[Top][All Lists]

[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

Re: [gpsd-dev] scons quirks

From: Eric S. Raymond
Subject: Re: [gpsd-dev] scons quirks
Date: Wed, 13 Nov 2013 18:10:43 -0500
User-agent: Mutt/1.5.21 (2010-09-15)

Hal Murray <address@hidden>:
> scons install runs as root, so if it has to make a file, it ends up owned by 
> root ...
> I expect "scons" to make everything that "scons install" will need.  Is that 
> correct?

Not necessarily.  See below.
> On Fedora, I end up with these:
> -rw-r--r-- 1 root   root      438 Nov 13 12:00 gps-3.10~dev.egg-info
> -r--r--r-- 1 root   root      204 Nov 13 12:00 libgps.pc
> -r--r--r-- 1 root   root      219 Nov 13 12:00 libgpsd.pc

Looking at this, it appears that these files are only created at installation
time  That's a flaw.  I've added this to the TODO:

    ** Build issues

    libgps.pc, libgpsd.pc and the egg-info file are only created at
    installation time and never cleaned.  Because the user will be
    running as root, these are created owned by root and annoying
    to remove later.

    SConstruct should be refactored so these are created as part
    of the normal build and normally cleaned.

This is not a 3.10 blocker.

> On NetBSD and FreeBSD, I end up with a zillion files owned by root.  I assume 
> that opening a file for writing updates the owner.
> -rwxr-xr-x  1 root    users   12705 Nov 13 12:18 gps2udp
> -rw-r--r--  1 root    users    7464 Nov 13 12:18 gps2udp.o
> -rwxr-xr-x  1 root    users   20163 Nov 13 12:18 gpsctl
> -rw-r--r--  1 root    users   16136 Nov 13 12:18 gpsctl.o
> -rwxr-xr-x  1 root    users   43669 Nov 13 12:18 gpsd
> -rw-r--r--  1 root    users   36284 Nov 13 12:18 gpsd.o
> ...
> I assume chrpath is writing the executables, but who/why is writing the .o-s?

Dunno.  Sounds like the BSD build chain is pulling some weird trick.  You
might be able to figure out where it's happening by watching the generated
installation commands and simulating them by hand.

I must say I really appreciate you being picky about these details.  It's
difficult for me to stand back enough to notice them, and that last 1%
of polish can make a significant difference in the perceived friction
cost of using the package.
                <a href="";>Eric S. Raymond</a>

reply via email to

[Prev in Thread] Current Thread [Next in Thread]