RE: bug in my view: temp directory hard-coded in cvs binary comes from the user that built the exe

[Ruben Diez]

> > I just found out to my cost that the temp directory hard-coded in the
cvs
> > binary comes from the user that built the exe. That's bad practice, and
I
> > think it should be considered as a bug. (I built under Linux
> > using the default makefile).
> >
> > I have implemented the -T work-around in xinetd.d/ etc. etc., but it's
> > still a bug in my view!
> >
>
> How else would you expect CVS to determine the temp directory to use at
> run time, aside from the -T option or the $TMPDIR environment variable?

I quote from the CVS manual, section "2.11 Temporary directories for the
server":

-------8<-------8<-------8<-------

They are located in the directory specified by the `-T' global option (see
section Global options), the TMPDIR environment variable (see section All
environment variables which affect CVS), or, failing that, `/tmp'.

-------8<-------8<-------8<-------

There's no mention there of any hard-coded directories in the CVS executable
that were grabbed from the environment of the user that built the executable
(/home/rdiez/temp in my case), which is in itself a bug in the manual
(although it's mentioned somewhere else).

I think CVS should not have any hard-coded temp directories inside. It
should just do what the manual says in that section.

The idea that I build a CVS executable and give it to a friend, or place it
in a server, and that it's got hard-coded inside some /home/rdiez/temp,
because my user account was rdiez when I built it, just feels wrong to me.

Cheers,
  Ruben