about selection

Alan Kilborn

It might be better to express regexes like this (inside of backticks) instead of the way you did it:

(?s)[^#]+\R\K.*?(?=\R[^#])

Whenever I see regexes with italics in them here, I think, uhoh, a * got consumed. Note that this is NOT the case for your posting (the regex is fine!), but the wacky italics and color is still “disturbing”.

Just a suggestion.

cisco779k

tanxs to both of them for your reg-ex.
however i don’t understand why to use reg-ex when on this simple case just enter in search field (or mark) ### …

Alan Kilborn

@astrosofista said in about selection:

here it goes another regex, but this one only selects the hashes ###,

Also, your choice of terminology here (“selects”) probably only confuses the OP. Probably the word “finds” would have been the better way to go. :-)

cisco779k

no confusion. my initial intent (clearly expressed) was a selection of the text from the ### fields.
but since this selection is not possible, then i use the mark to make a copy & paste.

astrosofista

@Alan-Kilborn

Now I see —aging eyes here—, thank you, didn’t noticed till now that italics being applied after the * — maybe we need an escape character for avoid that, ha ha.

OTOH, I like the message structure or disposition, were the regex is on its own line.

Well, will see what can I do.

astrosofista

@cisco779k

Three reasons I can think of:

Error prevention. Regex provides more control than the normal mode, so I can make sure to only match what it is needed —as I haven’t checked the actual file and as I only had seen a sample—, I don’t know if those three hashes are also located in the middle of the file.
More error prevention. Regex also provides a way to override options that the user inadvertently may have checked. Thus, in the last regex (?s) forces the regex to run a multiline search.
Personal habit. In my Notepad++ the regex mode is always checked, so my first attempt is to run even simple searches in that mode. I only select the normal mode in order to avoid regex metacharacters intrusion — and get back to regex realm as soon as possible :)

Alan Kilborn

@astrosofista said in about selection:

I only select the normal mode in order to avoid regex metacharacters intrusion — and get back to regex realm as soon as possible

Or you could just stay permanently in the regex realm (Search mode = Regular expression) and just lead off any “normal” searches you need to do (because of metacharacters) with a quick \Q – no need to bother with a trailing \E ! :-)

You can’t do a Whole word search this way, though, if you ever use that option, because that checkbox is disabled (but you could go to the trouble of adding \b before the \Q and at the other end).

Alan Kilborn

@astrosofista said:

…didn’t noticed till now that italics being applied after the * — maybe we need an escape character for avoid that

I think maybe we have a misunderstanding. Your regexes in this thread are fine, I was just suggesting

this method

instead of

this method

because the second method does some lexing (does it think it is C++ above because it has special highlighting for “this”?) and the first method just leaves everything alone (and thus is probably clearer).

astrosofista

@Alan-Kilborn said in about selection:

I think maybe we have a misunderstanding. Your regexes in this thread are fine, I was just suggesting…

Oh, yes. I thought you were suggesting inline regexes —that is, regular expressions in the middle of a paragraph—, which I really found confusing unless the expression is limited to a few characters.

So it´s clearer now. I will preserve the message structure while avoiding the multicolored format.

astrosofista

@Alan-Kilborn said in about selection:

Or you could just stay permanently in the regex realm…

Hmm, it looks very interesting. Will check the documentation, thank you.

Concerning the whole word issue, it wouldn’t necessarily be a problem. Because the Matching Brace feature is not available in the Find panel, a tool that I find almost indispensable to handle complex regexes, and both the Search and Replace fields are really short and the font size is too small for my needs, I usually write expressions on the Editor and use hotkeys to import and export them, so adding one or two options may not represent much trouble.

But let me first read about it and see if I can adapted it to my ways.

guy038

Hello, @cisco779k, @alan-kilborn, @astrosofista, and All

Of course, you certainly know that, in order to insert a code block, you must use, either the syntaxes :

```
Code_Text
```

or

~~~
Code_Text
~~~

But, do you know that you can add an indication, after the opening ``` or ~~~, to force a nice visual aspect of a specific language ?

Indeed, for example, you may use ```c to display a C code block, ```xml to visualize an XML piece of code and ```py to show a nice python script !

And to get only a black foreground text, with a white background, simply use any invalid indication. I usually write the ```z syntax !

Here are, below, the different indications that I’ve found out, yet, from the language names and extensions used :

    •---------------•--------------•---------------•
    |   LANGUAGE    |  MAIN form   |  OTHER forms  |
    •---------------•--------------•---------------•
    |  bash         |  sh          |  bash         |
    |  c            |  c           |  h            |
    |  c++          |  cpp         |  hpp          |
    |  c#           |  cs          |               |
    |  CSS          |  css         |               |
    |  DIFF         |  diff        |  patch        |
    |  HTML         |  html        |  xhtml        |
    |  HTTP         |  http        |  https        |
    |  ini file     |  ini         |               |
    |  javaScript   |  js / jsx    |  java         |
    |  JSON         |  json        |               |
    |               |  jsp         |               |
    |  Makefile     |  mk / mak    |  makefile     |
    |  Objective-C  |  mm          |  objectivec   |
    |  Perl         |  pl / pm     |  perl         |
    |  php          |  php         |  php3 / php4  |
    |  Python       |  py          |  python       |
    |  Ruby         |  rb          |  ruby         |
    |  SQL          |  sql         |               |
    |  XML          |  xml         |  xsl          |
    •---------------•--------------•---------------•

Note that I’m not sure, exactly, to which languages ( Java, JavaScript (embedded), JavaScript ) refer the indications js and jsp !?

Best Regards,

guy038

PeterJones

@guy038 said in about selection:

I usually write the ```z syntax

I always wondered why the z was included, and sometime between last year and now, I had dropped the z to become just ``` alone. Now that I know the purpose of the z (and the other options), I’ll have to start using those.

And if you use the </> button to fomat a block of text, you can still go back above and add in the option; for example, here I added pl to the ``` line after I generated it using </>.

my $default;
open my $fh, '>', 'out.txt' or die "Goodbye, cruel world: $!";
sub hehe { say "haha"; }
BEGIN {
    $default = Win32::Mechanize::NotepadPlusPlus::Notepad->_new();
}

yep, it properly highlights that. cool.

edits: fixed incomplete sentences/thoughts.

guy038

Hello, @cisco779k, @alan-kilborn, @astrosofista, @peterjones and All,

Oh…, I’m sorry ! I spoke a little too fast :-(( Peter, I did some tests, again, with your piece of Perl, a Python script, some HTML and C code and…, seemingly, as soon as you click on the </> icon and you paste the different pieces of code, NodeBB seems, each time, to detect correctly the language !?

Indeed, after I added the related indication, and used Ctrl + Z and Ctrl + Y, I did not be able to see any difference in the code !!

So, the table shown, in my previous post, is rather useless. The only interesting case would be if you would show for instance, a Perl script, with the Python syntax !

Could your confirm my tests ?

Cheers,

guy038

astrosofista

@guy038, @cisco779k, @alan-kilborn, @PeterJones, All

Thank you for sharing this interesting and useful info. And also can be mixed!

I played a bit with the available options and selected two alternative ways (js and mk) to the usual one to display regexes:

Search: (?s)[^#]+\R\K.*?(?=\R[^#])
Replace: [leave empty]

Search: (?s)[^#]+\R\K.*?(?=\R[^#])
Replace: [leave empty]

Search: (?s)[^#]+\R\K.*?(?=\R[^#])
Replace: [leave empty]

Both alternative displays are not multicolored and don’t mix font styles, so I think are fine. And aiming to create some contrast against text paragraphs, I am currently not sure but probably will use the second one. I hope @alan-kilborn won’t have strong opinions against it, ha ha.

@guy038, let me ask you a question: how do you create those nice tables? Is a free tool? I’m curious.

Best Regards

PeterJones

@guy038,

Could your confirm my tests ?

auto-identify:

my $default;
open my $fh, '>', 'out.txt' or die "Goodbye, cruel world: $!";
sub hehe { say "haha"; }
BEGIN {
    $default = Win32::Mechanize::NotepadPlusPlus::Notepad->_new();
}

force ```pl

my $default;
open my $fh, '>', 'out.txt' or die "Goodbye, cruel world: $!";
sub hehe { say "haha"; }
BEGIN {
    $default = Win32::Mechanize::NotepadPlusPlus::Notepad->_new();
}

force ```py (purposefully wrong syntax)

my $default;
open my $fh, '>', 'out.txt' or die "Goodbye, cruel world: $!";
sub hehe { say "haha"; }
BEGIN {
    $default = Win32::Mechanize::NotepadPlusPlus::Notepad->_new();
}

force ```z (blank)

my $default;
open my $fh, '>', 'out.txt' or die "Goodbye, cruel world: $!";
sub hehe { say "haha"; }
BEGIN {
    $default = Win32::Mechanize::NotepadPlusPlus::Notepad->_new();
}

indent

my $default;
open my $fh, '>', 'out.txt' or die "Goodbye, cruel world: $!";
sub hehe { say "haha"; }
BEGIN {
    $default = Win32::Mechanize::NotepadPlusPlus::Notepad->_new();
}

summary

yes, it auto-identifies if there’s nothing. But there are times when the text is ambiguous, or truly plaintext, so using the hardcoded filetype will enforce proper syntax highlighting in the forum.

ps

edit: sorry, @astrosofista , for hijacking your thread with this tangent.

Alan Kilborn

@astrosofista said in about selection:

I hope @alan-kilborn won’t have strong opinions against it, ha ha.

Well… :-)

Do it the way you want, but IMO this would be a more proper way, from your examples:

Search: (?s)[^#]+\R\K.*?(?=\R[^#])
Replace: [leave empty]

It doesn’t muddy what one would copy for the search regex, and since “[leave empty]” isn’t a regex it shouldn’t be specially highlighted.

But, really, it is all pointless. Whatever gets your point across.

Lately my ideal way is this:

Find what box: (?s)[^#]+\R\K.*?(?=\R[^#])
Backward direction checkbox: unticked
Match whole word only checkbox: unticked (and disabled)
Match case checkbox: unticked
Wrap around checkbox: unticked
Search mode radiobutton: Regular expression

…because I have a script that grabs that data right off the Find window, and formats it.

astrosofista

@PeterJones said in about selection:

edit: sorry, @astrosofista , for hijacking your thread with this tangent.

Nothing to be sorry about, and for sure the info is quite relevant, so I am paying attention.

astrosofista

@Alan-Kilborn

I get your points and I think they are quite reasonable. However and unfortunately I don’t clearly see the reddish text inside backtickets displayed on a black background —my better option—, so I feel the need to look for other color(s) to easy my reading.

Now, concerning your ideal way of extracting data from the Find window, I like the idea and depending on how much use I could give it, I will probably develop my own, thank you for the inspiration! :-)

Best Regards

Terry R

@astrosofista said in about selection:

reverse all the lines between hashes and only those

@cisco779k said:
@ astrosofista
you understand what I want to do

@cisco779k
I’m aware this problem has largely been solved but thought it an interesting enough exercise to see if another method could be obtained by keeping the file intact and not having to use a plugin or export lines to another file. I think I’ve come up with a solution. It’s not original, indeed I knew of anoher solution employing this technique provided by @guy038 some time ago to pad out numbers (right justify).

So my regex is
Find What:(?-s)(^[^#]{3}.+\R)?(^[^#]{3}.+\R)?(^[^#]{3}.+\R)?(^[^#]{3}.+\R)?(^[^#]{3}.+\R)?(^[^#]{3}.+\R)?(^[^#]{3}.+\R)?(^[^#]{3}.+\R)?(^[^#]{3}.+\R)
Replace With:\9\8\7\6\5\4\3\2\1

So at the moment it cannot work with more than 9 lines grouped together, however the concept can be expanded to more lines. Each capture group will store a line if AND only if the lines doesn’t start with 3 # characters. Then the Replace With field writes them back in reverse order. As some capture groups will be empty there will be nothing to write.

The only proviso I have is that there must be (in the current form of the regex) a blank line at the bottom of the file. This allows the last line to be captured as part of a group.

my 2cents worth
Terry

astrosofista

Hi @Terry-R, All

Clever idea, I like it. And what I like even more is that it is easy to communicate and to apply, just one step, a S/R replacement. My approach isn’t original as well, and it is really complex, as it involves the BetterMultiSelection plugin, Column Editor, and a macro that mechanizes a regex and a sort to reverse the lines.

However, your method has a curious drawback, as you constraint it to only 9 groups. The good news is that there is no reason to stop here.

Don’t know if I’m guessing too much, but I suspect you were misguided by a syntactic issue, as replacements use the backslash \. I don’t remember where I learn about it, but it is possible to create up to 99 replacement groups, as long as the dollar $ sign is used in them. Hope I am not wrong about this, @guy038 surely knows all the details.

As a test, just add two more groups to the Search field, and enter ($11)($10)$9$8$7$6$5$4$3$2$1 in the Replace box. Worked nice in this sample:

###
one one one one
two two two two
three three three three
four four four four
five five five five
six six six six
seven seven seven seven
eight eight eight eight
nine nine nine nine
ten ten ten ten
eleven eleven eleven eleven
###
eleven eleven eleven eleven
ten ten ten ten
nine nine nine nine
eight eight eight eight
seven seven seven seven
six six six six
five five five five
four four four four
three three three three
two two two two
one one one one
###

Best Regards