------------------------------------------------------------------------------ Title: TODO LIST FOR ADE Version: $HeadURL$ $LastChangedRevision$ ------------------------------------------------------------------------------ PROGRAMS/ADEGMT: make adegmt understand what I mean by this: dione$ cd /bldcfg dione$ adegmt -n .. . adegmt: ERROR: nothing was installed! ------------------------------------------------------------------------------ PROGRAMS/ADEGMT: saw this: dione$ pwd /home/alexis/doc/myth.net/ahdg dione$ adegmt --no-cvs .. doc/TODO doc/COPYING /home/alexis/bin/OSI/adegmt[1764]: cannot create /home/alexis/doc/myth.net/./doc/COPYING: No such file or directory /home/alexis/bin/OSI/adegmt[1764]: cannot create /home/alexis/doc/myth.net/./doc/TODO: No such file or directory Ok, I said '..' when it should have been '.' but the message is ugly. ------------------------------------------------------------------------------ PROGRAMS/ADEREL: Need to make a new aderel program. ------------------------------------------------------------------------------ PROGRAMS/ADESPP: adespp does not respects umask *above* source file permissions (i.e. if source file is [46]00 and umask is 022 and -nx is used, then result will be 644 when it should be 600. ------------------------------------------------------------------------------ PROGRAMS/ADEROOT: doesn't delegate early enough: 'aderoot -V' shows the wrong program version's version. ------------------------------------------------------------------------------ PROGRAMS/ADEMPX: write adempx (ADE man page extractor) to take man pages from sources. Then I can start documenting library functions. ------------------------------------------------------------------------------ TEMPLATES: The Makefiles need to be much more aligned with each other. The 'single-shell' and 'single-perl' should probably be dropped - adding things to the BIN_FILES list (e.g. backup) or subtracting things (e.g. fad) is done in inconsistent ways. Can prologue, settings and epilogue be merged? ------------------------------------------------------------------------------ TEMPLATES: ADE single shell and single perl script man pages should be .spp files, even if there is no need for them to be. Then aubdom - for which there is such a need - could have a standard make file (it doesn't at the moment) template aubdom uses a non-aubdom-specific Makefile to compile and install ------------------------------------------------------------------------------ DOC: write man pages for ADE shell routines - in the sources. ------------------------------------------------------------------------------ DOC: Need 'TUTORIAL' file. ------------------------------------------------------------------------------ DOC: write man pages for: adegmt, aderoot, adetest ------------------------------------------------------------------------------ LIBS: (*) write ade_???usage, which like ade_lck_lock will call a module- specified function to display the text but will then return back to itself to call exitdel (so that the user doesn't) (*) same for ade_listpaths() (*) version reporting should also be a ADE function, calling something in the program just to get the version. (*) the perl version of gep - unlike the shell one - does not reset the PATH, why? I have code in checkntp.spp to reset the PATH because there it is actually *required*, but shouldn't I align the perl and shell versions of gep in this respect? ------------------------------------------------------------------------------ LIBS: move the 'this code autogend' function from adespp.spp to utils.pl.spp and utils.sh.spp (and utils.c?) ------------------------------------------------------------------------------ LIBS: ade_ffd needs to support field labels on row one, i.e.: LOGIN:PASSWD:UID:.... ::... possibly with primary key markers: LOGIN/P1:PASSWD:UID/P2:.... meaning that 'LOGIN' is a primary key, and uid is a second unique field upon which lookups can be done. Consider a netstat db: SRCIP/P1:SRCPORT/P1:... where it is the *combination* of the identically-levelled 'P' fields which is the primary key this should be good for lcuinv, which can then identify fields by name instead of by field index and it should be used in 'backup' which can be extended to include 'date of allocation' and 'number of times used' ------------------------------------------------------------------------------ LIBS: the call to 'trap' should be done inside gep.*.spp. Template programs (and all programs) should only exit via return or error. ------------------------------------------------------------------------------ ADE 4.0 ------- Exit points will be: ade_msg_error _ade_exit_usage { 1 | 0 } return 0 ade_listpaths