[Distutils-sig] distutils charter and interscript
Tue, 01 Dec 1998 09:22:15 +1000
From the charter:
>will take the form of some standard modules, probably grouped in the
>'distutils' package, for building, packaging, and installing third-party
>modules. This includes both modules written purely in Python and
>extension modules written in C/C++.
Interscript is designed to do all this. Except, it has a much wider
scope: it isn't limited to building Python, and it includes testing
and documentation. (And a few other things -- some of which are implemented
already, and some of which are not [such as version control])
There's a generic description of requirements in the documentation at
From the topic list:
> dist - create a source distribution
Almost done: interscript creates a source tree in the doco.
It doesn't yet provide the ability to 'tar' up those files,
but that is fairly trivial.
> bdist - create a 'built distribution' (the equivalent of a 'binary
> distribution', except that binaries won't necessarily
> be present
As above, except that this time, a list of generated files is available.
Again, there is not yet a tool for taring up those files.
> install - install a built library on the local machine
This is MUCH harder. If you read my documentation, and examine the
sections on installation control (site-frame, platform-frame, user-frame)
you will see I have made fairly advanced provision for installation control.
I'm not using any of this yet.
My system discriminates the status of the package, and may install
it in different places depending on the status. For example,
when I download a new version of a package, I might put it into
a test directory and test it, before I install it into a more
widely accessible place. Furthermore, the install point is conditioned
by the authors evaluation: is it alpha, beta, or production software?
> gen_make - generate a Makefile to do some of the above tasks
> (mainly 'build', for developer convenience and efficiency)
Please don't! Please generate PYTHON script if it is necessary.
Python runs on all platforms, and it has -- or should have --
access to all the configuration information.
If this is not the case, we need to petition Guido to make that
information available, or to provide hooks that allow the client
to provide that information ONCE on installing python.
Make is extremely unreliable: it isn't available on all platforms,
it doesn't work properly on all platforms, it is almost impossible to
debug unless you are a Make programmer.
We need to get as far aware from this as possible.
We have something MUCH better already. Python. Lets use it.
>I will tentatively put forward March 1999 as a target for a working
>release to run on top of Python 1.5.2, which *might* require a
>patch-and-rebuild-Python step (or might find ways to work with the
>information available under 1.5). For a complete, tested, documented
>suite ready to bundle with Python 1.6, let's shoot for June 1999.
I don't believe a patch is necessary: what will almost certainly
be necessary, however, is that there be some kind of support module
that provides relevant information. And that module will have to be
written according to specifications, BY THE CLIENT. Of course,
standard defaults can and should be provided :-)
For example, the compilers module I provide happens to
use gnu gcc/g++. So it won't work on NT. Or a Solaris system
where the client wants to use the native cc. Etc etc.
>Other topics of interest:
> * encouraging module developers to write test suites by having
> a standard place for them in module distributions
Yes. Although interscript allows a slightly different approach:
test code is embedded directly in the source and is executed during
the build process. The test code isn't installed anywhere; it's
written to a temporary file for testing and then deleted afterwards.
This technique is not contrary to the idea of a stand-alone
test suite. BTW: 'standard place' needs to be more flexible that
a fixed subdirectory!
> * ditto for documentation -- although this is probably the job of
> the doc-sig, it would be nice to tie the two together at some point
Again, I agree. Interscript, as a literate programming tool,
emphasises documentation. And again, 'standard place' needs to be
more flexible than just a fixed subdirectory. [For example,
if you have HTML doco, and want networked clients of the software
to see it, it may have to be put into the httpd root: inside
the actual package would not work]
> * a standard for representing and comparing version numbers
I think this is very hard. It isn't enough to just use numbers.
RCS/CVS provides a tree. With labelled cut points. AFAIK, it cannot
handle the important thing: dependencies.
In other words, it isn't enough to just have a standard for
version identification, although that is a good start.
It is also necessary to determine what depends on what. RedHat RPM
tries to do that. It is fairly sophisticated, I'm told, but it
isn't very reliable: it doesn't know about non-rpm installation,
and it has stuffed up my Linux box badly.
> * social engineering in general, ie. convincing module developers
> to start using the system
I've already implementd half what this SIG's charter wants.
Now how can I convince you to use what I've done instead of
re-inventing the wheel?
> * the possible need for tweaks to the configure/build process for
> Python itself, and a place to hold configuration
> information (possibly a new built-in module, 'sys.config')
Yes. I believe the build needs to be split into two parts:
1) A nasty, hackers build of the bootstrap version
2) a clean python based build of the rest of the system
At present, the whole installation is in category (1): and it isn't
necessary. (1) cannot be dispensed with. But it can be reduced
for an initial installation, and reduced even more for an upgrade.
I think this target should be kept in mind, but otherwise deferred for a while:
if a suitably powerful and flexible system is developed, Guido will probably
use it. But this cannot happen until the system is tested,
distributed as standard, and been thoroughly ingrained in the
building of third party extensions. [IMHO: Guido mught surprise me:-]
That's because it is likely to be 'too risky' before that time:
it isn't so simple because the installation system has to be
built correctly _first_ before it can be used to build the rest of Python.
And it is likely that configuring the installation system is harder than
John Skaller email: email@example.com
snail: 10/1 Toxteth Rd, Glebe NSW 2037, Australia