On Sat, 25 Apr 2009, Joris van der Hoeven wrote:
This /bin/sh issue comes up every now and then; a pain.
In principle, as long as we no not use any bash-specific stuff,
then we probably should stick to /bin/sh.
Andrey, do you remember the previous conclusions on this subject?
Maybe, I don't remember all issues, this was discussed many times. One I
remember is 'type -p' vs. 'which'. Initially, I used 'which' in
maxima_detect. This was inconvenient for cygwin: it not always has 'which'
(I never used cygwin; my understanding (maybe wrong) is that it has some
package system, and 'which' is available in some package, but it is not
installed by default). 'which' was replaced by 'type -p', which is
bash-specific. More recently, 'type -p' was replaced (in many scripts in
/usr/libexec/TeXmacs/bin) by 'which', to make these scripts more
sh-compatible. Has the cygwin problem disappeared somehow?
In any case, currently (1.0.7.2) all scripts in /usr/libexec/TeXmacs/bin
use 'which', except tm_mathematica (it still uses 'type -p'). I think it
would be good to replace 'type -p' by 'which' in tm_mathematica, for
consistency. Though this script is hardly used by anybody: it worked
more-or-less well with mathematica-4.*, and the current version is 7, many
things don't work anymore :-(
I don't know what other bash-specific features are used in these scripts.
Probably, there are some, because there is some breakage on ubuntu. Ubuntu
users can investigate what scripts (and what lines in them) don't work in
the expected way.
Andrey
_______________________________________________
Texmacs-dev mailing list
[email protected]
http://lists.gnu.org/mailman/listinfo/texmacs-dev
