1

( https://trac.godzil.net/gcc4ti/ticket/6 )

This ticket is aimed at toolchain maintainers much more than at users, since this is the plumbery (used by a handful of persons only) of the documentation system.

It's no news that the programs handling the TIGCC documentation system (specific to TIGCC) has several drawbacks, and that Kevin has been contemplating rewriting the tools for a while. Actually, he already rewrote two of the simplest ones.
Obviously, it's desirable for both projects that TIGCC and GCC4TI cooperate in the making of the documentation system.

Before spending some time to rewrite the documentation tools (at some point in time), we feel that other approaches (e.g. DITA and DocBook, which depend on external tools, so that parts of the wheel don't have to be reinvented) should be evaluated at least. These approaches may prove completely unsuitable for the task, but maybe a small amount of time spent looking around at what people have done will save time in the end.
Typedefs part of multiple header files, and some checking code in the existing tools, may be hard to translate without a "real" programming language...

In case of a near-1:1 rewrite (besides fixing the several drawbacks, while we're at it, we should take advantage of the rewrite to improve the documentation, e.g. with ticket #5 if deemed worthwhile), high-level scripting languages (Perl, Python, Ruby, etc.) are just as portable, but tend to require less coding time than C / C++ using Glib, Qt4 or whatever other portable libraries.
Most toolchain maintainers use Linux, but on Windows, all of them require an extra installation step anyway.
avatar
Membre de la TI-Chess Team.
Co-mainteneur de GCC4TI (documentation en ligne de GCC4TI), TIEmu et TILP.
Co-admin de TI-Planet.