[etoys-dev] Re: Etoys-to-go goes AWOL?

[K. K. Subramaniam]

On Tuesday 11 Aug 2009 12:25:11 pm Bert Freudenberg wrote:
> It is useful in certain circumstances. But the preferred way is to
> have Etoys pre-installed on each machine...
Nothing in my proposal harms pre-installations. So there is no question of 
exclusion.

> Having this in the script for fixed installation makes no sense to me
> at all. If an installation is broken there should not be scripting
> magic to fix it. Rather, the packaging should be fixed.
Installers do not handle run-time issues. Auto-detecting display manager, 
framebuffer, desktop manager, sound subsystems, giving sensible dialogs on 
error conditions and so on are best handled in a script rather than a binary. 
These changes are expected when, say, Etoys is launched from a mounted 
volume/share by different clients.

> We are not using this batch file approach on Windows....
Looks like you skipped the initial part:
  "Blaming the breakage on Vista is unfair..."

Batch files are not required for pre-installed code or for situations where 
runtime environment does not change post-install (e.g XO, personal notebooks). 
In my situation and in the topic quoted, the runtime env was different from 
what Installers assumed for fairly valid reasons. Editing paths in a small 
files was something that teachers could handle so I opted for a short batch 
file.

Subbu