Experimental Physics and
Industrial Control System

Andrew Johnson <[email protected]> · Tue, 9 Dec 2008 12:11:22 -0600

Hi Remi,

On Tuesday 09 December 2008 11:30:18 Remi Machet wrote:
>
> I would like to add support for shell variables in convertRelease.pl
> (see example patch bellow that is against R3.14.9).

One problem with including this patch in Base is that it would affect the 
production environments at other sites as well as your development 
environment, so although it would make things easier for you, it could break 
someone else's production build system (they might now be relying on 
convertRelease.pl *not* reading environment variables, although I do admit 
this would be a little unusual).

> I have been told that this issue was already raised some time ago but
> nothing was changed. Looking at the archives I found a post from Ralph
> Lange on Tue, 27 May 2003 titled "Re: convertRelease.pl" which makes a
> good case against using shell variables.
>
> But in my opinion Ralph makes a case about why some of us should not use
> shell variables, not why no one should be able to ... I totally agree
> that using shell variables in a production environment can be dangerous
> and is in my mind not worth it, but in a development environment it can
> be extremely powerful (for example to find out at which version of base
> an IOC can work with).

I have a question about your environment - how are you setting those variables 
in the first place?  The convertRelease.pl script does support both include 
and -include statements in the configure/RELEASE file, so might there be some 
way to have those variables reflected in another file that just gets included 
by the RELEASE file?  I believe the path to the include file can be relative 
or absolute, so at one extreme you could have all your configure/RELEASE 
files just include a common file somewhere that just sets variables for all 
your installed support modules.

With a bit more information about what you're trying to achieve we may be able 
to come up with a solution for you.

> I am sorry if I am re-opening an already closed issue, but I believe
> that supporting shell variables, while not for every project could
> greatly help some of us.

Discussion is good, as long as it is informed by history - you did your 
research first, so no apology necessary.

- Andrew
-- 
The best FOSS code is written to be read by other humans -- Harold Welte

Experimental Physics and Industrial Control System

Experimental Physics and
Industrial Control System