Notices

What do you guys think? I have lots of extra space to fill after shrinking those bars and overlaying text on top of them. You can try my version from here if you want: xxxxx Not 100% sure if it installs cleanly, so it's at your own risk
(i will fix it for non power-kernel users, i wouldn't try this if you don't have that)

-edit- removed outdated link, get new one from project page

Couldn't you just share the conky.conf file? Or is this a later version of Conky? In which case are you the current Conky maintainer, or has someone else originally ported it, and if so, are they maintaining it at all?

Meanwhile: For the people following my problems above: I figured out why my battery values were screwed. Simply put bq27x00_battery and BME do not (at least with Advanced Power Monitor, but I don't think it's responsible for it) work well together. For the time being I've got bq27x00_battery being loaded at every boot, and am just rmmod-ing it for testing purposes and when I get annoyed by it.

Meanwhile, I have ask - do you know the command to fetch accelerometer stats in Conky? I have noticed that conky can easily be rotated - and watching how nicely some of the graphs resized themselves made me think about making conky display different bits of information below the end of the screen, which could then be seen when you rotate conky, while reserving the right 320-ish pixels of space for data you're willing to not see while looking at the data that's revealed when you rotated it.

This was too aesthetically unpleasing of a model, however, this made me realize that one could make an automatically portrait-mode supporting conky by merely enclosing the entire configuration text in an if-then statement, which polls the accelerometer, or whatever higher-level device-orientation state value can be grabbed, and if it is portrait, displays (portrait-mode layout), else (ie if it is landscape) displays (whatever layout you're currently using). or vice versa, since it probably doesn't make a difference which value is used as the "if" condition, since the other one will be the else condition.

Also, what's up with us getting shorted on all these neat features being unsupported by the kernel. Titan's still active isn't he? Shall we prod him about it? Especially since he'll probably have to release another Power Kernel for PR1.3, which is apparently confirmedly in the will-show-up-eventually-but-it-WILL-show-up category now?

yeah this is the newer version of conky i'm working on, nobody was maintaining it since the initial release so I asked to get the source and admin access to the monky project.
the conf file definately won't work with older versions in a useful way because i am making heavy use of "voffset" to have both left and right aligned text on the same line. I couldn't get it to work anyways, but I forget if it's even a feature of the old version or not.
I'm not sure I understand what you're saying about the battery, I'll have to re-read this stuff later when I have time.
I like your idea of having dual configurations for portrait / landscape mode, I'm not sure how to implement that but I'll definitely try to figure something out. I would think conky only reads the .conf file on load, so polling the system for the current orientation would have to be handled in the code somehow.
You should ask Titan if he disabled that on purpose to save... something. It would make sense to disable it in the kernel if tracking io stats is likely to consume too much cpu.

The battery thing was me explaining a problem I was having a while ago. When you have the bq27x00_battery module (comes in power kernel, it's what lets you see temperature and such) loaded, it seems to mess around with BME, so all the commands that fetch battery info (including the ones in healthcheck, netmon, and the Advanced Power Monitor replacement for the normal battery status menu applet) start getting glitchy, either underreporting the percent battery left, or completely failing to work (as in the case of Netmon). It doesn't have to do with Conky directly, which is what I was saying in case anyone was still paying attention to my earlier posts about such bugs.

As for the support for portrait by default - you don't really have to do any work, other than see if you can find a way for conky to run with rotation enabled as on or off based on a single line in the conf file (if possible, and if not, oh well, rotation can still be enabled manually for people who want it.) - I was just saying that if we have a conky command that will fetch device orientation (like healthcheck does), we can write our own layouts in the .conf file by enclosing everything in an if-then statement.

The implementation would still be for the user, as is writing your own .conf files if you don't like the default is already. My main question is: Do you know the code to fetch accelerometer data, and or to fetch the system's "device orientation" state on the N900 (The latter would make life easier, but the former works too)?

I'm not using the .conf that came with it, but I'm very happy to have a newer version of conky. Thanks.

I f'd up that release and the .conf is old, not matching my screenshot.
I'm making good progress getting it to work the same with the standard kernel.

-edit- if you use this with my latest build, remove the ${battery_percent}

Here are my conf files for conky.
One is for CPU and memory utilization, another one for system resources and kernel configuration, and third one for monitoring network traffic.
Thanks to all the posters on this thread who shared their configurations and gave me nice ideas

I also created a QBW widget to launch them simultaneously.Have a look at it here.

Kindly test and report back.

i hate that network conf you copied from, it just assumes that the connection is incoming or outgoing based on the local port number.
1-32767 show as incoming
32768-61000 show as outgoing
and anything 61000-65355 is not listed. GENIUS

@Creamy Goodness
Sorry don't have a thorough understanding of ports.Care to provide a fix please.
This was also used in a conf posted earlier, had I seen any report on this earlier would have fixed it.Can you help now?
I did notice that entire port range was not being covered but since no one brought this up earlier I thought I may have misunderstood something.Also I found the same thing on ubuntu forum leading me to believe there is some reason for it.

yeah, it is okay man, i know you are just copying it.
the thing is, it's fairly accurate as is, at least the connection count is 1-65535. but it might say there are 3 connections and only list two.
you basically have two choices:
1) be like me, change it to have just one large area (2 columns or however you lay it out) of 1-65535. I like this because it won't waste space, meaning, for example, it won't have half of your screen area devoted to incoming connections when you have more outgoing that incoming connections, and then it has not got enough room to display them all on the one side. Hopefully that made sense.
2) just fix the 32768-61000 to be 65535. I don't like this layout as I mentioned for the two reasons above - it's just guessing, even if it's mostly correct. And it potentially can't show all the connections even when it should.

wikipedia says the dynamically created outgoing connections are are called ephemeral ports and are 32768-61000 on "many linux systems"
i think its 49152-65535 on maemo, it says the file /proc/sys/net/ipv4/ip_local_port_range contains the range in use, and that is what was in there.

anyways i still don't like this idea of guessing the port direction just by the number, but if you want to do it that way at least use the right ranges!
1-49151 and 49152-65535