Creating and using a custom Linux kernel on Guix System https://www.gnu.org/software/guix/blog/2019/creating-and-using-a-custom-linux-kernel-on-guix-system/
During distribution development I often faced common tasks:
1. Change the C compiler for all packages.
2. Change optimization flags for all packages.
3. Add security flags like the "-pie" flag for all executables.
4. Investigate which binaries lacks requested compilation flags.
I tried to summarize my experience with this tasks in my new article https://www.mndet.net/2019/05/11/build-systems-from-distro-developer-view.html
This time I wrote without external editing, but I hope native English speakers can understand my Runglish.
I moved a menu block to the top, hided last posts, and reduced paddings to the screen borders for cases when a screen width less then 960px. Changes in paddings sizes looks ugly if we resize desktop browser window, but small paddings looks better on my phone.
I wrote the small article that show how to create custom compile-time warnings for C code with #gcc https://www.mndet.net/2019/01/07/c-custom-warnings.html
As result of this work I made my first contribution to GCC 🙂
Finally I prepared the small article on a #bitbake #mirrors mechanism https://www.mndet.net/2018/12/28/yocto-mirrors.html
This article should close the gap in the #yocto documentation
On this week I finally understood why I hate a Judeo-Masonic conspiracy theory. Conspiracy theories should create a new exciting reality that does not contradict the facts. "Conspirology" should be a sub-genre of "secret history" literature genre and a "totalitarian world government" conspiracy theories is the worst representative of the genre.
I finally extracted all my old meaningful posts for different blogs to https://www.mndet.net
Most interesting thing is #gcc optimization issue: https://www.mndet.net/2017/06/05/gcc-optimization-issue.html
Finally got some results according to #rust package for #guix reproducibility. It's looks like rust 1.25 and rust 1.27 releases steps on some llvm 6 non-deterministic behavior (there are several bugs in this field in llvm bugtracker), while rust 1.26, 1.28, 1.29 works well with it.
this rings so true omg "Contrary to common belief, the volume of face-to-face interaction decreased significantly (approx. 70%) in [two field studies transitions to open office plans], with an associated increase in electronic interaction. In short, rather than prompting increasingly vibrant face-to-face collaboration, open architecture appeared to trigger a natural human response to socially withdraw from officemates and interact instead over email and IM." http://rstb.royalsocietypublishing.org/content/373/1753/20170239
Looks like creating #guix package for big project it's easy way to find some spots that must be fixed.
While I worked on #rust package for guix I already committed two fixes to rust compiler and opened two issues for separate packages used by rust compiler.
While I worked on package for #poky https://github.com/mnd/guix-mnd-pkgs/blob/master/mnd/packages/yocto.scm I made 4 fixes for #openembedded code.
That feeling when only code that you write in #rust is patches to rust compiler itself.
Fixed #rust build for #guixsd with glibc 2.27 and created package for rust 1.26.2 release. Waiting for approve https://debbugs.gnu.org/cgi/bugreport.cgi?bug=31955
I'm going to migrate from my email@example.com #pumpio account to this place.
System programmer from Russia, Yekaterinburg
mastodon.at is a microblogging site that federates with most instances on the Fediverse.