[Git][reproducible-builds/reproducible-presentations][master] fosdem 2024: wip wip wip
Holger Levsen (@holger)
gitlab at salsa.debian.org
Fri Feb 2 18:05:36 UTC 2024
Holger Levsen pushed to branch master at Reproducible Builds / reproducible-presentations
Commits:
1fb90cef by Holger Levsen at 2024-02-02T19:05:17+01:00
fosdem 2024: wip wip wip
Signed-off-by: Holger Levsen <holger at layer-acht.org>
- - - - -
2 changed files:
- 2024-02-03-R-B-the-first-10-years/index.html
- 2024-02-03-R-B-the-first-10-years/todo
Changes:
=====================================
2024-02-03-R-B-the-first-10-years/index.html
=====================================
@@ -956,7 +956,7 @@ Warpforge.
<h2>Detour: more unexpected benefits of reproducible builds</h2>
<ul>
<li class="fragment">https://bootstrappable.org began as breakout session at the Reproducible Builds Summit 2016 in Berlin.</li>
- <li class="fragment"><em>as I understand it</em> is about bootstrapping toolchain binaries from sources <em>only</em>, so it starts with a handwritten 500 byte sized assembler code, which builds another assembler and then assembles some more, until it can build mes, which can build tinyCC, which can build an ancient GCC which then can build an another ancient GCC, which then can be used to build modern GCC and the rest of the universe.
+ <li class="fragment"><em>as I understand it</em> is about bootstrapping toolchain binaries from sources <em>only</em>, so starting from around 500 handwritten bytes of machine code, a very simple assembler is build and then assembled some more, until it can build mes, which can build tinyCC, which can build an ancient GCC which then can build another ancient GCC, which then can be used to build modern GCC and the rest of the universe.</li>
</ul>
</section>
@@ -1089,18 +1089,9 @@ Warpforge.
<h2>Debian policy</h2>
<ul>
<li>2017: packages <em>should</em> build reproducibly.</li>
- <li class="fragment">2024? reproducible packages <em>must not</em> regress.</li>
- <li class="fragment">2026? NEW packages <em>must</em> build reproducibly (to be allowed into <code>testing</code> and therefore into <code>stable</code>).</li>
- <li class="fragment">2028? packages <em>must</em> build reproducibly (to be allowed into <code>testing</code> and <code>stable</code>).</li>
- </ul>
- </section>
-
- <section data-background="images/FOSDEM_logo.svg" data-background-size="10%" data-background-position="93% 9%" data-transition="none">
- <h2>Debian policy</h2>
- <ul>
- <li>2017: packages <em>should</em> build reproducibly.</li>
- <li>2024? reproducible packages <em>must not</em> regress. NEW packages <em>must</em> build reproducibly (to be allowed into <code>testing</code> and therefore into <code>stable</code>).</li>
- <li>2027? packages <em>must</em> build reproducibly (to be allowed into <code>testing</code> and <code>stable</code>).</li>
+ <li class="fragment">2025? reproducible packages <em>must not</em> regress.</li>
+ <li class="fragment">2025? NEW packages <em>must</em> build reproducibly (to be allowed into <code>testing</code> and therefore into <code>stable</code>).</li>
+ <li class="fragment">2027? packages <em>must</em> build reproducibly (to be allowed into <code>testing</code> and <code>stable</code>).</li>
</ul>
</section>
@@ -1108,11 +1099,26 @@ Warpforge.
<h2>100%!</h2>
<ul>
<li class="fragment">100% reproducible is a political decision and nothing technical.</h3>
- <li class="fragment">Thus we need to change <code>debian-policy</code>!</li>
- <li class="fragment">Thus <em>Debian</em> needs to change <code>debian-policy</code>!</li>
+ <li class="fragment">We need to change <code>debian-policy</code>!</li>
+ <li class="fragment">But we can have whitelists in the beginning.</li>
</ul>
</section>
+ <section data-background="images/FOSDEM_logo.svg" data-background-size="10%" data-background-position="93% 9%" data-transition="none">
+ <h2>Debian <em>testing</em> migration</h2>
+ <ul>
+ <li>since the end of 2023, CI reproducible-builds results are included in the excuses output for Debian testing migration, but there is no penalty nor bonus yet.</li>
+ <li>in 2025 for Debian 14 "<em>forky</em>" however there could penalties for violating:
+ <ul>
+ <li class="fragment">reproducible packages <em>must not</em> regress.</li>
+ <li class="fragment">NEW packages <em>must</em> build reproducibly (to be allowed into <code>testing</code> and therefore into <code>stable</code>).</li>
+ </ul>
+ <li class="fragment">Surely in 2025 this would still need whitelisting some packages, but in less in 2027 and eventually we can stop those exceptions.</li>
+ </ul>
+ </section>
+
+
+
<section data-background="images/FOSDEM_logo.svg" data-background-size="10%" data-background-position="93% 9%" data-transition="none">
<h3>100% reproducibility in theory is not enough,<span style="font-size: 140%"> by far.</span></h3>
<ul>
=====================================
2024-02-03-R-B-the-first-10-years/todo
=====================================
@@ -1,8 +1,7 @@
-read policy slides
+still rough: around Then we need rebuilders.
+rebuilder-snapshot still not mentioned
2 big news 2023/2024: testing migration & rebuilder snapshot
-forky+1: unreproducible packages are still ok, but only as whitelisted exceptions?
-table trixie forky forky+1 +2
snapshot.d.o
@@ -17,6 +16,8 @@ sideremark:
someone should please do a reproducible binary fork of debian stable.
or another reproducible distro!
+Policy is not a stick to beat developers with.
+The release team is nicer: very nice package you're having there, wouldn't it be a sad if this package would not be allowed in Debian testing and thus stable?<
what is success?
View it on GitLab: https://salsa.debian.org/reproducible-builds/reproducible-presentations/-/commit/1fb90cef4f2ec1de3b3dcffb2abf12b226e4eb4d
--
View it on GitLab: https://salsa.debian.org/reproducible-builds/reproducible-presentations/-/commit/1fb90cef4f2ec1de3b3dcffb2abf12b226e4eb4d
You're receiving this email because of your account on salsa.debian.org.
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.reproducible-builds.org/pipermail/rb-commits/attachments/20240202/b6e40718/attachment.htm>
More information about the rb-commits
mailing list