<?xml version="1.0" encoding="utf-8"?>
<feed xmlns="http://www.w3.org/2005/Atom">
	<title type="html"><![CDATA[HyperForum — [Hyperbola] Packages listed to be risk with upgrades]]></title>
	<link rel="self" href="https://forums.hyperbola.info/extern.php?action=feed&amp;tid=1169&amp;type=atom" />
	<updated>2026-03-26T14:12:59Z</updated>
	<generator>PunBB</generator>
	<id>https://forums.hyperbola.info/viewtopic.php?id=1169</id>
		<entry>
			<title type="html"><![CDATA[Re: [Hyperbola] Packages listed to be risk with upgrades]]></title>
			<link rel="alternate" href="https://forums.hyperbola.info/viewtopic.php?pid=8784#p8784" />
			<content type="html"><![CDATA[<p><span class="bbu"><strong>Python</strong></span></p><p><strong>Draft solution:</strong> Keep track with <strong>python-setuptools</strong> and create our own pip-repository inside repo.hyperbola.info, modify pypi to point to our local pip-repository and finally (soft-)fork pip to enable <strong>python-build</strong> and <strong>python-installer</strong>.</p><p>Advantage within is a way forward, disadvantage is more amount of work to keep track for separated repositories and file-management for the tarballs to be downloaded at build-time.</p><p><strong>WARNING:</strong> This is neither an ideal nor a very safe way and just a first draft. This solution includes a possible violation of free software distribution guidelines and furthermore also a spot making packages later on part of the system users have no real influence and / or transparency about.</p>]]></content>
			<author>
				<name><![CDATA[throgh]]></name>
				<uri>https://forums.hyperbola.info/profile.php?id=347</uri>
			</author>
			<updated>2026-03-26T14:12:59Z</updated>
			<id>https://forums.hyperbola.info/viewtopic.php?pid=8784#p8784</id>
		</entry>
		<entry>
			<title type="html"><![CDATA[[Hyperbola] Packages listed to be risk with upgrades]]></title>
			<link rel="alternate" href="https://forums.hyperbola.info/viewtopic.php?pid=8783#p8783" />
			<content type="html"><![CDATA[<p>Hello, within this thread we want to give it a try (incremental) to list all packages going forward with a conflict in later releases for Hyperbola sooner or later. Out of experience and clear research this listing will surely grow over time and change further. So discussions are not helping here, should be done in IRC or elsewhere. Reasonings and information are nevertheless welcome and will be included. Please at that point with clear sources (links and articles) then.</p><p><strong>On that matter of using more than once programming-language at the same software-project:</strong> So-called &quot;modern&quot; implementations have a severe tendency to add just more than one language for building. So besides C and / or C++, Python, Perl, Rust and many more. Adding such complexity-level makes the whole project more complicated to audit, understand and ready for &quot;easy&quot; modifications. Users on the search to become their own developers get more blocks in the way and mandatory&nbsp; efforts are being blocked out technical and social. Then resulting also within more points to look on solutions being clearly no solution likewise LLM / machine-learning to overcome complexity - like an illusion running into more needs and dependencies. So on the statement of <strong>minimalism</strong> Hyperbola is making this should be at all costs kept out.</p><p><span class="bbu"><strong>High critical (build-infrastructure)</strong></span></p><p><strong>gcc:</strong> In need for a clear research as newer releases include Rust-definitions for a later inclusion of an own Rust-compiler, until unclear how dependencies should be resolved at build-time for Rust-projects later version of <strong>gcc</strong> (beyond version 12.x) need further evaluation</p><p><strong>python:</strong> Newer releases are surely possible, but since version 3.9 there is no further support for <strong>libressl</strong> exclusive provided and need patching, furthermore based packages on <strong>python</strong> are using <strong>python-installer</strong> and <strong>python-build</strong> with clear inclusion of <strong>pip</strong>, this collides with Hyperbola&#039;s essential notes (quote from blacklisting: <a href="https://git.hyperbola.info:50100/software/blacklist.git/plain/blacklist.txt)">https://git.hyperbola.info:50100/softwa … klist.txt)</a></p><div class="quotebox"><blockquote><p>python-pip::hyperbola:1294:[uses-nonfree] supports and recommends nonfree software</p></blockquote></div><p><strong>meson:</strong> Newer releases added an own package-repository for referencing at build-time (for reference: <a href="https://mesonbuild.com/Wrapdb-projects.html),">https://mesonbuild.com/Wrapdb-projects.html),</a> as Hyperbola has already a pakage-manager and for clear security-reasonings this should be completely excluded and is therefore in need for research if even possible to be complete removed, otherwise this package is another one in need to be frozen</p><p><strong>cmake:</strong> Newer releases will start adding an own package-management, for reference: <a href="https://www.phoronix.com/news/CMake-Tighter-Package-Integrate">https://www.phoronix.com/news/CMake-Tig … -Integrate</a></p><br /><p><span class="bbu"><strong>Medium critical (dependencies, frameworks and libraries)</strong></span></p><p><strong>git:</strong> With upcoming release of version 3.0 announced to make <strong>rust</strong> mandatory for build-infrastructure (see here: <a href="https://lore.kernel.org/git/20250904-b4-pks-rust-breaking-change-v1-0-3af1d25e0be9@pks.im/),">https://lore.kernel.org/git/20250904-b4 … @pks.im/),</a> as long as there is no finalized reference-implementation being clear free from trademark- and copyright-issues there is no way around that</p><p><strong>librsvg:</strong> Is in need of <strong>rust</strong> for building, therefore Hyperbola preserved the last working version without <strong>rust</strong> as <strong>librsvg-legacy</strong> which then resulted also as example for later on added so-called &quot;legacy-packages&quot; kept frozen in a concrete version-number</p><br /><p><span class="bbu"><strong>Low critical (applications)</strong></span></p><p><strong>inkscape:</strong> With version 1.1 added pre-defined templates for non-free services, therefore adapting non-free services on a social level and making a statement that their acceptance including their violation of user-rights and -freedoms is acceptable on different scale (in need for a decision of kept or handled different)</p><br /><p><span class="bbu"><strong>Uncritical (kept solely working)</strong></span></p>]]></content>
			<author>
				<name><![CDATA[throgh]]></name>
				<uri>https://forums.hyperbola.info/profile.php?id=347</uri>
			</author>
			<updated>2026-03-26T12:07:07Z</updated>
			<id>https://forums.hyperbola.info/viewtopic.php?pid=8783#p8783</id>
		</entry>
</feed>
