Print

Please login or register.
1 Hour 1 Day 1 Week 1 Month Forever
Login with username, password and session length

[roseway]

Only a small number of changes, but this version deals with the remaining comments from previous versions, so I hope to draw a line under it and start thinking about future enhancements.

Changes since v4.43.1:

- fixed "invalid typecast" error when scrolling the SNRM per band graph (DS) with the mouse scrollwheel
- fixed some issues with autosaving snapshots
- added option to fix the maximum Y value of the ES per hour graph
- improved the update notification procedure to accommodate 3-part version numbers
  (won't be effective until the next version release after this of course)


http://dslstats.plainroad.me.uk

[burakkucat]

Thank you Eric. I need to find the time to "have a play". 

[NewtronStar]

Only a small number of changes, but this version deals with the remaining comments from previous versions, so I hope to draw a line under it and start thinking about future enhancements.

Changes since v4.43.1:

- fixed "invalid typecast" error when scrolling the SNRM per band graph (DS) with the mouse scrollwheel
- fixed some issues with autosaving snapshots
- added option to fix the maximum Y value of the ES per hour graph
- improved the update notification procedure to accommodate 3-part version numbers
  (won't be effective until the next version release after this of course)


http://dslstats.plainroad.me.uk

The auto update is working again cheers.

[lucas-10]

Thank you Eric, works perfectly here

[roseway]

Thanks guys.

[les-70]

 Many thanks for all the changes. 4.4 fine here as well.

    When messing about with tests I have often wondered about faster than 15 second sampling as with fast samples some impacts are easier to see.
 Not many people may be interested but!! You could consider a stripped down version able to sample every one or two seconds showing a single screen with the main stats as numbers and just a time evolution graph of attainable speeds, ideally differential values relative to those at program start or relative to user set values.   With the differential values changes would be clearer than with full scale.  I suggest attainable as a graph rather than overall snrm as attainable is best thing to follow if a re-sync occurs to different sync speed.  Just a thought if you think it might be useful.

[roseway]

The main limiting factor in the sampling rate is the time taken to login to the telnet interface. I believe that Routerstats avoids this by staying logged in for the duration of the session. I've been rather unwilling to go down this route because it has the risk of something going wrong with the connection between the program and the modem, leaving an orphan login which never terminates.

I will have a look at a cut-down version with a short sampling period, and see where it takes me.

[les-70]

 I used dmt for a long period and I think that also stays logged in. I never noticed a problem with dmt or router stats.  A simple use at your own risk option might be the option to stay logged in and use faster sampling rates. That might be worth considering in the balance of your good work v.s gain.   I guess staying logged in would mean stopping HG612 modem stats if that was running.

[roseway]

I'll give that some serious thought.

[kitz]

Put dslstats back on today.   Everything seems to look good, like the sliders.. nice job eric 

[NewtronStar]

Just want to give Eric a good slap on the back for DSLstats V4.44, the Tones Graphs are great and this right click on the tones does all the calculations to show the Frequency of RFI, and I only saw this function last night. 

[roseway]

Thank you folks.

[renluop]

AOK, but the box giving types of Thomson router does not populate, yet graphs et al working

[roseway]

There should be a file called routers.dat in the same folder as the other DSLstats files. At startup, this data is loaded into the "Modem/router model" drop-down box. If that box is empty, then I can only guess that routers.dat is missing. But in any case, the box is only a convenience for setting up the login details; once the login is set up, the box serves no useful purpose.

[les-70]

   I have been comparing the average stats on the error tab in 24hours with those for the last 24 hours and with those in the telnet data .

 Is there any relationship between the HEC and CRC counts you reported on the error tab - on fast path?  I have never been sure whether the HEC count (which is not given for each reported time interval but only overall) is the number of packets where HEC failed to correct or the number that were corrected. I often seem to have more CRC in the telnet tab than on the dslstats error tab, but the total of CRC and HEC on the error tab seems a better match to the telnet tab CRC's.   is this correct?

I suspect all this too confusing to follow?? but I thought I would ask!