Re: [IBMiOSS] Python 2.7 and Cheetah

[Pete Helgren <pete@xxxxxxxxxx>]

Thanks John.  I haven't given up, I am just working on several component 
issues at the same time.  So, I appreciate your pointers.  In this 
particular case, I think I *might* be able to skip Cheetah completely 
because I don't plan to use that particular feature in the app I am 
attempting to install on IBM i.  I am not sure yet because the database 
portion of the app, the most important piece, can't be installed because 
of the way the MySQL was compiled on IBM i (the version included with 
PHP) and the alternative, sqlite, can't be used because of the way that 
Python was compiled for IBM i.  Kevin gave me a couple of workarounds 
that involved installing either a different MySQL or a different Python, 
but I am *trying* to use the "off the shelf" stuff that a typical IBM i 
OSS user would have.  It is already a bit daunting to get the IBM i OSS 
stuff running without having to install a bunch of workaround packages 
that make IBM i OSS less attractive (IMHO).

I am not a Python geek by any stretch but I always enjoy tackling a new 
project so right now I am hacking some of the database code for the app 
and seeing of I can get it to happily use the ibm_db module which would 
be cool, but not knowing the app's internals, might be a stretch.

I WILL use your approach if the app requires the use of Cheetah, but I 
think I can get away with not installing it.  I won't know until I can 
get a successful DB connection.....

Pete Helgren
www.petesworkshop.com
GIAC Secure Software Programmer-Java
Twitter - Sys_i_Geek  IBM_i_Geek

On 12/30/2016 9:42 AM, John Yeung wrote:
> Pete,
>
> Sorry for the late response; don't know if you already found a
> workaround or if you've given up.
>
> Cheetah is essentially a pure-Python package. There is one optional C
> module (which serves as a high-performance replacement for
> NameMapper.py). Cheetah runs fine without it. So if you just ignore
> that piece of it, you can "install" Cheetah the traditional low-tech
> way, which is to fully unpack the compressed tarball (you can find it
> at <https://pypi.python.org/pypi/Cheetah/2.4.4>), find the "cheetah"
> directory, and move or copy that to a place on the Python search path,
> typically the site-packages directory.
>
> (Note that there is a directory within cheetah that contains both
> "cheetah.h" and "Cheetah.h" as two distinct files. Clearly this was
> meant for a full-blown Unix or Linux, and can't exist in the IFS.
> Depending on what you're using to unpack the tarball, you may be
> prompted to overwrite one with the other. Go ahead and do so.
> Ultimately, that directory will be ignored anyway because it's just
> for building the C module I mentioned.)
>
> Now, bear in mind I don't actually use Cheetah myself. I'm just
> speaking as someone who (because I use iSeriesPython) only has the
> "dump the package source into site-packages" method of installation
> available.
>
> If you plan on using Cheetah in a standalone capacity, you may also
> want to put the contents of the included bin directory into a place on
> your QSH or PASE path (not the Python search path). They are intended
> to be executable scripts.
>
> Finally, I will note that while Cheetah was generally a very
> well-liked package, it hasn't been maintained in years, and never made
> the transition to Python 3 support (which I suppose is why you're
> using Python 2.7 for this). I take it you don't really intend to use
> Cheetah yourself, but rather need it as a dependency for something
> else? (Because if you want a templating engine for your own use, there
> are definitely other choices.)
>
> John Y.