I'm comparing an algorithm to other canned algorithms from the DVSDK (4_03_00_06), such as aachedec. When I build at the top-level, 'make codecs_clean', it does nice things, whereas my codec doesn't get cleaned. My algorithm is ultimately getting included in the clean, which boils down to a '$(XDC) clean' command with XDCARGS="prod". Unfortunately, with the default configuration of the algorithm, this command does nothing. (see below)
======== clean [packages/fluke/imagecor] ========
======== clean [packages/ti/sdo/codecs/aachedec] ========
rm -f package/ti.sdo.codecs.aachedec.sch
rm -f package/.vers_g160
rm -f package/.vers_r170
rm -f package/.vers_b150
rm -f package/.vers_c150
rm -f package/.xdc-u17
rm -f package/ti_sdo_codecs_aachedec.java
rm -f package/ti_sdo_codecs_aachedec.class
rm -f package/package_ti.sdo.codecs.aachedec.c
rm -f package/package.defs.h
rm -f package/package.doc.xml
rm -f package/ti.sdo.codecs.aachedec.ccs
rm -f ti_sdo_codecs_aachedec.tar
rm -f package/rel/ti_sdo_codecs_aachedec.xdc.inc
rm -f package/rel/ti_sdo_codecs_aachedec.tar.dep
rm -f .libraries .libraries,*
rm -f .dlls .dlls,*
rm -rf package
All of the pertinent 'clean' target information gets stored in a package.mak folder under the codec's directory (e.g. aachedec/package.mak). That file seems to be temporary, though. There is no equivalent for my wizard-generated algorithm.
How can I easily mimic what the canned codecs have in terms of the ability to be cleaned from the top-level 'make codecs_clean' command. There's some magic here that I'm not understanding.
I can continue to develop easily on my machine by hand-removing the output files when I want to rebuild the codec server. That works, it's just a bit clunky and isn't a good long-term solution if this is going to be worked on by more than just me, which is likely.