bashblog shouldn't rely on file date for post ordering

[cig0]

May be it would be best to store posts' index + date on a separate index tree.

[cig0]

A workaround could be:

1. Store .md and .html files time stamp
2. Do the rebuild magic
3. Touch -d $original_timestamp -m _file_

[cig0]

Just changed the name of the ticket as this has nothing to be with the 'rebuild' command.

These are however valid scenarios when the time stamps of the posts are updated - thus messing with the blog contents:

1. When reverting a commit (git revert, git reset, etc.)
2. When cloning the _user_.github.io repo

[cfenollosa]

Ordering files by timestamp is super convenient and fast and I'd like to keep it that was as much as possible. However, I understand that there are some scenarios which change timestamps and thus break ordering.

The post date is already stored inside the <div class="subtitle"> for each file, though it doesn't contain the time by default, just the day, month and year.

A middle ground solution could be to add a post timestamp to each file, inside a comment, and add a new bb.sh regenerate_timestamps to fix these after they have been modified externally.

What do you think of this option? Let's try to find a good solution that can work for the future before starting coding :)

[cig0]

Hi Carlos,
+1
That would be really useful, I'm very happy with the idea!

[cfenollosa]

Some more on this issue here: https://github.com/cfenollosa/bashblog/wiki/Regarding-file-timestamps-and-blogpost-dates

[cig0]

Grrrreat, now I can clone my blog on the different computers I work daily and blog from them all - thanks!