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

Re: FESCo Meeting Summary for 2007-02-08



Jesse Keating wrote :

> On Wednesday 14 February 2007 11:56, Matthias Saou wrote:
> > What's the rationale behind this? I fail to see the purpose, since this
> > is something that needs to be addressed better than it is currently,
> > and ideally from inside rpm itself (and apparently it's being done,
> > thanks Bill). It's even being subject to current discussions, like
> > using mktemp for it...
> >
> > As far as I'm concerned, any directory inside %_tmppath with a name
> > which is package _and_ version specific is fine. And before adding "user
> > specific too", I'd definitely add "arch specific" as I find it more
> > important, so this is a never ending and useless debate...
> >
> > I'm asking because the usual "you should use this buildroot" I usually
> > get from people reviewing my packages has become "you must", which does
> > annoy me because I fail to see the point for such a change :
> > https://bugzilla.redhat.com/228294
> 
> Because without using a mktemp style buildroot, there are technical risks of 
> overlapping directories from either multiple users or multiple arches 
> building on the same system.  Since all arguments around what BuildRoot 
> should be are centered around this, we need to pick _one_ that is safe, and 
> is the same as what the patches sent to rpm would use, once / if they get 
> integrated.  As stated, it is a simple one line change.

Why pick _one_ when it's trivial to evaluate the level of safety a given
buildroot has for the Fedora build system? That extra "id" execution is
totally useless with mock, and I simply dislike adding useless stuff.
Any useless stuff, anywhere.

%{_tmppath}/%{name}-%{version}-%{release}-root

Is shorter than what is now _mandatory_, just as safe/unsafe with mock,
which is what we encourage all users to use for builds anyway. And it
saves that useless "id" execution.

Which again is why I'd like to propose this for the short term, until
we can rip out all BuildRoot: lines for good :

> * Mandatory : The BuildRoot must start with
> %{_tmppath}/%{name}-%{version}-%{release}. The preferred value is
> %{_tmppath}/%{name}-%{version}-%{release}-root-%(%{__id_u} -n).

I like sane "guide" lines, not arbitrary "fixed" lines that make little
to no sense.

If a mktemp based buildroot which actually fixes real issues is found,
I'll have absolutely no problem switching to that!

Matthias

-- 
Clean custom Red Hat Linux rpm packages : http://freshrpms.net/
Fedora Core release 6 (Zod) - Linux kernel 2.6.19-1.2895.fc6
Load : 1.25 1.29 1.17


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