Hi, I saw in another post that loading of binary modules has been enabled in luatex (great!) that requires a stock lua51.dll on Windows. I wonder if this side-effect that will last or pass? I'm in the process of rewriting the generic script wrapper for TL on Windows to replace the current batch script(s). I didn't want to code everything in C, so I use only a thin binary stub that links with luatex.dll, sets up the argument list and calls dllluatexmain with 'texlua' and a predefined script as argv[0] and argv[1]. This is nice as I can implement most of the logic on the lua side but there is currently no elegant way to return to the C side (luatex exits rather than return) nor I can easily pass some data back. This is not a big problem for me ATM (after finding a workaround for os.spawn I don't need to return to the C side) but it could be useful in general to have more open access to (tex)lua interpreter from C. Is anything along these lines on the agenda? Cheers, Tomek
T T wrote:
Hi,
I saw in another post that loading of binary modules has been enabled in luatex (great!) that requires a stock lua51.dll on Windows. I wonder if this side-effect that will last or pass?
This is not easy to answer. Right now, it is clearly a side-effect. If we decide to keep the binary modules support (which is experimental right now) then in the future, it would be nice to build our 'own' lua dll. But that will require quite some changes to the build scripts as well as a conceptual change to the distribution process. Not easy to integrate that back into texlive. Best wishes, Taco
2009/11/24 Taco Hoekwater
If we decide to keep the binary modules support (which is experimental right now) then in the future, it would be nice to build our 'own' lua
That would be very welcome, indeed. It could make lua(tex) not only a common scripting solution, but also a common platform for building kpathsea/TeX aware applications. There are already a few interesting lua based applications, e.g. ps_view, ipe v7 (LaTeX friendly vector graphic editor). I expect projects like these could be quite interested in more open access to luatex and some very innovative solutions could emerge in the process.
dll. But that will require quite some changes to the build scripts as well as a conceptual change to the distribution process. Not easy to integrate that back into texlive.
I admit that I'm blissfully unaware of these hurdles, so I will welcome any advancements that can make luatex more accessible to us mere mortals ;) I see it as a major step in TeX development. Thanks for all the hard work you put into it. Cheers, Tomek
T T wrote:
This is not a big problem for me ATM (after finding a workaround for os.spawn I don't need to return to the C side) but it could be useful in general to have more open access to (tex)lua interpreter from C. Is anything along these lines on the agenda?
we're now about version 0.50 so you can expect more opening up but we focus on the tex machinery. also, we're currently aiming at a stable version 0.50, and most of next year will be used for cleanup and further stabelizing (directionality, backend, etc), cwebbing, as well as enhancing mplib which is associated with this project Hans ----------------------------------------------------------------- Hans Hagen | PRAGMA ADE Ridderstraat 27 | 8061 GH Hasselt | The Netherlands tel: 038 477 53 69 | fax: 038 477 53 74 | www.pragma-ade.com | www.pragma-pod.nl -----------------------------------------------------------------
participants (3)
-
Hans Hagen -
T T -
Taco Hoekwater