Re: The multisubstitute commands in the run script generated by s6-usertree-maker are in a wrong order
Great, thanks! MMS On Fri, Jan 22, 2021, 17:28 Laurent Bercot wrote: > >As shown above, the multisubstitute command that contains XDG_RUNTIME_DIR > is put after the one that contains USER, HOME, UID, GID, and GIDLIST. If > for example XDG_RUNTIME_DIR=/run/user/$UID, the $UID here will not be > substituted with the user's UID since by the time $UID is substituted, > $XDG_RUNTIME_DIR hasn't been substituted yet. So perhaps the order of these > two multisubstitute should be inverted. > > You're right, of course. I remember testing it, and it *working*, so > I did not think any further, but in retrospect it appears my test was > incorrect. > Thanks for the report! Fixed in git head. > > I'll give it the week-end, in case more bug-reports come in, then > I'll release 2.10.0.1. > > Note that skalibs-2.10.0.1 is out already, it fixes a bug that > manifests in execline's 'emptyenv -c' command. > > -- > Laurent > >
Re: The multisubstitute commands in the run script generated by s6-usertree-maker are in a wrong order
As shown above, the multisubstitute command that contains XDG_RUNTIME_DIR is put after the one that contains USER, HOME, UID, GID, and GIDLIST. If for example XDG_RUNTIME_DIR=/run/user/$UID, the $UID here will not be substituted with the user's UID since by the time $UID is substituted, $XDG_RUNTIME_DIR hasn't been substituted yet. So perhaps the order of these two multisubstitute should be inverted. You're right, of course. I remember testing it, and it *working*, so I did not think any further, but in retrospect it appears my test was incorrect. Thanks for the report! Fixed in git head. I'll give it the week-end, in case more bug-reports come in, then I'll release 2.10.0.1. Note that skalibs-2.10.0.1 is out already, it fixes a bug that manifests in execline's 'emptyenv -c' command. -- Laurent
