CODEWITHSUNDEEP
xmlparsingxml-parsingctagsexuberant-ctags
rafmartom
rafmartom

Reputation: 1

Unable to define a working mtable structure in ctags

I have a simple XML-like filetype in which I have two kinds of tags. First, we find the Block tags, then inside them the in-lines.

A simplified example text is like the following

messing.dan

<B=003>/introduction/changelog.html
& Changelog &

------------------


Built with Sphinx using a theme provided by Read the Docs .

</B>
=====================
<B=004>/introduction/executingScripts.html
& Executing Scripts &
<I=1>

-----------------------

<I=2><I=3>

Installing scripts in the Scripts menu


</B>
=========================

As you can see, the in-lines are always nested inside . I ultimately want to parse both and , adding to the the scoping of which are they nested to

So following these examples I have

[Capturing variables in sequences] (https://docs.ctags.io/en/latest/optlib.html#capturing-variables-in-a-sequence) Example 3

dan.ctags

--langdef=dantags
--langmap=dantags:.dan
--kinddef-dantags=i,inline,inlines

--_tabledef-dantags=toplevel
--_tabledef-dantags=blocktab
--_tabledef-dantags=inlinetab

--_mtable-regex-dantags=toplevel/^<B=([a-zA-Z0-9]+)>.*\n/\1/b/{tenter=blocktab}
--_mtable-regex-dantags=toplevel/.//

--_mtable-regex-dantags=blocktab/<\/B>//{tleave}
--_mtable-regex-dantags=blocktab/<I=([a-zA-Z0-9]+)>/\1/i/
--_mtable-regex-dantags=blocktab/.//

Yet all I get in my file is just the first been parsed, that's it.

Upvotes: 0

Views: 14

Answers (1)

Masatake YAMATO
Masatake YAMATO

Reputation: 1300

scope= mtable-regex-flag may help you. Ctags uses an internal stack to track scope.

Specify {scope=push} when entering a scope. Specify {scope=pop} when leaving a scope. Specify {scope=ref} for assigning the scope at the top of the stack to a newly created tag. See also scope= in https://docs.ctags.io/en/latest/man/ctags-optlib.7.html

$ cat input.dan 
<B=003>/introduction/changelog.html
& Changelog &

------------------


Built with Sphinx using a theme provided by Read the Docs .

</B>
=====================
<B=004>/introduction/executingScripts.html
& Executing Scripts &
<I=1>

-----------------------

<I=2><I=3>

Installing scripts in the Scripts menu


</B>
=========================


$ cat dan.ctags 
--langdef=dantags
--map-dantags=.dan
--kinddef-dantags=b,block,blocks
--kinddef-dantags=i,inline,inlines

--_tabledef-dantags=toplevel
--_tabledef-dantags=blocktab

--_mtable-regex-dantags=toplevel/^<B=([a-zA-Z0-9]+)>[^\n]*\n/\1/b/{tenter=blocktab}{scope=push}
--_mtable-regex-dantags=toplevel/.//

--_mtable-regex-dantags=blocktab/<\/B>//{tleave}{scope=pop}
--_mtable-regex-dantags=blocktab/<I=([a-zA-Z0-9]+)>/\1/i/{scope=ref}
--_mtable-regex-dantags=blocktab/.//


$ ctags --options=NONE -o - --options=/tmp/dan.ctags /tmp/input.dan 
ctags: Notice: No options will be read from files or environment
003 /tmp/input.dan  /^<B=003>\/introduction\/changelog.html$/;" b
004 /tmp/input.dan  /^<B=004>\/introduction\/executingScripts.html$/;"  b
1   /tmp/input.dan  /^<I=1>$/;" i   block:004
2   /tmp/input.dan  /^<I=2><I=3>$/;"    i   block:004
3   /tmp/input.dan  /^<I=2><I=3>$/;"    i   block:004

What I changed from the original version.

  • (must) using [^\n] instead of . in the first regex in toplevel table not for consuming whole the file contents
  • (must) adding {scope=...} flags
  • (optional) defining block kind explicitly
  • (optional) deleting inlinetab because it is not used
  • (optional) using --map-<LANG> option instead of --langmap option

--verbose option is helpful to see how ctags consumes the input.

Upvotes: 0

Related Questions