Re: Potential bug with "--path-style through" and GTAGSLIBPATH
Aymeric Agon-Rambosson <[email protected]> Wed, 07 May 2025 15:14:33 +0200
| Newsgroups | gmane.comp.gnu.global.bugs |
|---|---|
| Message-ID | <87selgbm12.fsf@X570GP> |
Hello, Sorry for the late answer. Le vendredi 2 mai 2025 =C3=A0 22:58, Shigio YAMAGUCHI <[email protected]>=20 a =C3=A9crit : > The original purpose of this option (--path-style=3Dthrough) was > to allow external programs to do the path manipulation that > global(1) does. I think it would be easier for them to use it > without referencing GTAGSLIBPATH. The point of the question was in fact to ascertain what the=20 original purpose of the option was, and how it is to be used. I was asking because an emacs package was using it, which resulted=20 in unwanted behaviour (see=20 https://github.com/Ergus/gtags-mode/pull/8). You've been clear that this option is not to be used with=20 GTAGSLIBPATH, which is enough. > How about this? > > else if (!strcmp(path_style, "through")) { > type =3D PATH_THROUGH; > unsetenv("GTAGSLIBPATH"); <=3D=3D add > } > > global(1) > ... > =C2=B4through=C2=B4 means the relative path from the project > root directory (internal format of GPATH). > When this style is used, GTAGSLIBPATH is ignored. <=3D=3D=20 > add This is the simplest solution. There is no possibility now that=20 users could possibly misinterpret the output path now. Best, Aymeric