines.
But the TRichMemo.Lines should give #13#10 on Windows systems (and be the same as TStringList and Delphi).

As a test I edited TRichEditManager.GetTextW to include GTL_USECRLF (and that works) but I see that it must be implemented in a lot of other places in win32richmemoproc.pas too (everywhere where the TGETTEXTEX.flag is used).

I not sure why TRichMemo.Text doesn't just use TStrings.GetTextStr like Delphi does.

https://forum.lazarus.freepascal.org/index.php/topic,60956.msg457944.html#msg457944

[paule32]

is the Text a PChar ?

Because, W-Functions: #13#00#10#00

So, PChar is null-terminated ... ?

[Thaddy]

Pascal strings are also #0 terminated, because of interoperability with PChars.
(Except shortstrings)

[paule32]

what are the relevant Part of ShortStrings ?
PChar is relevant if you use Library Calls that was build with a C/C++ Compiler where the Strings can have more Characters than the 255 Limit - and terminated by #0.

So, if you use {$mode FPCOBJ}
you have to explicit add {$H+} for Huge String's...

In {$mode DELPHI}
you don't need to explicit add {$H+} which is by default.

And when H+ Option is on, FPC use AnsiString instead, which are RefCount'ed.

For me I get the Opinion, that PChar and AnsiString are nearly the same (if we don't Focus the Encoding) ?