Hi, I could reproduce the problem starting a new site from scratch. 1) `nikola init site`, accepting all the defaults, no comments system. 2) Change the configuration as follows: a) use "" as the destination folder for stories b) change `INDEX_PATH` to "posts" 3) Add a couple off dummy posts 4) Add a new "home page" for the site a) Set its slug to `index`, so it works as a proper home page b) Change its date to "1970-01-01 00:00:00" to avoid the race condition that was mentioned before (maybe this is unrelated... but still). c) Add a `post-list` directive to include the last 5 posts 5) `nikola build` The whole site is built as expected. 6) Add a new post 7) `nikola build` The file output/index.html (i.e., the home.rst rendering) is NOT rebuilt, and as such, the post list in it is not updated; I'm attaching: - conf.py - test.log: the stdout of the whole process (stripping some of the first build), with a couple of comments highlighting the points where I made changes - packages.txt: my list of packages (`pip list` from the virtualenv I'm using). This is on Ubuntu 4.10, with Python 3.4, with `nikola[extras]` installed in a virtualenv. If you confirm this is a bug and you want me to open an issue in GitHub, please let me know. Also, I *might* get at shot at it if someone give me some pointers. My python-fu is way too rusty nowadays :'(, but who knows... Thanks! On Tue, Apr 28, 2015 at 2:33 PM Mariano Draghi wrote: > On Tue, Apr 28, 2015 at 2:04 PM Roberto Alsina > wrote: > >> On 28/04/15 13:32, Mariano Draghi wrote: >> > In fact, that problematic page (`pages/home.rst` in my case) was not being >> rebuilt, and because of that when I added the last post >> (`posts/rearrancando.rst`) it didn't get linked from the home until y >> manually made some dummy edit int `pages/home.rst` to force a rebuild of it. >> >> >> That looks like a huge bug. If someone can provide me steps to reproduce >> this sort of things from scratch I will take a look in my next free time :-( >> >> > > I'll try to give it a shot later using a new site, and see if it's > reproducible in a predictable way. > >