Commit Graph

9 Commits

Author SHA1 Message Date
Jeremy Rand
b89ad278dc Improve the build to better support building the babel fish translator. Fix the bugs in the filter code for the translator. Based on testing, I think it is working now and matches any file with a name *.md. 2021-05-13 23:28:53 -04:00
Jeremy Rand
9e8e470ed9 Using the latest version of GoldenGate, the teachRez script workaround is no longer required. The md2teach exec works as well now under GoldenGate as it does under GNO/ORCA (which hopefully means it works well, TBD). 2021-05-11 23:35:26 -04:00
Jeremy Rand
e54c22fdda Add more to the makefiles to support building both the shell command and the babelfish translator which is the next step in supporting markdown. 2021-05-06 23:43:31 -04:00
Jeremy Rand
76d20f76e6 Turn on the -r argument in the default execution rule in the makefile. Some minor cleanup. 2021-05-05 23:33:59 -04:00
Jeremy Rand
34fd580448 Implement the rest of the direct writing of resources to the file. I have tested this under GNO (as opposed to Golden Gate where resource writes are not yet supported) and it does seem to work. I also tested it at 2.8MHz and the large test file takes about 30 seconds to convert. I may look at turning on the optimizer in the compiler to try to get a bit better performance on real HW. I decreased the stack size to 8K for now because 32K was too big for a "real" machine as opposed to the emulated Golden Gate. 2021-05-05 00:23:28 -04:00
Jeremy Rand
9e3562dd5f Implement a workaround for writing resources. Write the resources I want to a .rez file and use a script in the build itself to turn that into resources and attach it to the output file. This seems to result in a styled teach file for my test input that loads and looks pretty close to right. 2021-04-29 00:24:27 -04:00
Jeremy Rand
8b97df852c Use the --mem argument in Golden Gate to get an accurate stack usage value. Beginnings of work on writing resource information. 2021-04-28 23:20:20 -04:00
Jeremy Rand
effa777037 Improve the pure text output of the converter. It now handles numbered and unnumbered lists. It adds a horizontal line for HR. Spacing is better between blocks. It no longer inserts a leading blank line at the top of the document.
I have extended the execution rule to put the resulting text file onto a disk image and launch an emulator so I can see the text file on the target machine.  This should make iteration easier.

Even though I am asking to print carriage returns in the output, I am getting line feeds.  Now sure what is doing the translation.  It could be the ORCA standard library (not likely) or Golden Gate (maybe).  I should next change to direct use of GS/OS for the writes and avoid the ORCA library anyway.
2021-04-21 00:00:01 -04:00
Jeremy Rand
ad2a57f4d4 Initial Commit 2021-04-13 15:10:02 -04:00