markFeatureWriter improvement for Indic

[pauldhunt]

When running markFeatureWriter with the options -m -t -c -i, on my Devanagari project it would be preferable for the Indic mark-to-mark lookup to be written at end of the abvm file, instead of to a separate mkmk file.

I believe that the -i and -m switches should do different things. Rules for Indic mark-to-mark should be written by default when the -i option is invoked to their respective abvm or blwm features. The -m switch should only apply to non-Indic mark-to-mark rules when they are both activated so that Indic and non-Indic mark-to-mark rules and lookups get written to the files that correspond with the correct mark features.

Running the markFeatureWriter in this case also produces empty mark.fea files. This should not happen.

Lastly, a sample of the output to the mkmk file when run as described is as follows:

lookup MKMK_MARK_abvmBNDU {
 	lookupflag MarkAttachmentType @MC_abvmBNDU;

	pos mark dvEsign <anchor -53 683> mark @MC_abvmBNDU;
	pos mark dvAisign <anchor 2 683> mark @MC_abvmBNDU;
	pos mark dvEcandrasign <anchor -149 734> mark @MC_abvmBNDU;
	pos mark dvEshortsign <anchor -9 683> mark @MC_abvmBNDU;
	pos mark dvAysign <anchor 1 683> mark @MC_abvmBNDU;
} MKMK_MARK_abvmBNDU;

the line lookupflag MarkAttachmentType @MC_abvmBNDU; actually prevents the following rules from forming.