[Updated May 23] Conky 1.9.7 n900 edition - Page 14 - maemo.org

GomerS | # 131 | 2011-01-13, 16:29 | Report

And what information to add in a file conky.conf that in line Cell Radio that showed also CellID and LAC? At me version 1.8.5-1

The Following User Says Thank You to GomerS For This Useful Post:
AgentZ

fnordianslip | # 132 | 2011-01-14, 15:32 | Report

Creamy poked me by email yesterday and reminded me to push the Conky 1.9.6-1.2 package from extras-testing to extras, which I've now done.

I've managed to get tolua++ built now and can build Conky with full Lua support in my scratchbox, but only if I install the tolua++ deb package manually with dpkg. If I install the new tolua++ package (that I have pushed through the autobuilder) from extras-devel, then Conky will not build in my scratchbox, and it will not build on the autobuilder either, failing during the configure stage with a missing dependency on libtolua++.

I think this is possibly due to optification, I'm not sure yet but hopefully I'll have it figured out soon, and can push a version with full lua support through the repo system.

Kudos to Creamy for all his hard work.

Oh, and one final thing. I've pushed a preliminary package of conkyforecast into extras-devel if anyone wants to have a play with it. I know nothing about it yet and may not have time to try it in the near future, but if anyone feels like dabbling with it, please report back here.

Last edited by fnordianslip; 2011-01-14 at 15:36. Reason: Fixed typos.

The Following 8 Users Say Thank You to fnordianslip For This Useful Post:
Creamy Goodness, efekt, gazza_d, Helmuth, Mentalist Traceur, stevomanu, theonelaw, Xagoln

helex | # 133 | 2011-04-20, 23:26 | Report

With ${voffset} and ${offset} we are able to move the cursor relative to the current position.

With ${goto} we are able to move the cursor to a absolute horizontal position.

Is there something like ${goto} to move the coursor to a specific vertical position?

A simple home position would help, too...

The Following User Says Thank You to helex For This Useful Post:
Creamy Goodness

Creamy Goodness | # 134 | 2011-04-27, 19:16 | Report

Sorry, I was away for a few days, didn't have time to reply until now. The official Conky probably doesn't support this, just google the official docs and double check, but I don't remember anything like that.
The Conky source code is fairly confusing to me, I'm not used to pointers and custom data structres so I can't really make any quick changes to the flow of the program. I have to reinstall all the development tools first anyways, so I can't even try at the moment. Let me know if the official Conky gets updated though, I can update ours with any new features they add unless they make it incompatible somehow.

The Following 2 Users Say Thank You to Creamy Goodness For This Useful Post:
helex, Mentalist Traceur

helex | # 135 | 2011-04-27, 22:57 | Report

Originally Posted by Creamy Goodness View Post

Sorry, I was away for a few days, didn't have time to reply until now.

No problem. Thanks for your reply.
Writing later and thoughtful is even better than writing in a hurry but just in time.

Originally Posted by Creamy Goodness View Post

The official Conky probably doesn't support this, just google the official docs and double check, but I don't remember anything like that.

No, checked several times the past weeks...

I hoped for something undocumented or a dirty trick.

A Idea I haven't thought about before... a command that resets the coursor position by accident perhaps?

Originally Posted by Creamy Goodness View Post

The Conky source code is fairly confusing to me, I'm not used to pointers and custom data structres so I can't really make any quick changes to the flow of the program. I have to reinstall all the development tools first anyways, so I can't even try at the moment. Let me know if the official Conky gets updated though, I can update ours with any new features they add unless they make it incompatible somehow.

Thanks, thats good to hear!

But I doubt that there exists demand for such a feature at the desktop Conky... or it would exists since ages.

I'm currently trying to squeeze out of Conky as much information as possible. But the N900 screen is far to small. Because of that I began developing my small project Conky Layout Switcher that makes it possible to switch the Layout live within the config file and without closing conky! - but it seems to get to heavy. The CPU usage is still very high even if it doesn't draw all 4 different layouts at the same time. Perhaps you're interested how it works...

Sadly it seems that I'm unable to store information from a file in a variable, too?

I have to check a lot of times the same files if it has a specific content. I'm able to optimize this a bit (outline the blocks) as soon as the layouts are settled. But this wouldn't fix the main problem fully. Are such query somehow cached within conky or does conky execute them each time?

misiak | # 136 | 2011-05-11, 02:13 | Report

Hello! As you might have already heard, new power kernel (47) reports temperature in different units than earlier kernels ('old' way was reporting temperature in 1C and 'new' way is reporting in 0.1C). So conky shows that device's temperature is for example 310C and users get crazy that their devices are burning and melting (while the temperature is in fact 31C and it's just conky and other apps reporting it wrongly

). I created a quick temporary fix for this case untill packages in repos are updated or the situation is solved in other way. The temperature must be divided by 10 and voltage must be divided by 1000. I compiled updated version of 1.9.6-1.2 (newest avaible in extras-devel) and called it temporarily 1.9.6-1.2-misiak1 (Edit: and 1.9.6-1.2-misiak2). The changes i made (now i 1.9.6-1.2-misiak2 changes suggested by Creamy Goodness):

In file linux.c:
in function "void get_battery_stuff(char *buffer, unsigned int n, const char *bat, int item)":

Originally Posted by original linux.c

while (!feof(sysfs_bat_fp[idx])) {
//we are looping through the data here
char buf[256];
if (fgets(buf, 256, sysfs_bat_fp[idx]) == NULL)
break;
if (strncmp (buf, "POWER_SUPPLY_PRESENT=1", 22) == 0)//does current row match this?
strcpy(present, "yes");//n900 always yes
else if (strncmp(buf, "POWER_SUPPLY_VOLTAGE_NOW=", 25) == 0)//keep checking possible matches
sscanf(buf, "POWER_SUPPLY_VOLTAGE_NOW=%d", &voltage);
else if (strncmp(buf, "POWER_SUPPLY_CURRENT_NOW=", 25) == 0)
sscanf(buf, "POWER_SUPPLY_CURRENT_NOW=%d", &present_rate);
else if (strncmp(buf, "POWER_SUPPLY_CAPACITY=", 22) == 0)
sscanf(buf, "POWER_SUPPLY_CAPACITY=%d", &remaining_capacity);
else if (strncmp(buf, "POWER_SUPPLY_TEMP=", 18) == 0)
sscanf(buf, "POWER_SUPPLY_TEMP=%d", &temp);
}
fclose(sysfs_bat_fp[idx]);
sysfs_bat_fp[idx] = NULL;

last_battery_volts[idx] = voltage;
last_battery_temp[idx] = temp;

changed to

Originally Posted by modified linux.c

while (!feof(sysfs_bat_fp[idx])) {
//we are looping through the data here
char buf[256];
if (fgets(buf, 256, sysfs_bat_fp[idx]) == NULL)
break;
if (strncmp (buf, "POWER_SUPPLY_PRESENT=1", 22) == 0)//does current row match this?
strcpy(present, "yes");//n900 always yes
else if (strncmp(buf, "POWER_SUPPLY_VOLTAGE_NOW=", 25) == 0)//keep checking possible matches
sscanf(buf, "POWER_SUPPLY_VOLTAGE_NOW=%d", &voltage);
else if (strncmp(buf, "POWER_SUPPLY_CURRENT_NOW=", 25) == 0)
sscanf(buf, "POWER_SUPPLY_CURRENT_NOW=%d", &present_rate);
else if (strncmp(buf, "POWER_SUPPLY_CAPACITY=", 22) == 0)
sscanf(buf, "POWER_SUPPLY_CAPACITY=%d", &remaining_capacity);
else if (strncmp(buf, "POWER_SUPPLY_TEMP=", 18) == 0)
sscanf(buf, "POWER_SUPPLY_TEMP=%d", &temp);
}
fclose(sysfs_bat_fp[idx]);
sysfs_bat_fp[idx] = NULL;
if(voltage>10000)
last_battery_volts[idx] = voltage/1000;
else
last_battery_volts[idx] = voltage;
if(temp>100)
last_battery_temp[idx] = temp/10;
else
last_battery_temp[idx] = temp;

And I can confirm it works (tested on my N900, PR1.3, kernel-power 47). I also added lines to changelog and to description informing about changes I made and about compatibility only with kernel-power 47 and newer and broken compatibility with other kernels. If you feel like, you can upload it to extras-devel (I posted link to needed files near the end of this post). It's your decision, as power kernel 47 is not yet so popular, but it might be worth uploading to extras-devel (as version working with older kernels is in extras already). You may also disregard my changes and create new version by yourself ignoring my changelog and description additions, your decision.

PS. I know that it could be coded better and adding hardcoded /1000 and /10 is just brutal and I should feel ashamed of myself, I just didn't want to dig deeper into the code and wanted to release a quick fix until package maintainer does something more

So treat it as just a temporary fix and install on your own risk.

Sources are available at: http://www.miskiewicz.org/files/conky-n900/

EDIT: New version 1.9.6-1.2-misiak2, it should work now with older power kernels as previously

But I haven't tested it. So plase download and install conky-n900_1.9.6-1.2-misiak2_armel.deb now if you read this post

Attached Files

	conky-n900_1.9.6-1.2-misiak1_armel.deb (316.7 KB, 196 views)
	conky-n900_1.9.6-1.2-misiak2_armel.deb (316.7 KB, 262 views)

Last edited by misiak; 2011-05-14 at 17:56.

The Following 10 Users Say Thank You to misiak For This Useful Post:
anapospastos, Erazor, Estel, fw190, handaxe, Helmuth, J4ZZ, Mentalist Traceur, shawwawa, victorevo

Creamy Goodness | # 137 | 2011-05-11, 07:06 | Report

I didn't know that, I haven't tried the new kernel yet.
You should check if the voltage is > 1000 or the temp is > 100 before trying to divide it.

misiak | # 138 | 2011-05-11, 09:50 | Report

Originally Posted by Creamy Goodness View Post

I didn't know that, I haven't tried the new kernel yet.
You should check if the voltage is > 1000 or the temp is > 100 before trying to divide it.

I know, this way it would be compatible with every kernel existing nowe for N900. But I just wanted to publish a quick fix for people, as I am not a maintainer of this package. It's his decision what to do now

Of course

Originally Posted by

if(voltage>10000)
last_battery_volts[idx] = voltage/1000;
else
last_battery_volts[idx] = voltage;
if(temp>100)
last_battery_temp[idx] = temp/10;
else
last_battery_temp[idx] = temp;

Would be better. But it's not my decision :P Anyway, I will recompile package as you suggested now, it's good idea to have compability with both kernels (stock and new power kernel), I was just to sleepy to think about that at 4A.M. :P

Creamy Goodness | # 139 | 2011-05-13, 16:08 | Report

Well, I am pretty much the maintainer. I wrote that chunk of code you modified, and I was trying to get you to fix it properly and test it since you have that 47 kernel installed and I don't. This weekend I will double check it works fine on the existing kernel, and see if we can get the change pushed out to everyone.

misiak | # 140 | 2011-05-14, 18:00 | Report

Originally Posted by Creamy Goodness View Post

Well, I am pretty much the maintainer. I wrote that chunk of code you modified, and I was trying to get you to fix it properly and test it since you have that 47 kernel installed and I don't. This weekend I will double check it works fine on the existing kernel, and see if we can get the change pushed out to everyone.

Oh, sorry, I didn't notice you are maintainer at the moment, my bad :P I can confirm that:
- on power kernel 47 the package attached to my last post (1.9.6-1.2-misiak2) shows appropriate temperature and battery voltage
- on stock kernel it shows correct voltage and -1C temperature (but after I downgraded to version from the repos it still shows -1C on stock omap1 kernel, so I guess is it normal? Anyway, it behaves the same as previous package)
I haven't tested it on power kernel 46 and earlier only (well, it should be tested also using power kernel 42 as it's last version considered "stable" - it's in Extras)