a requirement for fbgs, a wrapper script for viewing ps/pdf files on the
framebuffer console using fbi, an image viewer for the linux framebuffer
console.
* add a depend line.
* check that python is installed.
* call python directly, rather than using $(PYTHON) (I could find
no use of the $PYTHON variable in any other packages.
This package contains the cmavo, gismu, lujvo, and rafsi, wordlists for
Lojban.
This is reference material when installed on the NanoNote, and is
also a build-time dependency for jbofihe.
I've updated my NanoNote to the latest source, and having fixed the
patch for this package, wordgrinder is now working.
This program was crashing on the ~2010-06-17 image, delivered with
my NanoNote.
Only compile&install a minimum host Gforth for bootstrapping the target
Gforth, removing strange emacs-dependencies that cropped up earlier.
Image file can now be built on the nanonote using 'gforth-update-image'
command. Also, Gforth is now run through a wrapper that calls
'gforth-update-image' automatically, if necessary.
Also first attempts at supporting libcc C-interface libraries pre-compiled on
the host.
Tested to compile on amd64 and i386.
I believe upstream modified the wordgrinder-0.3.3 release in place.
Only whitespace is different between this patch and the previous
patch.
This now applies cleanly.
ElementTree is not available by default. (it is in pyexpat, which is
a non-default module.)
since ElementTree is only used by the Updater code, and this code is
not designed to run on the NanoNote (it is only run as part of the
build process), move the import statement inside the Updater in the
hopes that the user on the NanoNote never has to deal with it.
In the future I will make pyexpat a default module for the Lojban
config, but this change will still see a performance gain in the
most typical use case for makfa.
ElementTree is not available by default. (it is in pyexpat, which is
a non-default module.)
since ElementTree is only used by the Updater code, and this code is
not designed to run on the NanoNote (it is only run as part of the
build process), move the import statement inside the Updater in the
hopes that the user on the NanoNote never has to deal with it.
In the future I will make pyexpat a default module for the Lojban
config, but this change will still see a performance gain in the
most typical use case for makfa.
I originally choose /var/lib/... because makfa.dump can be updated by
the makfa program. /usr/share/... is for read-only data, which this
most certainly is not.
The best solution for this problem would be to have some kind of
/ovelay/var/... directory that is copied at boot to /var/... This
would allow files that need to be persistent in /var/... to be so,
but would permit the flexibility currently seen for I/O heavy uses
of /var/...
Ater a discussion with xiangfu, we decided that modifying the
base-package to support an overlay on /var is too much work to do
right now, and that we would use /usr/share/... for the time being.
Update is not intended to be used on Ben NanoNote, so I'm moving the
import of pyexpat to the Updater class. pyexpat is not included in
the default build, and this patch saves me from having to install
it.
I will in the future include pyexpat and enable full support for
generating new dump files, but this patch will still function in
that case.
