For this html
<!DOCTYPE HTML PUBLIC "-//W3C//DTD HTML 4.01 Transitional//EN" "http://www.w3.org/TR/html4/loose.dtd">
<html lang="cs">
<head>
<meta http-equiv="content-type" content="text/html; charset=windows-1250">
<meta name="generator" content="SynWrite editor">
<title>Untitled Glossary</title>
</head>
<body>
<h1>
<a id="GlossTop">Glossary of Terms</a>
</h1>
<hr>
<!-- Glossary Toolbar -->
<textarea readonly disabled ></textarea>
<h3>
<button type="submit"></button>
<a href="#GlossA >A</a>
<a href="#GlossB">B</a>
<a href="#GlossC">C</a>
<a href="#GlossD">D</a>
<a href="#GlossE">E</a>
<a href="#GlossF">F</a>
<a href="#GlossG">G</a>
<a href="#GlossH">H</a>
<a href="#GlossI">I</a>
<a href="#GlossJ">J</a>
<a href="#GlossK">K</a>
<a href="#GlossL">L</a>
<a href="#GlossM">M</a>
<a href="#GlossN">N</a>
<a href="#GlossO">O</a>
<a href="#GlossP">P</a>
<a href="#GlossQ">Q</a>
<a href="#GlossR">R</a>
<a href="#GlossS">S</a>
<a href="#GlossT">T</a>
<a href="#GlossU">U</a>
<a href="#GlossV">V</a>
<a href="#GlossW">W</a>
<a href="#GlossX">X</a>
<a href="#GlossY">Y</a>
<a href="#GlossZ">Z</a>
</h3>
<hr>
<!-- A -->
<h1><a id="GlossA"> A </a></h1>
<dl>
<dt><a id="Algol"><b>Algol</b></a>
<dd>Algorithmic Oriented Language.
</dl>
<small><a href="#GlossTop">Top</a></small>
<hr>
some plain text here.
<!-- B -->
<h1><a id="GlossB"> B </a></h1>
<dl>
<dd>
</dl>
<small><a href="#GlossTop">Top</a></small>
<hr>
<!-- C -->
<h1><a id="GlossC"> C </a></h1>
<dl>
<dd>
</dl>
<small><a href="#GlossTop">Top</a></small>
<hr>
<!-- D -->
<h1><a id="GlossD"> D </a></h1>
<dl>
<dd>
</dl>
<small><a href="#GlossTop">Top</a></small>
<hr>
<!-- E -->
<h1><a id="GlossE"> E </a></h1>
<dl>
<dd>
</dl>
<small><a href="#GlossTop">Top</a></small>
<hr>
<!-- F -->
<h1><a id="GlossF"> F </a></h1>
<dl>
<dd>
</dl>
<small><a href="#GlossTop">Top</a></small>
</body>
</html>
plugin finds only these words
['body', 'button', 'content', 'disabled', 'equiv', 'head', 'href', 'html', 'http', 'lang', 'link', 'meta', 'name', 'readonly', 'rel', 'script', 'small', 'src', 'textarea', 'title', 'type']
Ie words from tags.
Wanted: plugin finds words out of html tags, ie plain text.
Maybe also words from HTML comments.
Plg must handle HTML specially.
So a separate options needed like no_strings and no_comments, but only for html? Am I understanding correctly?
Not separate options, I suggest to find text words in HTML by special algorithm- ignoring lexer stuff, parse HTML, and find plain text.
(all lexers which names begin with 'HTML')
Isn't that what those options do? With them turned off I get all the words from the html doc
but then, autocompletion for <a h|ref suggests something, and <b| suggests something. the idea - that plugin must not mess with usual HTML autocompletion.
So for html.* in strings and comments autocomplete from strings and comments, and outside - autocomplete without strings and comments?
In strings - do or don't autocomplete (not sure why we need to complete in "val")
In text out of tags - do
In comments - do
Others- skip plugin's work
Sorry for the pause. Let us continue? :)
The terminology is not clear to me... am I understanding correctly, plugin needs to auto-complete only in marked places with all the words from the whole document?

picture is OK. it has 2 marks -
a) position of plain text. (my terminology).
b) position in comment.
on autocomplete, plg must do this-
a) don't work if inside tags (pass this work to Pascal auto-complete)
b) if outside tags, or if in comment, (your 2 marks), work: scan all document outside of tags, and find all words. suggest found words to user.
@eltonfabricio10 @miroslavmatas @JairoMartinezA @ildarkhasanshin
New plugin is made, to solve the issue.
It can work together with 'Complete from text' and with builtin HTML completion.
'Complete HTML Text'.
Pls see if interested, report bugs in new topics.