On 03/09/2010 15:03, Dirk-Willem van Gulik wrote:

Op 3 sep 2010, om 15:55 heeft Adam Armstrong het volgende geschreven:

On 03/09/2010 14:52, Dirk-Willem van Gulik wrote:

Op 3 sep 2010, om 15:33 heeft Adam Armstrong het volgende geschreven:

It needs a bit of cleaning up (the table and stuff is kinda ugly), but it's useful :)

Aye - agreed. I took it from the other scripts in that section - as I was unsure how much freedom I had to 'fix' that.

Dw.

We generally don't mind people rewriting things, so long as the code retains a level of simplicity that means most people can work on it (i.e. no OOP or complex structures), as most of us aren't programmers primarily :)

Ok.

There is a lot of very old and dirty code which should be rewritten,

Yes - and a lot of security/injection problems.

I've fixed about a trillion, feel free to fix any others you find. We have shortened mysqlrealescapestring to mres() for using variables in MySQL.

so it's not always a good idea to look at existing stuff for examples. Personally, I can't write user-interaction stuff to save my life, so it's usually terrible.

Ok. Meanwhile - would you mind me doing a bit of a clean (like the patch for freebsd/darwin/solaris) of a few days ago ? Or is that not something you are willing to merge in ? In which case it is easier for me to ignore patch cleanup.

Feel free to clean whatever up. It's often better to ask others what they think of changes, as there are quite a few others with ideas on how we should change things. Most discussion is on the IRC channel on Freenode.

Some of our primary goals are ease and intuitiveness of UI (we quite often won't implement things until we know exactly how it should look in the web interface), and enforcement of good practice in SPs (we dislike implementing things to work around the lazy or bad practices of others, as it only helps to make our own staff lazy!)

My private agenda is a bit if 'ignore' flags (disks, interfaces) where needed, disk thresholds, SNMPv3, netgear switches (see patch), bit more power discovery on APC and perhaps an 'aggregate/stack bandwidht' display for switches as an option in the all-devices-graph-popup - as for a lot of swtiches uptime/cpu displays make little sense.

We certainly need more manual ignores. I also think we perhaps shouldn't delete things so readily, rather mark them removed but retain data and settings, incase they spontaneously reappear.

adam.