smbgroupedit patch

Bradley W. Langhorst brad at langhorst.com
Mon Aug 12 17:28:00 GMT 2002


On Fri, 2002-08-09 at 19:43, Andrew Bartlett wrote: 
> "Bradley W. Langhorst" wrote:
> > 
> > I propose the following changes to smbgroupedit in HEAD
> > to keep others from getting burned as I did by a
> > subtle change of group type during a unix association
> > 
> > I can't believe how long it took me to figure out why my
> > my groups weren't showing up.
> 
> Yes, the tools isn't the most intuitive - and that's *after* I allowed
> SIDs to be specified by name...
I wasn't trying to make a comment about the tool - more about my own
shortcoming...  I think it is pretty intuitive now, just that little
quirk about reverting to unknown group type was tough. 

> > -       if (sid_type != SID_NAME_UNKNOWN) {
> > -           if (sid_type != SID_NAME_WKN_GRP) {
> 
> Why not just do an explict test on type builtin then...  
i thought that was a test on builtin groups... 
i screwed the direction of that diff up though so maybe that was
confusing. 

> >         /* If a new Unix group is specified, check and change */
> >         if (group!=NULL) {
> > +           if (strncmp(group,"-1", strlen(group)) == 0){
> > +               printf("Dissociating unix groups from ntgroup: %s\n",
> > map.nt_name);
> > +               map.gid=-1; /*dissociate the unix group*/
> 
> Looks like a good idea - but we might to better to just have an option
> to 'disassociate'.
i thought about that but I wasn't sure how to fit it in 
maybe a -c "group name" -d option? 


> > +                       printf("The UNIX group: &s does not exist\n",
> > group);
> 
> Did you actually compile this?  That looks like an aperstand to me, not
> a % (percentage, format qualifier...).
ack that's a typo 
it does compile  but just prints out &s instead of "users", forgot to
test that bit of code 

brad 




More information about the samba-technical mailing list