On Tue, May 6, 2008 at 4:29 PM, George N. White III wrote:
On Sat, May 3, 2008 at 7:00 AM, Mojca Miklavec
wrote: some time ago some MikTeX users have complained that ConTeXt doesn't work on MikTeX. It seems to be a problem with MikTeX itself since this workaround seems to solve the problem, at least temporary: ruby "C:\Program Files\MiKTeX 2.7\scripts\context\ruby\texmfstart.rb" texexec
I'm not so quick to blame MiKTeX -- I suspect texmfstart.rb would need to provide some path searching tailored to MiKTeX. It seems to rely on $0 with a tetex/texlive directory structure where SELFAUTOPARENT can be used.
Christian has indeed replied that he uses the binary version of texmfstart from CTAN (which is a bit older that the one on PRAGMA). But I still don't understand what can go wrong "in the middle of nothing", without any major changes. So there is either a problem in MikTeX or in ConTeXt.
The only criticism of MiKTeX is that it is silly to distribute a broken package, but then it is up to ConTeXt users to report breakage and, ideally, supply fixes.
I completely agree. Except that I have no MikTeX any more. And I don't have the slightest idea what could have caused the problem. I don't see any serious changes in the texmfstart code that could break the 3-week-old version, while it has worked OK before. Mea culpa. I should have forced Hans to fix it during BachoTeX ... :) I will try to misuse some windows computer next week to see if I can find any bugs (but it's difficult to test since one cannot simply modify the ruby script as on other platforms).
(But apparently there are not many MikTeX users on the list anyway. Else more would have complained.)
Don't complain, just switch to a platform/distro where ConTeXt works and use mpm (since few others provide updates with the speed and convenience of MiKTeX).
I think MiKTeX should remain focusing on the packages and binaries, and rely on upstream to fix broken packages.
I agree. But I have thought that some change in MikTeX might have occurred in the meantime. Mojca