On 11/04/2012 01:50, Lex Trotman wrote:
I notice that docbook isn't highlighted, should we make it use the Scintilla XML lexer? (since all the docbook I have is machine generated it has the extension xml any way so I had to set it to docbook)
It was highlighted before, looks like it got broken in the highlightingmappings.h change. One problem is the filetype file uses [styling=XML], which is recursive so should be =HTML. Another is that the style strings in the header don't match. I've fixed these locally but I'm not yet sure if docbook styling needs to be separate.
- The symbol tab now shows the sections and chapters in a separate
list, and that way the hierarchy of the tags (the structure of the file) is lost. It is possible to display the list in a tree?
The current tagmanager code is not very good at hierarchy handling (although it does it, it isn't good) and there is work under way to replace it. Until then I would say the answer is theoretically yes
Is there? By who? Surely it would still need to parse tags with the CTags parsers even if it's managed differently than TagManager.
Columban is doing some work on it as he gets time and yes his current intention is to re-use the current parsers (at least when I asked a couple of weeks ago).
Also not sure what you mean by tagmanager not being good at hierarchies.
Well, to be explicit I guess its more to do with the problems with repeated names (although Colombans recent patch has greatly improved it). At the moment the docbook parser uses the id as the unique key and only shows entities with an id, but not all entities must have an ID. To show the entire hierarchy it would have repeated<chapter> and especially<section> <paragraph> etc it is going to need a better "uniquification" of the repeated name. Colomban uses line number to uniquify the key, but with XML you often get multiple things on a line so something else is needed, eg character number.
Sounds like CTags/Tagmanager could just store a character number field rather than actually *replacing* tagmanager.
But I'm not sure it's really worth it - can you show a usual code sample where we actually want to show things in the symbol list that are on the same line number which have the same name?
Nick