On 8/21/2018 10:00 AM, Procházka Lukáš Ing. wrote:

When building a document (or more precisely: documentation which consists of many documents), I'm widely referring to other files via \input, \component etc., I'm widely using \defs, \defines and \buffers defined in separate files (as these defs are used at more places - so let them being defined in one source file) and I'm often generating texts by Lua (e.g. tables - being read and parsed from a text in a \start/\stopluacode block or e.g. Excel named range read via LuaXls or from a .xml - and typeset by a Lua function, which allows simply change all \NC into \VL in a particular place).

The lengthy process is to EDIT (almost) finished docs/documentation. And at this place, it would be very handy if synctex worked. So I vote for a better click-in-PDF-go-to-source support.

in principle you can push / pop input states but for such a variety of input you have to do that yourself in the styles ... and it will quickly become a performance issue then

the core will not get that kind of hard coded synctex flip-flopping

fwiw, the xml sub mechanism does support that kind of functionality as option (only because a collegue uses xml with deeply nested inclusions spread all over a direcory structure)

I don't know anything deep about how sync is provided. But my layman point of view would be that whenever a character (or a "box") is to be placed on the page (or into output stream), ConTeXt should know which is the "deepest - currently read" source file, the line, moreover the column - and that position could result into "pick-and-go" position.

only partly ... you don't know where a macro comes from (or what lua code generated something) and it makes no sense either to go bakc to some macro ... in fact, synctex support in context blocks going to styles because those who edit files are not supposed to change styles

things like headers and footers come from styles, not user input

and even much structure stuff comes from elsewhere (like titles of sections: they go via lua so there we already need to cheat input registration)

In this approach - "boxes" generated by Lua-in-ConTeXt should "jump" on \startluacode, \ctxlua, \cldcommand statement...

won't happen ... way too much overhead and it would polute the source too (i follow the principle that any new mechanism that gets added will not slow down compilation in a measurable way)

Anyway, i can cheat at the tex/lua end if needed and support e.g. titles
but i'm not going to pollute the code with every place where we come
from lua (also because in most cases there is no relation with the
source anyway then). Also, I will not add code that can have an impact
on performance when synctex is turned off.

Synctex should be turned on/off before source files are compiled into PDF - via command line option "--synctex" or \enabledirectives[system.synctex] before \starttext (\enabledirectives[system.synctex] after \starttext (\startcomponent) should be ignored?) - that could cause hooking picked Lua functions (or taking their "synctex-on" alternatives) - so normally (with synctex off) there wouldn't be any performance impact; worse performance would come only with synctex on - but it's user's choice.

But for a subset of
constructs that are relatively short in usage it is doable (but doesn't
really qualify as fun -- also, i don't use it myself).

Anyway, a better support for "pick-in-PDF-and-go-to-source" (synctex), namely when user "tunes" the docs to their final stage, would be great.
I think that the current approach to hard code synctex libraries into a viewer is a limitation (instead an editor should call an external program with the page and coordinates so that an external program can then look at the synctex file and decide where to go in the editor. That way more sophisticated support is possible than the now hard coded heuristics. Keep in mind that these heuristics are tuned for latex and from context we generate the real minimal amount of synctex code that works without clashing with these heuristics. If that were not that case is would be unuseable.

Anyhow, there's only so much i can do about it (i'm not going to patch viewers). Currently synctex is mostly working for decent structured source (can be multiple files).


                                          Hans Hagen | PRAGMA ADE
              Ridderstraat 27 | 8061 GH Hasselt | The Netherlands
       tel: 038 477 53 69 | www.pragma-ade.nl | www.pragma-pod.nl
If your question is of interest to others as well, please add an entry to the 

maillist : ntg-context@ntg.nl / http://www.ntg.nl/mailman/listinfo/ntg-context
webpage  : http://www.pragma-ade.nl / http://context.aanhet.net
archive  : https://bitbucket.org/phg/context-mirror/commits/
wiki     : http://contextgarden.net

Reply via email to