Re: Big, fat statically-linked gnumeric executable

[Simon Taylor <simon unisolve com au>]

Jody, (and others)

> While it is feasible to generate a staticly linked version of
> gnumeric you will need to support dynamic linking to use the
> plugins.
>
> The number of shared libs is a _good_ thing.  It means that we are
> actually reusing code finally.  Yes it is a bit daunting if none of
> the gnome libraries are installed, but it is not as bad as it seems.
> Once the compilation process works for one of them the rest should
> fall like dominos.

It's not that I have problems building each new gnumeric as it comes out, I'm 
a software developer and though I don't know much at all about the gnome 
environment, I can usually get things built OK.

What I'm trying to do however, is to set up a web page where people who are 
not developers can simply download and evaluate the latest gnumeric 
executable.

RPMs are ok, but sometimes these introduce dependency issues of their own, 
which can be quite daunting for the kind of new user I want to help.

It's funny how many developers don't seem to understand the need for this. We 
do need to help unsophisticated users (unsophisticated only in terms of 
software development that is), to see what a great spreadsheet gnumeric is.

A single hyperlink that lets such a user download the latest stable version of 
the program and then *run* it,  is what I want to provide, (or have someone 
else provide). 

In the interests of fairness, I've just downloaded the most recent non-source 
gnumeric rpm I can find, and here is what it gives me on a Red Hat 7.2 
machine.

# rpm -Uvh gnumeric-1.0.5-3.i386.rpm
error: failed dependencies:
        libbonobo-print.so.2   is needed by gnumeric-1.0.5-3
        libbonobo.so.2   is needed by gnumeric-1.0.5-3
        libbonobox.so.2   is needed by gnumeric-1.0.5-3
        libgal.so.19   is needed by gnumeric-1.0.5-3
        libgnomeprint.so.15   is needed by gnumeric-1.0.5-3
        liboaf.so.0   is needed by gnumeric-1.0.5-3

This is quite reasonable of course, but it's too much for the casual user, 
IMHO.

As for the use of plugins, I think this might be a fair trade off. If a new 
user likes the spreadsheet enough, this'll give them the motivation they need 
to build it from rpms, or perhaps to install Ximian Evolution.

Anyway, my 10c worth. I'm grateful for the feedback I'm getting, and hopefully 
someone will give me a few tips here and there... 

Regards,

Simon Taylor
-- 
Unisolve Pty Ltd - Melbourne, Australia
+61 3 9568 2005