[Ecls-list] why not make a release?

[Juan Jose Garcia-Ripoll]

On Tue, Nov 27, 2012 at 4:01 PM, Anton Vodonosov <avodonosov at yandex.ru>wrote:

> In other words the version advertised on ECL web site as the release should
> be the best known of existing versions.
>

Let me clarify the criterion I have used so far for making releases:

* Sufficient number of fixes.
* I have managed to test the release on all supported platforms.
* I have managed to test and not increase the number of broken libraries.
* Maxima builds and tests fine.

This has to be done because accumulating fixes does not mean everything
stays the same -- sometimes one fix breaks other things --. And so far,
with the current git/CVS tree, points 2-4 from this list are not checked.

In principle one could do more frequent releases but verifying cycles 2-4
usually take several days for me, specially when I cannot rely that
automated tests produced the right results. cl-test-grid helps with #3
currently and hopefully I will find a free weekend during my approaching
holidays to do #2 and #4, but you realize it is not that easy -- which is
why other projects have a release manager.

Juanjo

-- 
Instituto de Física Fundamental, CSIC
c/ Serrano, 113b, Madrid 28006 (Spain)
http://juanjose.garciaripoll.googlepages.com
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <https://mailman.common-lisp.net/pipermail/ecl-devel/attachments/20121128/cf7bcb70/attachment.html>