Red Hat Leftovers
-
Lubomír Sedlář: Today I Learned: Searching in Vim
Ok, I lied. Searching in Vim is something I do all the time, so it’s not such a new thing. However there is a feature that I only need every once in a while, and always forget how to do it.
The goal is to highlight something in the code, you can just search for it and if
hlsearchis on, it will shine with the light of thousand suns. -
Lubomír Sedlář: Introducing entr
When writing code, my workflow is usually to make edits and re-run the test suite (that is, if there is one). This is actually a lot of work that I don’t want to do.
Thankfully, it’s quite easy to automate. All we need is a utility that would watch files for change and then run the tests.
On Linux, we can somewhat get by with inotifywait, but it is rather tedious. There must be a better way; and there actually is.
-
Lubomír Sedlář: A compose what?
Kushal wrote a nice post about how package updates get into repositories. However he glosses over rawhide by saying it happens automatically. This is true, but I want to elaborate on the mechanism of how it happens..
A package build for Rawhide is not available in the repositories immediately. It has to wait until a nightly Rawhide compose runs that updates the repo.
Let’s take a dive into what exactly the compose does and is.
Terminology
A compose is a snapshot of a release with a specific identity (which is determined by the actual release and date when it was created, like
Fedora-Rawhide-20170710.n.0). It consists of at least one variant (such asServer,WorkstationorEverything). A variant is a subset of the content from a release that is aimed for particular use case.Generally a variant in the final compose contains a repo with RPMs and other artefacts like images of various kinds.
-
Lubomír Sedlář: Today I Learned: Handling source tarballs
Fedora packages are generally built from tarballs provided by upstream developers. What’s actually happening to these tarballs and how do they move around?
How are sources uploaded?
When a new tarball is supposed to be used, the packager will upload it to lookaside cache (because storing big binary files directly in git is bad). This is done with
fedpkg uploadorfedpkg new-sources. These commands make aPOSTrequest to a CGI script running on the dist-git server. This requests includes the name of the file, a checksum and the actual file contents.Up until today (Dec 12) MD5 hash has been used. Now it’s SHA512.
-
Lubomír Sedlář: Clickable Pungi logs
When debugging problems with composes, the logs left behind by all stages of the compose run are tremendously helpful. However, they are rather difficult to read due to the sheer volume. Being exposed to them quite intensively for close to a year helps, but it still is a nasty chore.
The most accessible way to look at the logs is via a web browser on kojipkgs. It’s just httpd displaying the raw log files on the disk.
It took me too long to figure out this could be made much more pleasant that copy-pasting stuff from the wall of text.
How about a user script that would run in Greasemonkey and allow clicking through to different log files or even Koji tasks?
-
Fedora Project ☛ Fedora Community Blog: Fedora datacenter move later this year (2025 version)
The Fedora Project has instances in a number of datacenters and clouds all over the world, but a majority of instances are in a datacenter located in Virginia, USA. This datacenter space, along with the majority of servers in it, were generously provided by our primary sponsor, Red Hat. We moved to our current space from another Red Bait datacenter back in 2020, and now it’s time to move again.
So why would we want to move? Well, there’s a number of reasons: [...]
-
Jonathan Dowland ☛ Jonathan Dowland: FOSDEM 2025
I'm going to FOSDEM 2025!
As usual, I'll be in the Java Devroom for most of that day, which this time around is Saturday.
-
Xe's Blog ☛ Life pro tip: Oracle GNU/Linux is the best local VM for MacBooks
No, seriously. It sips battery life. I don't get it either.