summaryrefslogtreecommitdiff
path: root/website/posts
diff options
context:
space:
mode:
authorLudovic Courtès <ludo@gnu.org>2019-05-17 19:17:39 +0200
committerLudovic Courtès <ludo@gnu.org>2019-05-19 23:23:00 +0200
commite391f06f514d6f148253f0f628f0107408c1bae2 (patch)
tree5e12a90f22401eb2786ff382a3cf3497d56ed4da /website/posts
parentwebsite: Add minor changes to the 1.0.1 post. (diff)
downloadguix-artwork-e391f06f514d6f148253f0f628f0107408c1bae2.tar.gz
website: guix-1.0.1: Minor updates.
* website/posts/gnu-guix-1.0.1-released.md: Change time, add link to Debian's preseeds.
Diffstat (limited to 'website/posts')
-rw-r--r--website/posts/gnu-guix-1.0.1-released.md9
1 files changed, 5 insertions, 4 deletions
diff --git a/website/posts/gnu-guix-1.0.1-released.md b/website/posts/gnu-guix-1.0.1-released.md
index 1914756..81bbc9d 100644
--- a/website/posts/gnu-guix-1.0.1-released.md
+++ b/website/posts/gnu-guix-1.0.1-released.md
@@ -1,5 +1,5 @@
title: GNU Guix 1.0.1 released
-date: 2019-05-17 17:17
+date: 2019-05-19 23:30
author: Ludovic Courtès
slug: gnu-guix-1.0.1-released
tags: Releases, System tests
@@ -22,7 +22,7 @@ Guix users can update by running `guix pull`.
It’s been just over two weeks since we [announced
1.0.0](https://www.gnu.org/software/guix/blog/2019/gnu-guix-1.0.0-released/)—two
-weeks and 702 commits by 39 people already!
+weeks and 706 commits by 40 people already!
This is primarily a bug-fix release, specifically focusing on issues in
the graphical installer for the standalone system:
@@ -171,8 +171,9 @@ The lesson here is that: manual testing should _also_ look for issues in
the graphical UI. The Debian and Guix installer UIs are similar—both
using the [Newt](https://pagure.io/newt) toolkit. Debian tests its
installer using
-[“pre-seeds”](https://wiki.debian.org/DebianInstaller/Preseed), which
-are essentially answers to all the questions and choices the UI would
+[“pre-seeds”](https://wiki.debian.org/DebianInstaller/Preseed)
+([code](https://salsa.debian.org/installer-team/preseed)), which are
+essentially answers to all the questions and choices the UI would
present. We could adopt a similar approach, or we could test the UI
itself at a lower level—reading the screen, and simulating key strokes.
UI testing is notoriously tricky so we’ll have to figure out how to get