<!DOCTYPE HTML PUBLIC "-//W3C//DTD HTML 4.01 Transitional//EN">
<html>
<head>
<meta content="text/html; charset=UTF-8" http-equiv="Content-Type">
</head>
<body bgcolor="#ffffff" text="#000000">
Hi Hilaire,<br>
<br>
On 5/22/2024 12:04 PM, Hilaire Fernandes via Cuis-dev wrote:
<blockquote cite="mid:4a4185da-52b2-4947-ab00-d3ff00ddaddc@free.fr"
type="cite">
<meta http-equiv="Content-Type" content="text/html; charset=UTF-8">
<p><font size="4">Hi, <br>
</font></p>
<p><font size="4">I think we need some clarification and facing
reality on what can be done.</font></p>
</blockquote>
<br>
Ok. I like your realistic approach.<br>
<br>
What I was suggesting (thanks Ken and Mariano for commiting your
time to that) is more ambitious, difficult, and perhaps not needed.<br>
<br>
<blockquote cite="mid:4a4185da-52b2-4947-ab00-d3ff00ddaddc@free.fr"
type="cite">
<p><font size="4">Regarding a given Cuis stable release and a
given package, being part of Cuis-Smalltalk organisation or
not, I think about two different aspects of the package:</font></p>
<p><font size="4">1. being compatible, known to work, with the
given Cuis stable release</font></p>
<p><font size="4">2. being maintained for this given Cuis stable
release, its code will evolve within this branch, I guess this
is what you call LTS<br>
</font></p>
<p><font size="4">I have absolutely zero resource to do #2, like
maintaining LTS for Cuis-Smalltalk-UI for example . I have
difficulty to find free time to work on DrGeo or Dynabook.<br>
</font></p>
</blockquote>
<br>
We can try #1. That is, right after we freeze the stable release
(that should be done in no more than a couple of weeks), we can
review 3rd party packages, and after some testing and quality
assessing, we can tag their current commit as "Cuis7.0Stable ready".
Maybe that is enough. We don't have yet enough experience to know if
more than that is needed.<br>
<br>
If we actually need to go #2, yes, we'll need to open branches on
the affected repos.<br>
<br>
<blockquote cite="mid:4a4185da-52b2-4947-ab00-d3ff00ddaddc@free.fr"
type="cite">
<p><font size="4"> </font></p>
<p><font size="4">Discussion continue below.<br>
</font></p>
<div class="moz-cite-prefix">Le 21/05/2024 à 15:25, Juan Vuletich
a écrit :<br>
</div>
<blockquote type="cite" cite="mid:664CA0CD.8010100@cuis.st"><br>
But I think it is more important to find out which '3rd party'
packages we want (and are able to!) support for a stable
release. Stable releases are meant to be high quality, and users
expect them to work for a long time. They are "LTS" releases
(Long Term Support). We can only include packages we know are in
good shape, and someone has committed to provide support for
them as part of the stable release. We haven't been explicit
about this before, basically because I'll be doing that for the
packages in the Cuis7-0 repository. But I can't do that for
packages I don't use and don't know too much about how they
work. <br>
</blockquote>
<p>Sure you can't and you shouldn't do that. </p>
<p>A maintainers of a package should decide if they want to
inform about #1 and/or volunteer to #2 or none, in this later
case package only evolve with -dev version. </p>
<p>I have no idea how this should be done but a common rule,
description how to do it will be useful to both maintainer and
user of any given package for Cuis. I guess you wan to rely on
git capability.<br>
</p>
</blockquote>
<br>
Yep.<br>
<br>
<blockquote cite="mid:4a4185da-52b2-4947-ab00-d3ff00ddaddc@free.fr"
type="cite">
<p> </p>
<blockquote type="cite" cite="mid:664CA0CD.8010100@cuis.st"> <br>
The next step needs to be to identify those packages. I know,
(this is just an example), that GeographicInformationSystems is
cool. But I haven't used it in quite some time. So unless
someone else has, it doesn't qualify as "well maintained and
routinely used". Obviously, if there is some project actually
using it, the situation is completely different. So we need that
list of packages to work on. <br>
</blockquote>
<p>Even if not used, such package could be tagged to be known to
work, or may be just load ?<br>
</p>
</blockquote>
<br>
Yes. But I'd devote at least one hour, after Cuis7.0Stable is out,
to check that it loads, all tests pass, and example scripts do work.
Then we can tag it.<br>
<br>
<blockquote cite="mid:4a4185da-52b2-4947-ab00-d3ff00ddaddc@free.fr"
type="cite">
<p> </p>
<blockquote type="cite" cite="mid:664CA0CD.8010100@cuis.st"> <br>
I took a not-too-old DrGeo image I had on my machine, and looked
at the packages it has loaded. Hilaire please correct me if this
is wrong or incomplete. Some of these packages are already in
the Cuis7-0 repo. Others are specific to DrGeo. I'm not listing
any of them in what follows. <br>
<br>
Many packages are in other (3rd party) repos in the
Cuis-Smalltalk organization. So, these are the repos / packages
we need to work on: <br>
- Numerics / 'LinearAlgebra' <br>
- SVG / 'SVG' <br>
- Parsers / 'PetitParser' <br>
- Erudite / <br>
'PetitParserBinding' <br>
'Erudite' <br>
- Cuis-Smalltalk-UI / <br>
'UI-Click-Select' <br>
'UI-Core' <br>
'UI-Entry' <br>
'UI-Panel' <br>
'UI-Preference' <br>
'UI-Widgets' <br>
<br>
I'll support Numerics / 'LinearAlgebra' and SVG / 'SVG' under
the repo structure we decide for "3rd party packages for Cuis
Stable Releases". I think they are mature and stable, and
haven't had issues in a long time. <br>
<br>
So, Hilaire and Ken, do you think that the Cuis-Smalltalk-UI
packages are stable enough to be in a stable release with long
term support? If needed, will you give support for them, not
only for the current Rolling Release, but also for Stable
Releases? These will be different package files, and most likely
on separate repos. <br>
</blockquote>
<p>As I wrote I can't do #2 on these packages.<br>
</p>
</blockquote>
<br>
It is OK. We'll find our way.<br>
<br>
<blockquote cite="mid:4a4185da-52b2-4947-ab00-d3ff00ddaddc@free.fr"
type="cite">
<p> </p>
<blockquote type="cite" cite="mid:664CA0CD.8010100@cuis.st"><br>
I apologize for asking bluntly, but we need to be confident,
before stamping "LongTermSupport'd" on any packages. <br>
</blockquote>
<p>I think an intermediate status as "known to work" with will be
more realistic.<br>
</p>
Hilaire
<pre class="moz-signature" cols="72">--
GNU Dr. Geo
<a moz-do-not-send="true" class="moz-txt-link-freetext" href="http://gnu.org/s/dr-geo/">http://gnu.org/s/dr-geo/</a>
<a moz-do-not-send="true" class="moz-txt-link-freetext" href="http://gnu-drgeo.blogspot.com/">http://gnu-drgeo.blogspot.com/</a></pre>
</blockquote>
<br>
Yes. Thanks Hilaire for your honest and thoughtful answer. This
discussion is really useful!<br>
<br>
Cheers,<br>
<pre class="moz-signature" cols="72">--
Juan Vuletich
cuis.st
github.com/jvuletich
researchgate.net/profile/Juan-Vuletich
independent.academia.edu/JuanVuletich
patents.justia.com/inventor/juan-manuel-vuletich
linkedin.com/in/juan-vuletich-75611b3
twitter.com/JuanVuletich</pre>
</body>
</html>