On Feb 7, 2007, at 7:59 AM, Mojca Miklavec wrote: > On 2/7/07, Paul Fons wrote: >> I am a little late it would appear to enter into the discussion on >> fink's internal tex conflicting with external texs, but I thought I >> would offer an idea to potentially solve the problem. First, I think >> that everyone would agree that the (obsolete and discontinued) >> version of teTeX included with fink is a poor choice for people who >> use tex for writing. > > For those who don't require much and for compiling the documentation, > teTeX might still be usable, but for anyone with some reasonable > demands, it's out of question. > >> On the other hand, teaching fink above every >> version of TeX that people may use would seem to be a waste of energy >> as well. > >> Proposal: if fink only needs a rudimentary TeX to typeset >> documentation, why not 1. require a fink version of tex that is >> always in the same place on all systems (e.g. binary compatibility) >> and 2. hide the installation of this tex by placing it off the user's >> path (e.g. place the tex binaries in a "standard" location within the >> fink /sw subtree that only fink knows about? > > That's possible, although I cannot really imagine the amount of work > needed to fix all the packages, so that they will become aware of that > change. > > Although - when I think about it again: if one would set the necessary > TeX variables and add the fink's "hidden" TeX executables to PATH only > during installation of new packages (my impression was that most > packages need it for compiling documentation, but I may be wrong), > that should work, but on the other hand you still have people who > would indeed be interested in using those binaries. So you need to > have two variants then (one hidden and one visible, depending on user > needs.) > >> I noted that it was >> also suggested to change the path order, but this is a questionable >> solution and would likely cause lots of problems down the road (a >> debugging nightmare). As the fink tex solution is really not useful >> outside of fink, why not try the solution above? > > I had serious problems if /sw path was at the beginning, I don't > experince problems now when I installed fink's teTeX (and moved /sw at > the end). > > In the long term it would surely make sense to integrate TeX Live. I > went researching a bit, but there is one big problem: it's size, and I > really have no idea how to cope with that. Unless we want to end up > with 2GB zip, one should make a fink-specific zip somewhere on CTAN > and adapt everything needed for it (research and remove files that are > not needed, split contents into multiple packages - the most important > would be to create a "minimal" tree, which would allow compiling the > documentation, and a few additions, which could serve as a serious > alternative to other distributions). > > There is still an enormous mount of work that would have to be spent > for that if one would want to do it properly (and I still doubt that I > could do it better than the current gwTeX, which, after all, is still > being somehow-maintaned and updated). One can of course take the whole > tree as well, but that wouldn't be the best idea. > > teTeX, gwTeX, fpTeX ... maintainers probably gave up with a good > reason (too much work involved) ... Creating a (fink) package withouth > the intention of maintaining it, is probably not worth the time > investment, otherwise it will end up like teTeX now - unusable for any > serious work. I was thinking about creating a package, but there's too > much work involved and I'm not capable of maintaining it. Esp. because > gwTeX still does it's best. > > > But there's still something that I don't understand: what exacly does > "support multiple external TeX distributions" mean and why is that a > problem? Most packages which need TeX to compile their documentation, > need only a working version "pdflatex". (I must have missed something, > but I don't know what.) > > Which package needs more than that? (Perhaps XeTeX, but that's another > story. One would not want to istall fink's XeTeX over "system-tex" > anyway.) > > To sum my thoughts up: > - I still think that system-tex makes sense until one creates a > texlive package (after all, the maintainer of gwTeX has "quit" after > he did a major update of the system, not before it, so the system is > still usable) > - texlive might be the long-term solution to go for, but it needs a > bit of tweaking > > Mojca
Hi, just a quick remark in case drm (the tetex maintainer) should read this and come to the conclusion that nobody appreciates fink's tetex: I've been using fink's tetex almost exclusively for 5 to 6 years and see nothing wrong with it for any "reasonable" use. I've probably customized it a little, but I rarely ever have to worry about my tex installation because it "just works". I'm not trying to say tetex is "the only tex you'll ever need" - but it's just not accurate to say fink's tetex is "unusable". So thank you very much for keeping the tetex package around, drm & co! Jens ------------------------------------------------------------------------- Using Tomcat but need to do more? Need to support web services, security? Get stuff done quickly with pre-integrated technology to make your job easier. Download IBM WebSphere Application Server v.1.0.1 based on Apache Geronimo http://sel.as-us.falkag.net/sel?cmd=lnk&kid=120709&bid=263057&dat=121642 _______________________________________________ Fink-users mailing list [email protected] https://lists.sourceforge.net/lists/listinfo/fink-users
