Still investigating... It looks like the issue isn't caused by the actual gtags invocation. Something (perhaps gtags-mode) calls gtags with the --config option before the actual scan. This in turn invokes gtags_hook, too. I am already maintaining a list of options in my gtags_hook that should bypass generating the file list, e.g --help, --dump, --single-update. Now adding --config. Let's close this for now.
Thanks, Marcus On Tue, Apr 21, 2020 at 11:22 PM Shigio YAMAGUCHI <shi...@gnu.org> wrote: > Hello, > Could you please give me a reproduce method to understand the problem? > > Thanks, > Shigio > > 2020年4月21日(火) 22:26 Marcus Harnisch <marcus.harni...@verilab.com>: > > > > Hi Shigio > > > > It's been a while but related to this, I was wondering if > GTAGS_COMMANDLINE isn't supposed to include ‘--single-update’ when > specified. In my hook script, I am using this to determine whether to start > a long-running process of rebuilding the file list. > > It looks like the switch doesn't make it into GTAGS_COMMANDLINE. Did > this change in one of the recent releases? Could this be impacted by > something in the config file? I am almost certain that it used to work the > way I expected. > > $ global -u --single-update foo/bar > > GTAGS_COMMANDLINE=gtags --skip-unreadable --sqlite --compact > --config=ctagscom > > > > Thanks, > > Marcus > > > > -- > Shigio YAMAGUCHI <shi...@gnu.org> > PGP fingerprint: > 26F6 31B4 3D62 4A92 7E6F 1C33 969C 3BE3 89DD A6EB >