Uh oh!
There was an error while loading. Please reload this page.
Ignore the existing broken URLs - #200
Conversation
Data cleanse - correct owner names and strip errant spaces (pidcodes#176)
Resync with master
Escape the colons in the title (pidcodes#180)
Resync with master
peternewman
commented
Dec 3, 2016
I should probably double check if any have just moved before we give up on them entirely. |
Arachnid
commented
Dec 3, 2016
Linkrot sets in fast! I predict a lot of updates to this file. :/ |
peternewman
commented
Dec 3, 2016
Heh, well that's the other question, do we make it so this test has to pass for a merge or not? I'm inclined to say yes for a bit and see how it goes. The fact we've only had 8 errors across all the PIDs so far bodes rather well I'd say. |
peternewman
commented
Dec 3, 2016
@Misfittech you seem to have deleted your Gerbers by accident. Also would you mind checking the links I've updated to are the correct ones please? |
Arachnid
commented
Dec 17, 2016
Is this ready for review & merge? |
peternewman
commented
Dec 18, 2016
Not yet, I need to manually set all the other links to be proper links, then fix any new broken addresses. I think I'm going to add some new optional fields to PID and org, so expect those soon first. |
peternewman
commented
Jan 4, 2017
Okay @Arachnid this is now green. For now this just fixes or blacklists any broken URLs, but you'll still need to manually look at Travis allowed failures to see if a new PR introduces broken links. Perhaps run it like that for a bit and see how many false positives it gives/how many existing sites disappear, then potentially look at integrating it as a required test? |
Arachnid
commented
Jan 9, 2017
I assume I need to look at allowed failures and see if the count goes up? Eg, there's no easy way to see if the failures are new to the current PR? |
peternewman
commented
Jan 10, 2017
The result should be green currently, as the count is zero. If an existing link breaks, and then a new PR is made, it will appear that the PR broke it, there's no trivial way to differentiate that. Although doing things like https://docs.travis-ci.com/user/cron-jobs/ and blacklisting broken links as soon as they appear could help. I guess it should be possible to write something that runs during the PR, generates a list of all links, or all broken links from master at that point, and then uses that list to blacklist all existing links, then use that blacklist to run against the PR (so it only checks the PR links). But that's probably rather complex for the sake of a few web URLs... As I say, the balance is between removing it from allowed failures, so you don't have to check, and the Travis merge will complain if there are broken links, balanced against dealing with random links dropping off as false positives. |
Arachnid
commented
Jan 10, 2017
Right, gotcha. Thanks again for all the travis help! |
No description provided.