Follow-up Comment #9, bug #64450 (group groff): [comment #7 comment #7:] > In the year-plus since I last commented, I've seen more > _nroff_-using _make_(1) files in the wild
I have not done these Makefile safaris, so the following opinions are untainted by fact. As far as I can tell, groff's nroff has never used -e or -s for any purpose; it came out of the womb ignoring them for (presumably) compatibility. (The first nroff.sh, added in 1992 in [http://git.savannah.gnu.org/cgit/groff.git/commit/?id=b04d345ed commit b04d345ed], was a lean 51 lines that included logic to ignore -e, -q, and -s.) So the only scripts we need to worry about are those written against some other nroff and unmodified when groff's nroff replaced it--hopefully a small number, but here's where my ignorance starts to show. To that, I'd observe that it is generally harmless to run unneeded preprocessors. However, there may be some concern about running a preprocessor again on input that has already had that preprocessor run via a pipe. I believe you've tried to make groff's preprocessors more forgiving of this, but it may not be perfect (yet, or achievable in all cases). An "enabler" option works actively against the goal of being able to use the same command line for troff and nroff. But that may be a cost worth incurring, if the cost of an unwanted eqn or soelim is too high. > That means selecting an option letter that _groff_ already uses > that no _nroff_ user would ever employ. I sympathize with the reasoning that got you here. On the other hand, having the same option letter mean different things in different contexts makes the system as a whole harder to learn. Marginally, in this instance, but marginal difficulties accrue. (And I realize this would hardly be the first such overloading.) _______________________________________________________ Reply to this item at: <https://savannah.gnu.org/bugs/?64450> _______________________________________________ Message sent via Savannah https://savannah.gnu.org/
