had to change lexconv perl script to accomodate flex naming...

moved stuff from Makefiles to Imakefiles

added -ldbm to makefile command lines

phopely 3/11/98



steam position was lost during dictman operation;
I saved and restored it over the critical sections, I think...

This may be a problem w/ the flex-generated scanner; it is unclear now.

May need to try different command line options during flex generation...



Changed the Makefile script for datum/ to include creation of a dummy
1st entry for the dictman compilation process...  For some reason, the first
record appears to be dropped.


On the nice side, it appears that the dictman system can now dump dictionary
contents.  Next, to check the tailer...

O.k. - the first real data element and tailer appear to be
in the 'dump database' dictman output.

So, it appears that you can now create and dump a dbm file using
the dictman tool under linux.


phopely 3/13/98


looks like moa is working under linux?!


phopely 3/13/98.2


*** Altered src/token/lexconv to, in theory, seemlessly work w/ 'standard' lex
(elsif clause) or w/ flex (initial conditional clause)...  Other lex-es
should be accomodated here by adding elsif clauses that match
against the yylex function declaration pattern of your lex's generated
lex.yy.c file...  Just be careful not to overmatch the pattern :)


phopely 3/17/98


Appears to run from CD live under linux now :)


phopely 3/18/98


Altered dictman.c to retry on EMANYARG or EUNKNOWNCOM because these errors
occur during dictionary construction process when fp pointer gets clobbered
by lex call - spooky...  Resets the hold_pos to previous value (stored
in hold_pos2).

Dictionary construction still has MAdictNInsertPOS entry redundancy errors -
this is a little scary also...

These symptoms may be a reflection of a greater memory management problem
or perhaps a bug in my linux kernel?  This is unclear...


phopely 3/19/98


Developed environment variable wrapper script system (Kim, Kim.2, Kim.proto,
Kim.2.env.proto, alter.c), installation code, etc., to permit the Kim system
to be used easily.

Wrote addendum READMEs, etc., like what you are reading right now :)


phopely 3/22/98