Kitz ADSL Broadband Information
adsl spacer  
Support this site
Home Broadband ISPs Tech Routers Wiki Forum
 
     
   Compare ISP   Rate your ISP
   Glossary   Glossary
 
Please login or register.

Login with username, password and session length
Advanced search  

News:

Pages: [1] 2

Author Topic: rs-ux and rs-w v2.1 released  (Read 5883 times)

roseway

  • Administrator
  • Senior Kitizen
  • *
  • Posts: 39632
  • Penguins CAN fly
    • DSLstats
rs-ux and rs-w v2.1 released
« on: February 04, 2013, 02:44:28 PM »

Bugfixes, tidying-up of some details, and some small new features

- when a re-sync occurs, or no data is received from the router, zero values are plotted and sampling continues instead of pausing as it did before
- fixed bug whereby a change in graph colours wasn't propagated to all the graphs
- fixed bug whereby a right-click below the bitloading graph didn't clear the tone values
- if recording is paused for a long period, the program no longer hangs when attempting to resume recording
- an issue with button positioning and graph resizing has been resolved
- an option has been added to specify the connection type as Annex M so that tweaking or re-syncing works correctly with this type of connection
- the program now takes account of different decimal separators used in different locales
- confusing items removed from the colour selectors used for graph configuration
- the zero-byte 'test' file used to verify the snapshot directory is now deleted after use

http://rsux.plainroad.me.uk
« Last Edit: February 04, 2013, 02:46:35 PM by roseway »
Logged
  Eric

les-70

  • Kitizen
  • ****
  • Posts: 1254
Re: rs-ux and rs-w v2.1 released
« Reply #1 on: February 04, 2013, 04:33:04 PM »

  Annex m option works!  :) Thanks
Logged

roseway

  • Administrator
  • Senior Kitizen
  • *
  • Posts: 39632
  • Penguins CAN fly
    • DSLstats
Re: rs-ux and rs-w v2.1 released
« Reply #2 on: February 04, 2013, 04:37:05 PM »

Thanks for letting me know. :)
Logged
  Eric

broadstairs

  • Kitizen
  • ****
  • Posts: 3300
Re: rs-ux and rs-w v2.1 released
« Reply #3 on: February 04, 2013, 06:02:06 PM »

Been running fine here since mid-afternoon.

Stuart
Logged
ISP:TalkTalk Connection:FTTC Cab:ECI Router:Netgear D6220

roseway

  • Administrator
  • Senior Kitizen
  • *
  • Posts: 39632
  • Penguins CAN fly
    • DSLstats
Re: rs-ux and rs-w v2.1 released
« Reply #4 on: February 04, 2013, 06:39:21 PM »

Thanks Stuart. I'll be interested to hear if you get the occasional "No stats received" error, and rs-us handles it as intended.
Logged
  Eric

burakkucat

  • Global Moderator
  • Senior Kitizen
  • *
  • Posts: 29845
  • Over the Rainbow Bridge
    • The ELRepo Project
Re: rs-ux and rs-w v2.1 released
« Reply #5 on: February 04, 2013, 09:55:56 PM »

Thank you for this ongoing development, Eric:)

Only just downloaded the latest release to The Cattery. It will be tested 'when the moment is ripe'.
Logged
:cat:  100% Linux and, previously, Unix. Co-founder of the ELRepo Project.

Please consider making a donation to support the running of this site.

burakkucat

  • Global Moderator
  • Senior Kitizen
  • *
  • Posts: 29845
  • Over the Rainbow Bridge
    • The ELRepo Project
Re: rs-ux and rs-w v2.1 released
« Reply #6 on: February 06, 2013, 12:16:35 AM »

The moment 'ripened' and I have been able to test this latest release, version 2.1, on a 64-bit Linux kernel based system.

Everything now looks good. There is just one small niggle. The expansion/contraction feature of the graphs. For the Bitloading graph, ticking the 'Expanded view' box (and its converse) works exactly as expected. However for the Hlog, QLN and SNR per tone graphs that feature does not work. Those three graphs remain in their 'Contracted view'.  :(

As an example of both cases, two Bitloading and two Hlog graphs are attached. Please note the state of the 'Expanded view' tick-boxes for each graph.
Logged
:cat:  100% Linux and, previously, Unix. Co-founder of the ELRepo Project.

Please consider making a donation to support the running of this site.

roseway

  • Administrator
  • Senior Kitizen
  • *
  • Posts: 39632
  • Penguins CAN fly
    • DSLstats
Re: rs-ux and rs-w v2.1 released
« Reply #7 on: February 06, 2013, 07:28:44 AM »

Thanks for the report. The expanded/contracted behaviour of the three graphs you mention is deliberate, although I accept that it may be rather confusing. In the default expanded view the horizontal scale is one pixel per tone, so every tone is shown, with no gaps between. When "Expanded view" is unchecked, the horizontal scale is compressed as necessary in order to make it possible to see the entire tone range without scrolling. In the case of ADSL (2, 2+) connections, only 512 horizontal pixels are needed, so no compression takes place.

The bitloading graph is different - in the expanded view there are three pixels per tone, so each tone can be individually inspected, so even on ADSL connections the compressed view is narrower.

I'm not really sure what's best to do here - to leave it as it is, but add some explanation to avoid the confusion, or to redimension all the graphs in question to behave like the bitloading graph. Does anyone have a view on this?
Logged
  Eric

les-70

  • Kitizen
  • ****
  • Posts: 1254
Re: rs-ux and rs-w v2.1 released
« Reply #8 on: February 06, 2013, 09:35:40 AM »

 At risk of being very fussy I wonder if the behavior after a PC standby period could be better.  I use standby quite a lot simply to save power. 

At the moment on speed and SNR sometimes I get a straight line and sometimes a dip to zero, the inconsistent behavior is confusing.  On on errors I get high peak equal I think to error sum since it last connected.   

The plots below have three standby period one long and two short.  The error peak is after the long one and dip in snr and speed after one of the short ones.

Personally I would prefer nothing to be plotted when the time interval between samples is more than it should be.  i.e just gaps in snr and speed and no large scale distorting value with errors.

A dsl disconnection gives a nice drop to zero in snr and speed and to my surprise preserves the snr tweak.   I like it preserving the snr tweak  :) . The disconnection does however render the error plot useless by adding such a large peak that rest vanishes at zero. Unless you avoid plotting the initial error after a resync that may be hard to avoid.

 
Logged

broadstairs

  • Kitizen
  • ****
  • Posts: 3300
Re: rs-ux and rs-w v2.1 released
« Reply #9 on: February 06, 2013, 10:12:06 AM »

Just a thought. When your PC returns from standby I wonder if the differences you see could be caused by rs-ux attempting to contact the router before the network has reconnected and sometimes it may not try until after the network is up. I'm not convinced Eric will be able to do much about this if that is the case, unless it is possible to detect the network is reconnected.

Stuart
Logged
ISP:TalkTalk Connection:FTTC Cab:ECI Router:Netgear D6220

roseway

  • Administrator
  • Senior Kitizen
  • *
  • Posts: 39632
  • Penguins CAN fly
    • DSLstats
Re: rs-ux and rs-w v2.1 released
« Reply #10 on: February 06, 2013, 10:21:34 AM »

To be honest, I haven't even considered standby, and I have no real idea what happens to running applications during standby. Assuming that they go completely quiescent, then your suggestion to plot nothing when the time between samples is more than it should be sounds like a good one. But as Stuart says, there may be a situation while the PC is emerging from standby, which I won't be able to deal with.

I think I can change what happens during a disconnection, by making the behaviour the same as when the router fails to return a sample. What happens in this case is that zero values are plotted, and the error graphs start counting from the next real sample, so there's no spike plotted. I think I need to do the same thing during a pause.
Logged
  Eric

broadstairs

  • Kitizen
  • ****
  • Posts: 3300
Re: rs-ux and rs-w v2.1 released
« Reply #11 on: February 06, 2013, 10:30:07 AM »

I realise that the more your excellent program is used the more suggestions you get but here goers with another  ;)

Relating to plotting zeros or not during pause etc there could be arguments for doing it both ways, either zeros or a straight line between plot times. Might I suggest one way to decide which to do? I would suggest that somewhere you record the actual DSL up time at each successful sample and then after a pause or other time gap compare the new DSL up time against the last recorded value and if the new value is less then plot zeros on all timeline graphs as there was obviously a re-sync or reboot during the gap, if the new time is greater then simply plot a straight line between the last and new time steps. At least this would then mean it was obvious from looking at the graph that a re-sync or reboot (of router) happened.

Stuart
Logged
ISP:TalkTalk Connection:FTTC Cab:ECI Router:Netgear D6220

roseway

  • Administrator
  • Senior Kitizen
  • *
  • Posts: 39632
  • Penguins CAN fly
    • DSLstats
Re: rs-ux and rs-w v2.1 released
« Reply #12 on: February 06, 2013, 11:18:16 AM »

These are good suggestions. I can't use the DSL uptime as the determinant though, because not all Broadcom-based routers return the actual DSL uptime, and as we've discussed before, the AS figure seems to vary a bit in what it actually represents. But what I can do is look for "Showtime" in the stats and use that as the determinant.

I think there's some merit in the idea of plotting zero values (or possibly spaces) during downtime, and straight lines between samples during pauses or missing samples, but taking into account the issue with errors as mentioned by les-70. More head scratching to be undertaken. :)
Logged
  Eric

burakkucat

  • Global Moderator
  • Senior Kitizen
  • *
  • Posts: 29845
  • Over the Rainbow Bridge
    • The ELRepo Project
Re: rs-ux and rs-w v2.1 released
« Reply #13 on: February 06, 2013, 08:25:23 PM »

Thanks for the report. The expanded/contracted behaviour of the three graphs you mention is deliberate, although I accept that it may be rather confusing. In the default expanded view the horizontal scale is one pixel per tone, so every tone is shown, with no gaps between. When "Expanded view" is unchecked, the horizontal scale is compressed as necessary in order to make it possible to see the entire tone range without scrolling. In the case of ADSL (2, 2+) connections, only 512 horizontal pixels are needed, so no compression takes place.

The bitloading graph is different - in the expanded view there are three pixels per tone, so each tone can be individually inspected, so even on ADSL connections the compressed view is narrower.

Thank you for the explanation, Eric. I now fully understand. (Until next time . . .  :-\  )

Quote
I'm not really sure what's best to do here - to leave it as it is, but add some explanation to avoid the confusion, or to redimension all the graphs in question to behave like the bitloading graph. Does anyone have a view on this?

I do not have a strong opinion on this. Perhaps you could show us a few examples of what, say, an ADSL2+ Hlog graph would look like when it is expanded as like the bitloading graph?
Logged
:cat:  100% Linux and, previously, Unix. Co-founder of the ELRepo Project.

Please consider making a donation to support the running of this site.

roseway

  • Administrator
  • Senior Kitizen
  • *
  • Posts: 39632
  • Penguins CAN fly
    • DSLstats
Re: rs-ux and rs-w v2.1 released
« Reply #14 on: February 06, 2013, 10:42:15 PM »

I do not have a strong opinion on this. Perhaps you could show us a few examples of what, say, an ADSL2+ Hlog graph would look like when it is expanded as like the bitloading graph?

I'll put that on the list. :)
Logged
  Eric
Pages: [1] 2
 

anything