Follow-up Comment #3, bug #44714 (group groff): [comment #1 comment #1:] > I don't understand why '.do tm' is going to the next input line > to collect arguments. 'tm' is not documented in CSTR #54 as > behaving this way.
It's not the behavior of .tm that's at issue, but of .do. This is the bug that's referred to in this comment in [http://git.savannah.gnu.org/cgit/groff.git/tree/tmac/andoc.tmac tmac/andoc.tmac]: "Due to a bug in GNU troff it necessary to have a no-op line between `.do' and `\*'." This comment was added in [http://git.savannah.gnu.org/cgit/groff.git/commit/?id=42c866a3f commit 42c866a3f] (which was part of the fix for bug #44708) and refers to these changes in the commit: -. als TH reload-man -\\*[Dd]\\ +. do als TH reload-man +. +\\*(Dd\\ -. als Dd reload-doc -\\*[TH]\\ +. do als Dd reload-doc +. +\\*(TH\\ Indeed, removing these two empty control lines wreaks havoc, showing that a .tm isn't necessary to trigger the bug. Probably until this is fixed (which may be a while if it proves as intractable as Werner predicts), the andoc.tmac comment should cite this ticket rather than vaguely referring to "a bug in GNU troff." [comment #2 comment #2:] > Why does '.tm' need a '.do'? It doesn't; that's just a minimal test case to illustrate the bug. The situation in andoc.tmac is a real-world case but not a minimal one. _______________________________________________________ Reply to this item at: <https://savannah.gnu.org/bugs/?44714> _______________________________________________ Message sent via Savannah https://savannah.gnu.org/