Adding an automated check to catch broken links #116
Conversation
|
@MirkoBonadei what does adding .travis.yml achieve in this change? |
|
We're not using Travis CI as you might have guessed from the error messages. Also, I'm not sure what this htmlproofer does exactly. Can you elaborate? |
|
Hi @x1ddos, my goal is to check all the links to catch broken links as soon as they break (maybe enabling a TravisCI cronjob in the future to check them once a day).
Yeah, sorry for the build, I forgot to update the |
|
Not sure why there's a need for a cron job.
Regardless, what I'm asking is how this htmlproofer works: does it start a
local server to check for broken internal links or is it specifically
desgined for jekyll; how does the tool check for external links? - a lot of
links are go outside of the site's scope. If this tool doesn't check
external links, then I question the cron job need even more, it would seem
absolutely useless in this case.
Finally, doesn't jekyll has already a builtin tool to check for internal
links maybe?
…On 28 Dec 2016 1:12 pm, "Mirko Bonadei" ***@***.***> wrote:
Hi @x1ddos <https://github.com/x1ddos>, my goal is to check all the links
to catch broken links as soon as they break (maybe enabling a TravisCI
cronjob in the future to check them once a day).
htmlproofer is a tool to validate HTML, we can run it on the _site
directory after a jekyll build and as you can see from the --only-4xx we
are only interested in stopping the build if we spot a link which returns
4xx.
Yeah, sorry for the build, I forgot to update the Gemfile.lock.
—
You are receiving this because you were mentioned.
Reply to this email directly, view it on GitHub
<#116 (comment)>,
or mute the thread
<https://github.com/notifications/unsubscribe-auth/AABjPZKSWK25OQVym6MN1r1LfYheLqb_ks5rMlIogaJpZM4LW28X>
.
|
|
Ok, I've read a bit on the html-proofer and this answers my question about external links. I also see that it's just accepts HTML as input. Will try to incorporate it in our current CI system. |
|
Yes, from the documentation it checks also the external links (which are the ones that are breaking the most, especially when they link to the codebase). The meaning of the cronjob is to catch the broken (external) link as soon as possible, since we cannot control the external source then it would be cool to fix the link as soon as it breaks. I have no clue about the internals but it seems to parse the HTML generated by Jekyll and then it performs lots of checks (we can select the checks that we want). I am not aware of something similar in jekyll. |
|
Ok, cool. We posted at the same moment. Thanks, let me know if you need something else. |
|
Right. The thing with cron job is, who do you notify in case of build failures and how often. Or will there be someone checking the build status now and then, from time to time? (note that this site is developed completely in the open) |
|
We could a new github issue created automatically for a newly broken link but I don't believe anything suitable exists. It would have to be implemented. |
|
For others webrtc projects (samples, AppRTC, ecc...) we are setting up something similar and we send notifications to |
|
Ah, that's good to know! Thanks. |
|
Yes, it would be great so we can help with the maintenance of the build. |
This is a proof of concept to understand if it is possible to catch broken links "at build time".