[mythtv-users] MythicalLibrarian, symlinks and Myth: Will this work or will Myth lose its mind?

Raymond Wagner raymond at wagnerrp.com
Wed Feb 1 21:12:44 UTC 2012


On 2/1/2012 15:32, Matt Emmott wrote:
> If you'd like to see how MythicalLibrarian does its thing, here's the 
> output from my log as it figured out a recording of In Living Color - 
> Posted to pastebin at http://pastebin.com/HqPq265Q

I know the basic operation.  I spent a decent amount of time with the 
author about a year and a half back working out some bad things he was 
doing to the database, and trying to encourage him to learn a more 
capable and more sensible language.  Of course I failed, because it was 
already written in BASH, "which everyone knows is designed for textual 
data".  All that built in capability for textual data must be why the 
script is a mess of calls out to curl, sed, grep, and awk to fake XML 
processing.

This script really kicked off my hatred of bash.  There's nothing 
inherently wrong with bash, per se.  If you're doing some limited 
variable manipulation, and just intend to string together or run 
external utilities, it's great.  That, plus the fact that it's light 
weight, and can be statically linked, is why it's used for init scripts 
so often.  On the other hand, I see far too many instances on this 
mailing list where people write absurdly long bash scripts, doing all 
sorts of database manipulation, because bash is what they know and 
they're afraid to leave their comfort zone.  That's not to say more 
advanced interpreted languages are immune to that, as there are plenty 
of other examples of things that would do well to be implemented 
directly into the code base, but are left in perl/python because 
"compiled languages are hard!"


More information about the mythtv-users mailing list