Perl language syntax highlighting troubles (bug or limitation ?)

Ekopalypse

Don’t bother too much w/ my pb, concentrate on your job

:-) That’s for sure - we are allowed to surf the web as long as work related stuff is
priority 1 and we take this very seriously.

bit confused.

Actually me too.
If you don’t mind - would you try to do the following.

Restart npp (Only one single instance should run)
Click Plugins->PythonScript->Show Console
Click Plugins->PythonScript->Scripts->NAME_OF_YOUR_SCRIPT
-> check if there is an error reported in the console, if so - stop here and post it
If no error has been written to console
Open a perl file or a new file and change the language to PERL and put some code in it
-> check if there is an error reported in the console, if so - stop here and post it
If no error has been written to console
execute the following in the console run edit box

editor.getLexer()

Does it return 6 as specified in the script (BUILTIN_LEXER_ID = 6)?
If it does, execute the following in the console run edit box

EnhanceBuiltinLexer

What does it return?

Gilles Maisonneuve

restarted NPP, only one instance: checked
python console:

Python 2.7.15 (v2.7.15:ca079a3ea3, Apr 30 2018, 16:30:26) [MSC v.1500 64 bit (AMD64)]
Initialisation took 109ms
Ready.
Plugins/PythonScript/Scripts->EnhancePerlLexer.py: checked. Result: NOTHING in the console, no error but no confirmation that the py script as run either, is it normal?
Opened a new window, made it Perl syntaxing. . Did not do it. Here is the code for the colors in your script that I modified:

regexes[(1, 5)] = (r’\bq[rwqx]{0,1}\b([^\h]).?\1|(\bq[rwqx]{0,1}\b\h+(\w).?\3)', [0])

regexes[(2, 5)] = (r’\bq[rwqx]{0,1}\b\h*((.+?)|[.+?]|{.+?})', [0])

own defined color and non-default match group(s) used

regexes[(1, (0,0,0))] = (r’\bq[rwqx]{0,1}\b([^\h]).?\1|(\bq[rwqx]{0,1}\b\h+(\w).?\3)‘, [0])
regexes[(2, (0,0,0))] = (r’\bq[rwqx]{0,1}\b\h*((.+?)|[.+?]|{.+?})‘, [0])
regexes[(3, (0,0,0))] = (r’(?s)((<<)"(\w+?)";.?\3)‘, [2])
regexes[(4, (128,128,128))] = (r’(?s)((<<)\h+“(\w+?)”;.?\3)', [2,3])

Executed editor.getLexer(), got a return of 1; executed EnhanceBuiltinLexer, results:

Python 2.7.15 (v2.7.15:ca079a3ea3, Apr 30 2018, 16:30:26) [MSC v.1500 64 bit (AMD64)]
Initialisation took 125ms
Ready.
>>> editor.getLexer()
1
>>> EnhanceBuiltinLexer
<class '__main__.EnhanceBuiltinLexer'>

Seems that I get the wrong return code… don’t I ?

Ekopalypse

@Gilles-Maisonneuve

Seems that I get the wrong return code… don’t I ?

Indeed - would you mind executing the following in the console run edit box?

editor.getLexerLanguage()

no error but no confirmation that the py script as run either, is it normal?

yes.

Ekopalypse

@Gilles-Maisonneuve

I get 1 returned if the language is set to Normal Text.
Sorry for asking, but sure you have set the language to Perl?

Gilles Maisonneuve

@Ekopalypse

>>> editor.getLexerLanguage()
'perl'

Would it be (more) convenient for you to connect directly on my machine ? (DWservice) ?

Ekopalypse

@Gilles-Maisonneuve

Could it be that you have two views open, one with a perl document and one with a normal text document?

Gilles Maisonneuve

I have many documents opened in different Windows ([edited]TABS). Is this what you mean by “views” ?

No double Windows at a time like in “compare” or so.

Ekopalypse

@Gilles-Maisonneuve

No double Windows at a time like in “compare” or so.

Ok, then it seems we have learned that the id is not always the same.
Gimme a second to see what needs to be changed to use editor.getLexerLanguage() instead.

Gilles Maisonneuve

all right, take your time.
i’ll be there late tonight.

about DWservice, it’s GNU, hosted in DE I guess, not smthg like Teamviewer. But perhaps you are right, one should always be paranoiac.

Ekopalypse

@Gilles-Maisonneuve

I’m living in DE but have to admit, that I haven’t heard of DWservice before, shame on me :-)

Changes
line 9 to BUILTIN_LEXER = 'perl'
line 111 self.lexer_name = BUILTIN_LEXER.lower()
line 219 to self.doc_is_of_interest = True if editor.getLexerLanguage().lower() == self.lexer_name else False

As python is very picky about whitespaces make sure that you either use
spaces or tabs for indentation only. Best python practice is to set
Settings->Preferences->Language->TabSettings->Python
Tab size = 4 and check replace by space checkbox
(if this isn’t your default setting, of course)

Gilles Maisonneuve

@Ekopalypse

Oops, I must have made some mistakes but can’t find where… Here is the console log:

Traceback (most recent call last):
  File "C:\Users\gm\AppData\Roaming\Notepad++\plugins\Config\PythonScript\scripts\startup.py", line 1, in <module>
    import EnhancePerlLexer
  File "C:\Users\gm\AppData\Roaming\Notepad++\plugins\Config\PythonScript\scripts\EnhancePerlLexer.py", line 283, in <module>
    EnhanceBuiltinLexer().main()
  File "C:\Users\gm\AppData\Roaming\Notepad++\plugins\Config\PythonScript\scripts\EnhancePerlLexer.py", line 280, in main
    self.on_bufferactivated(None)
  File "C:\Users\gm\AppData\Roaming\Notepad++\plugins\Config\PythonScript\scripts\EnhancePerlLexer.py", line 237, in on_bufferactivated
    self.check_lexer()
  File "C:\Users\gm\AppData\Roaming\Notepad++\plugins\Config\PythonScript\scripts\EnhancePerlLexer.py", line 224, in check_lexer
    self.doc_is_of_interest = True if editor.getLexerLanguage().lower() == self.lexer_name else False
AttributeError: 'EnhanceBuiltinLexer' object has no attribute 'lexer_name'
Python 2.7.15 (v2.7.15:ca079a3ea3, Apr 30 2018, 16:30:26) [MSC v.1500 64 bit (AMD64)]
Initialisation took 343ms
Ready.

Traceback (most recent call last):
  File "C:\Users\gm\AppData\Roaming\Notepad++\plugins\Config\PythonScript\scripts\EnhancePerlLexer.py", line 73, in <module>
    EnhanceBuiltinLexer().main()
  File "C:\Users\gm\AppData\Roaming\Notepad++\plugins\Config\PythonScript\scripts\EnhancePerlLexer.py", line 280, in main
    self.on_bufferactivated(None)
  File "C:\Users\gm\AppData\Roaming\Notepad++\plugins\Config\PythonScript\scripts\EnhancePerlLexer.py", line 237, in on_bufferactivated
    self.check_lexer()
  File "C:\Users\gm\AppData\Roaming\Notepad++\plugins\Config\PythonScript\scripts\EnhancePerlLexer.py", line 224, in check_lexer
    self.doc_is_of_interest = True if editor.getLexerLanguage().lower() == self.lexer_name else False
AttributeError: 'EnhanceBuiltinLexer' object has no attribute 'lexer_name'

Line numbers don’t match because I already commented out some of your lines but kept them in the file, and duplicated them with my own changes. But I did do the changes at the places you told me to do them.

Ekopalypse

@Gilles-Maisonneuve

check line 111 - it defines the lexer_name

Gilles Maisonneuve

OOPS, yours :== self.lexer_name, mine :== self_lexer_name, I am really a dumb when dealing with OO programming, can’t realize that ‘self’ is the current object and of course separated by a dot.

Colour has changed for q* keywords and there text (black on dark blue, can’t read but now just need to ajust the colors).
No change for here docs, but don’t know if I properly set the colors, have to check.

Send you a screen copy in a few minutes.

Gilles Maisonneuve

@Ekopalypse

All right, nearly done: with the following regexp in your python code:

regexes[(1, (255,0,128))] = (r'\bq[rwqx]{0,1}\b([^\h]).*?\1|(\bq[rwqx]{0,1}\b\h+(\w).*?\3)', [0])
regexes[(2, (255,0,128))] = (r'\bq[rwqx]{0,1}\b\h*(\(.+?\)|\[.+?\]|\{.+?\})', [0])
regexes[(3, (0,0,0))] = (r'(?s)((<<)"*(\w+?)"*;.*?\3)', [2])
regexes[(4, (0,0,0))] = (r'(?s)((<<)\h+"(\w+?)";.*?\3)', [2,3])

I get the following colors:

q* colors OK, here docs no

Q* colors are good {well I might have an uggly taste in colors but at least they match ;-)) }

Would you have any clue about why the here docs= are still not handled properly ? They should be black, I think.

Ekopalypse

@Gilles-Maisonneuve

the regexes assumes double quotes and semicolon directly attached to EOT.
Like

print << "EOT";

--------------------- separation line ------------------

EOT

Is there a rule how this is specified?

Gilles Maisonneuve

I think I found why.
Your regexp says :
r'(?s)((<<)"*(\w+?)"*;.*?\3)'
would not it be better if :
r'(?s)(\h*(<<)\h*"*(\w+?)"*\h*;.*?\3)'

???

To answer your question:

Perl allows

<<TEXT,
<< TEXT
<<‘TEXT’ / << ‘TEXT’
<<“TEXT” / << “TEXT”

meanings differ in each case…

Ekopalypse

To be honest - I’m not a regex expert at all :-D
If you, as a perl developer, say so I would absolutely believe it is :-)

Gilles Maisonneuve

@Ekopalypse

In your Python regexp, what’s the meaning of:

“\3”
“, [2]” and “[2,3]” ?

If I can understand what I think I could translate a Perl regex code into python (for this case at least).

Ekopalypse

What about using this
(?s)((<<)\h+(["|'])(\w+?)\3\h*;.*?\4)

Ekopalypse

is the boost:regex convention to denote match group 3
and
defines which match group actually should be painted

Like if you have:

r'(word1)(word2)(word3)', [2,3]

would mean that only word2 and word3 would be painted
whereas if you would specify

r'(word1)(word2)(word3)', [0]

everything would be colored.

Does this makes sense to you?

Perl language syntax highlighting troubles (bug or limitation ?)

regexes[(1, 5)] = (r’\bq[rwqx]{0,1}\b([^\h]).?\1|(\bq[rwqx]{0,1}\b\h+(\w).?\3)', [0])

regexes[(2, 5)] = (r’\bq[rwqx]{0,1}\b\h*((.+?)|[.+?]|{.+?})', [0])

own defined color and non-default match group(s) used