Issue 7 in orapig: automate creation of sample packages/modules
codesite...@google.com
http://code.google.com/p/orapig/issues/detail?id=7
Comment #4 by marhar:
looks good, and I like the idea of being able to uninstall to a pristine
state. I asked a windows friend about Make, and he said it's a
reasonable thing to expect.
Issue attribute updates:
Status: Fixed
--
You received this message because you are listed in the owner
or CC fields of this issue, or because you starred this issue.
You may adjust your issue notification preferences at:
http://code.google.com/hosting/settings
Henning
> I asked a windows friend about Make, and he said it's a
> reasonable thing to expect.
A typical windows installation does not contain make.
Python is usually installed from a binary distribution.
Why use a makefile when a plain script would do?
I'm not an expert for distutils, but perhaps it's possible
to call sqlplus via setup.py?
If it's not supported straightforward, one could still use
the subprocess module.
OTOH, a user can still run the SQL scripts manually,
so this is not much of an issue.
Henning
Mark Roddy
From: Mark Roddy <mark...@gmail.com>
Date: Jun 3, 2008 11:24 AM
Subject: Re: [orapig-dev] Re: Issue 7 in orapig: automate creation of
sample packages/modules
On 6/3/08, Henning <h.von...@gmail.com> wrote:
>
> marhar wrote:
> > I asked a windows friend about Make, and he said it's a
> > reasonable thing to expect.
>
>
> I don't think so.
> A typical windows installation does not contain make.
> Python is usually installed from a binary distribution.
I didn't think we would need to consider whether or not most windows
users will have 'make' install, but rather whether most windows users
who will be running the unit tests will have make installed. In which
case they're going to be some sort of developer, from which I would
infer that they will have it installed or be able to.
This is of course since I was thinking of the contents of the samples
directory as tests. We may want to consider separating into a test
and samples directory so as not to enforce any constructs on users
that simple want to try out the samples.
> Why use a makefile when a plain script would do?
In the case of an everyday user wanting to install the sample packages
to try out I agree.
When working on testing it provides a lot of benefits. It will be
much less terse then a script as it internally handles dependencies.
A script would have to contain all the code for checking for file
changes and building up dependency graphs to have all the features
that 'make' does. Also, there's a bunch of commands that have to be
executed to perform tests. A makefile allows everything to be done
for you by executing:
make install-samples
make test
In addition, if there are an changes, say to file names or adding a
new sample package, we don't have to worry about any new commands that
need to be executed as the makefile can be updated.
> I'm not an expert for distutils, but perhaps it's possible
> to call sqlplus via setup.py?
> If it's not supported straightforward, one could still use
> the subprocess module.
I haven't tried to execute an entire sql script from python, but that
doesn't seem out of the range of posiblities. I'm all in favor of
using distutils to make installers, but can it be made so that
installing samples is optional? I wouldn't want to force that on
users.