This repo contains the source files for the bufferbloat.net web site (https://www.bufferbloat.net). The site is built using the Hugo static site generator (https://gohugo.io) and most of the content is automatically converted from the old bufferbloat.net redmine installation.
Contributions to improve the content are very welcome. See below for some pointers on how the content is organised and how to contribute.
The content shown on the site is all in the content directory. Each
subdirectory is a project, which in turn has subdirectories
corresponding to the different types of content. Currently, these are
wiki for Wiki pages, news for news items and issues for the
exported bug reports from the old site. The latter is mainly kept for
archival purposes, and is probably not worth editing.
The scripts used to convert the content from the old site live in the
export directory, which also contains the csv files with the data.
Hopefully, there shouldn't be a need to run these anymore. The
old-projects directory contains exported data for projects that it was
not deemed relevant to move over to the new site. This data is here for
archival purposes and is not used when generating the site.
Finally, the static directory contains resources such as images and
CSS files used for the site layout, and the layouts directory contains
the templates Hugo uses to build the site.
For more information on how the latter works,
consult the documentation at https://gohugo.io.
To contribute to the site, simply clone the repository, make your changes to the markdown file corresponding to the page you wish to edit, commit the results and open a pull request against this repository. It is also possible to file issues to point out errors or omissions, or give suggestions for the content.
The site is build using the Hugo static site generator. Version 0.16 (not higher!) is required to build the site. See the Using Docker for development section below.
When making changes, you can see a local version of the site by
installing Hugo and running hugo serve in this directory. That will
run a web server on localhost, which you can navigate to and see a live
version of the site.
You can use the hugo binary to create a new news item with appropriate front matter (the stuff in the top that tells Hugo that this is a news item). First, decide which project the news item should belong to, then issue a command like
hugo new -k news bloat/news/2016-06-12-my-news-heading.md
to create the new file (the -k news tells Hugo to create a news item).
The new item will be marked as a draft and appear in the
content/bloat/news directory. You can then edit the file (don't forget
to the title and add your name as author). While you are editing, you
can run hugo -D serve to see the changes in real-time (the -D means
'show drafts'). When you are done you can either manually change the
draft status at the top of the file, or you can use the hugo undraft
command to do it for you. Then add the file to git and submit a pull
request to have your news item included on the site.
The Bufferbloat.net site relies on Hugo 0.16 capabilities. (Newer Hugo versions give WARNING: blackfriday's sourceRelativeLinksEval is deprecated ... error messages for the hundreds of internal links in the repo.)
Docker provides an easy way to run Hugo 0.16 reliably on any OS. The codycraven/docker-hugo repo has a container for creating and running a Hugo server in Docker.
The Docker instance operates on the current directory,
so normal editing, git commands, etc. work as desired.
This command runs the Hugo server, listening on port 8080.
To use it, cd <directory-containing-your-Hugo-files>
then issue the docker run... command each time you edit the site.
cd <directory-containing-your-Hugo-files>
docker run --rm -it \
-v /tmp:/tmp \
-v $(pwd):/site \
-w /site \
-p 8080:8080 \
codycraven/hugo:0.16 \
server \
--bind=0.0.0.0 \
--port=8080 \
-w
Follow the link to localhost:8080/projects, and you'll see the blog as it appears. As you edit the underlying pages, the web display will update live.