0 users browsing Programming. | 6 bots
Pages: 1
| Screwtape |
Posted on 18-11-09, 03:57
|
|
Full mod
 Post: #25 of 443 Since: 10-30-18 Last post: 1101 days Last view: 172 days |
The redo build system kind of blew my mind when I came across it years ago; simple to understand and implement but surprisingly powerful and flexible, it's exactly the kind of software I want to write and to use. Since it was designed by DJB, the same guy who invented the daemontools service monitor and the UCSPI-TCP network server framework, it's inspiringly elegant and more thoroughly Unixy than the actual official Unix tools it replaces. I just came across redo, buildroot, and serializing parallel logs which talks about a new redo feature that records and caches build-logs so that you always get live-updating progress while the project is building, but build logs are always in a deterministic, sensible order no matter how much parallelism is involved. Once again, the trick's pretty simple to understand (and implement) once you've had it explained to you, but it sounded pretty magical to me at first. I really wish I had some kind of project where I could use redo to its full potential, but pretty much everything I'm working on these days is either Rust (which already has its own smart build-system) or Python (which resists having a build-system at all). The biggest niche for redo right now is probably for meta-build-systems that build a bunch of components written using other build-systems, like the "buildroot" project he mentions that builds a bootable Linux disk image from scratch... but I really have no reason to work on anything like that. :/ The ending of the words is ALMSIVI. |
Pages: 1
