ATM we copy release logs to a dedicated S3 bucket, the logs are publicly accessible, but they are not browsable. Maybe we can use clodups infra and "mount" this bucket to some directory under archive.m.o.
What about generating an index.html for each directory? Every time a log is uploaded we could check for an index.html and create it if it doesn't exist. If it does exist, we overwrite the file with the new log to reflect the change.
I wonder how people do this... Generating index.html may lead to race conditions, so you may end up with missing files. I quick google search points me to https://github.com/rgrp/s3-bucket-listing, looks promising!
I put the index.html and list.js to the bucket and followed the instruction and it looks like it's browsable now: http://mozilla-release-logs.s3-website-us-east-1.amazonaws.com/
Nice! It looks like it's only browsable when you start from the index, though. For example if you were looking at a log and took the 'filename' out, you would not be able to see all related logs in a listing (ie this doesn't do what you would expect http://mozilla-release-logs.s3.amazonaws.com/mozilla-beta/firefox-49.0b8/build1/) If you think it's worth doing, I could change the log upload procedure to include adding an index.html file with the script to any new 'directories' so you can see the related files without having to click all the way up the tree.
Apparently the script handles that too! http://mozilla-release-logs.s3-website-us-east-1.amazonaws.com/jamun/firefox-45.2.5esr/build3/ wfm now. I needed to change S3BL_IGNORE_PATH to false to make it work
Status: NEW → RESOLVED
Closed: 3 years ago
Resolution: --- → FIXED
You need to log in before you can comment on or make changes to this bug.