Extracting Content Automatically

[ideasman42]

To make it easy to maintain, it should be possible to extract content automatically, formatted for inclusion.

• Commits made during the week.
• Newly added tasks and diffs.
• Tasks marked as resolved.

[ideasman42]

This script gets formats using RST for easy copy-pasting.

START="HEAD~20"
END="HEAD"

git log origin/master \
        --no-merges --first-parent \
        --stat \
        --decorate \
        --format="%n- %C(yellow):rev:\`%s <B%H>\`%C(reset)%n%n%w(120,2,2)%b" \
        $START..$END | \
        perl -e "s/\bT(\d+)\b/:task:\`\$1\`/" -p | \
        perl -e "s/\bD(\d+)\b/:diff:\`\$1\`/" -p

Example output: http://members.iinet.net.au/~ideasman42/typescript.html

---

Edit, added tool to extract this info that only uses Python & Git: b93777464a66164508511f4927adcbf4246c4ae3

[Blendify]

We can generate a list of new diffs by sorting https://developer.blender.org/api/differential.revision.search and using the title and maybe authorPHID which can be ran through https://developer.blender.org/conduit/method/user.search/ to get the real name/username a bit cumbersome but will save a lot of time if we end up using this. This can be done either through arc/php/python(https://github.com/disqus/python-phabricator) curl can also be used but there are length limits with that.