> While ctags has the option to use regex to parse the source files, it
> is not what is being used for the vhdl module of ctags. The regex is
> only brought into the equation when the tags file is written for
> vhdl.

Ok.  Thank you for digging that.  But note that the parser doesn't deal
with protected type (nor any vhdl 2008 features).

> A solution involving ctags would require at least a partial re-parse
> and rewrite of the ctags file every time the file buffer is changed.
> (not a pretty solution)
> 
> note: (I use token to mean any file data that needs to be
> kept-track-of.)
> 
> I think a good solution for fixing this problem would be: rewriting
> the parser to include token scope, etc. and changing the tags file
> format, so the last tab delimited field per line would be a call to
> an external program. (do-able but, a bit bulky, and cumbersome)

I suppose that any solution based on spawning a process when the file
is modified would be too slow.

> The program would output the location of the tag regardless of
> white-space reformatting or comments.
> 
> This would break compatibility I'm almost sure of it.
> 
> 
> If we were to use GHDL to output all the token locations for each
> file and noting scope, the external program for locating the current
> location of the token regardless of white-space or comments would
> still be an essential tool.

GHDL has that capability (although this is not fully implemented).

Regards,
Tristan.

_______________________________________________
Ghdl-discuss mailing list
[email protected]
https://mail.gna.org/listinfo/ghdl-discuss

Reply via email to