Boost logo

Boost :

Subject: Re: [boost] what happened to the "tools/regression" directory - and the stuff in it?
From: Rene Rivera (grafikrobot_at_[hidden])
Date: 2015-03-31 21:58:28

On Tue, Mar 31, 2015 at 2:08 PM, Robert Ramey <ramey_at_[hidden]> wrote:

> Beman Dawes wrote
> > On Thu, Mar 26, 2015 at 1:39 PM, Robert Ramey &lt;
> > ramey@
> > &gt; wrote:
> >> I repackaged this is a git project in the form compatible with
> > <boost
> >>
> > directory>/tools/...
> >>
> >>
> >>
> >> So you can clone it into this directory, build it via boost build and/or
> >> CMake, copy the executable to a convenient place and start generating
> >> your
> >> local test table.
> >>
> >> If there's interest, we can add it back into the boost tools directory.
> >
> > Robert,
> >
> > Why on earth you do that rather than just cloning
> > git@
> > :boostorg/regression.git and using it directly?
> Rene had told me that the directory was obsolete and for that reason it was
> dropped.

Not exactly what I said :-)

I found it and resurrected it on my system.
> I looked at and it contains a huge number of files whose purpose I couldn't
> understand. Also the directory structure was quite different that than
> of the other tools files which I wanted to replicate.

I've been restructuring (but very carefully as to not interrupt the release
schedule) in order to separate out the "user" level tools vs. the
regression testers and reports users. With the hope of making it easier to
maintain and use. And that's still a work in progress. So most things are
not obsolete in the regression.git repo. The one thing that is obsolete
(and will removed soonish) is the XSLT code... It's all been replaced by a
combination of Python and C++ by now.

> It seemed to me
> that I would have to undo Rene's delete then do ... what exactly?.
> I foresaw a disproportional amount of git ju-jitso in order to resurrect
> (and merge in recent changes) something that had already been
> deprecated and no one seemed to have interest in anyway. Then
> I would have to move the files and directories around. I actually
> did consider it. But it all just looked to hairy for something no one
> had an interest in but me.

I did promise to clean up the library_status docs and make it "easy" to
use.. Once I got my new machine (I have it finally and spent all weekend
restoring from my almost dead old one). So I will be looking at what you
did and incorporating some of it back into the Boost version :-)

> So I just cut the gordian knot and did it.
> Actually, it's not even clear that it will be part of boost. No one seems
> to
> be interested in it and it's never been reviewed. It was dropped and no
> one
> complained
> except me.
> That's why I did what I did what I did. I still don't seen anything
> wrong with it.

I don't see anything wrong with it either. It might sound like a waste of
time.. But since it helps you out while the Boost version of the tool is
back to usability I'd say it's worth a little waste.

> If there are problems with
> > git@
> > :boostorg/regression.git, it
> > is fine to fork it, fix the problems, then submit pull requests. But
> > there is no reason it needs to be part of the Boost super-project. I
> > guess we should have left a dummy boost-root/tools/regression
> > directory in the super-project to point people to the new location,
> > and provided explicit build instructions. I'll try to do that in the
> > next few days, but it won't happen for 1.58.0.
> Of course it's up to you, but it still doesn't seem worth your time
> to me.

Yea.. Doesn't need to be part of the super-project. I think some pointers
to it in the library author directed documentation should be enough. And it
is at least worth my time if someone else wants to spend time on it that's
fine too.

-- Rene Rivera
-- Grafik - Don't Assume Anything
-- Robot Dreams -
-- rrivera/ (msn) - grafikrobot/aim,yahoo,skype,efnet,gmail

Boost list run by bdawes at, gregod at, cpdaniel at, john at