2014-05-28 Pipeline Factory » History » Version 4
Tom Clegg, 05/28/2014 02:59 PM
1 | 1 | h1. 2014-05-28 Pipeline Factory |
|
---|---|---|---|
2 | |||
3 | 2 | Tom Clegg | Retrospective |
4 | * Velocity a bit more steady |
||
5 | * Early velocity probably more indicative of rejected stories than real points |
||
6 | * Updating time-remaining could still use more diligence |
||
7 | * Unforeseen dependencies (e.g., new keep server + new sdk) |
||
8 | * Auto-deploy latest Python SDK might be wrong? |
||
9 | * Could avoid some problems if test suite tested old components against new components |
||
10 | * Write tests against the APIs you're *using*, not just the ones you're *building* |
||
11 | * Include story points and acceptance tests for docs&tests for new bits (and filling in deficit) |
||
12 | 3 | Tom Clegg | * Need to separate "staging" from "master" instance |
13 | * Need live integration test |
||
14 | * Stable/master doc sites? (no, later) |
||
15 | * Nicer name for workbench, without qr1hi etc. |
||
16 | 2 | Tom Clegg | * git merge party hairball |
17 | 4 | Tom Clegg | * commit-message rules |
18 | ** local commit hook should apply the rules (copy story# from branch name to commit, say refs #1234 on merge) |
||
19 | ** only follow left parent when enforcing commit message rules ("1 parent, or has refs #1234" + "never says 'merge master'") |