VERSION 3.6 DEVELOPMENT BUILD JUNE 2021

[Mark Liversedge]

Hello

A new development build, which is going to be close to the release build in July is now available. It contains lots of bug fixes since the January build and enhancements for multisport.

https://github.com/GoldenCheetah/GoldenCheetah/releases/tag/v3.6-DEV2106

Please use this thread to notify of any build related issues.

Thanks

Mark

[Manuel Oberti]

I upgraded GC from january build and every time I upload a ride GC works well, but when I pass to activity panel GC crash.

Installed in a different folder, Vc installed , Restart but GC crasher every time

Thanks

[Guillaume TOFFOLI]

Hi Mark,

I updated GC with this last relase build and I'm having an issue you can find in the attached files.

Could you help me to fix it.

Thanks a lot.
Guillaume

[Guillaume TOFFOLI]

https://drive.google.com/file/d/1pojY-dP5yLpLJXDt8DMdGQh4ZBdAz6cO/view?usp=sharing

You can find the dmp file in the link if needed too.

Thanks for your help.

[Ale Martinez]

El miércoles, 9 de junio de 2021 a la(s) 13:08:48 UTC-3, manuel...@gmail.com escribió:

I upgraded GC from january build and every time I upload a ride GC works well, but when I pass to activity panel GC crash.

Installed in a different folder, Vc installed , Restart but GC crasher every time

Can you attach a crash*.log file or paste their content to see where the crash is happening? 

[Ale Martinez]

El miércoles, 9 de junio de 2021 a la(s) 16:00:50 UTC-3, guillaume....@gmail.com escribió:

Hi Mark,

P

I updated GC with this last relase build and I'm having an issue you can find in the attached files.

Looking at the crash log the problem seems to be happening in a R chart, I would check the R version is compatible and present in PATH environment variable running:

> R --version

from a command prompt, it should return R version 3.4 or higher.

To start GC without R embedding see https://github.com/GoldenCheetah/GoldenCheetah/wiki/Disable-R-Or-Python

[Manuel Oberti]

Sure Ale , this is the crash log

THANKS A LOT 

[Guillaume TOFFOLI]

Yes, the problem is happening with the function "ggplot" but it worked perfectly with the previous build.

I've seen that GC is compatible with R up to 4.0, is it maybe the problem because I've the lastest version ?

Thanks for your help.

[Ale Martinez]

El jueves, 10 de junio de 2021 a la(s) 03:42:01 UTC-3, guillaume....@gmail.com escribió:

Yes, the problem is happening with the function "ggplot" but it worked perfectly with the previous build.

I've seen that GC is compatible with R up to 4.0, is it maybe the problem because I've the lastest version ?

I am not aware there is a problem using 4.1, but 4.0 is the last version I tested, may be others can comment.

Know issues related to R embedding are:

1) Windows PATH pointing to the wrong location after R upgrade: https://cran.r-project.org/bin/windows/base/rw-FAQ.html#Rcmd-is-not-found-in-my-PATH_0021

2) ggplot2 doesn’t work on scripts: https://github.com/GoldenCheetah/GoldenCheetah/issues/2235

[Ale Martinez]

Regrettably the crash log is not helpful in this case.

You could try creating a new athlete from Choose an Athlete dialog and importing some files, to try to understand if this is a general problem, or one related to config of the crashing athlete in particular.

[Manuel Oberti]

Thanks Ale , I'll try.

and then I'll try another test ........

 to reset the graphics do I have to delete the home-layout file?

Thanks Ale

[Manuel Oberti]

I tried with a new athlete and now GC work fine......when I click on RIDE I see the charts and no crash

Ale, how could I fix everything without losing all my workouts?

Thanks

[Ale Martinez]

El viernes, 11 de junio de 2021 a la(s) 02:16:15 UTC-3, manuel...@gmail.com escribió:

I tried with a new athlete and now GC work fine......when I click on RIDE I see the charts and no crash

Ok, this seems to indicate install and metrics are not the source of the problem.

Now we would need to identify the chart where the crash is happening:

is it always the same chart and which one?

does it happens for any activity of specific ones?

Other useful context information are the sport of the activity(ies) and your list of possible values for the Sport tag.

 

Ale, how could I fix everything without losing all my workouts?

If you import activities from json files all data and metadata are preserved, except for zones and measures which are stored on separated files on athlete's config folder. 

[Manuel Oberti]

Thanks Ale,

GC crash with all activities when I click on tab RIDE .

I use this sports BDC,MBT,SPIN and I had set all the zone for each sports

[Ale Martinez]

El viernes, 11 de junio de 2021 a la(s) 13:47:43 UTC-3, manuel...@gmail.com escribió:

Thanks Ale,

GC crash with all activities when I click on tab RIDE .

I use this sports BDC,MBT,SPIN and I had set all the zone for each sports

Has that Ride chart any User Data series? I so, which are the formulas?

If you can't open it, please attach your config\analysis-layout.xml file for debugging.

BTW, you can create a new "Ride" chart from scratch using View > Add Chart > Performance

[Manuel Oberti]

HI Ale,

this is my file analisys-layout

thanks

[Ale Martinez]

The problem is caused by estPower User Data formula: 

((1-(7/100))^(-1))*((9.8067*sin(atan(SLOPE/ 100))*(Athlete_Weight+ 12))+(9.8067*cos(atan(SLOPE/100))*(Athlete_Weight+12)*0.005)+(0.5*0.88*(0.0276*1.75^0.725*Athlete_Weight^0.425+0.1647)*((101325*((1-((9.806*ALTITUDE)/(1007*288.15)))^((1007*0.0289)/(8.314))))/(287.058*(TEMPERATURE+273.15)))*(SPEED/3.6)^2))* (SPEED/3.6)

The expression seems correctly formed and  log shows the crash is due to exhaustion of c++ execution stack in expression evaluator:

GoldenCheetah.exe!__chkstk+0x38 vctools\crt\vcstartup\src\misc\amd64\chkstk.asm:109

GoldenCheetah.exe!Leaf::eval+0x23 Core\DataFilter.cpp:3024

GoldenCheetah.exe!Leaf::eval+0x139e0 Core\DataFilter.cpp:6413

GoldenCheetah.exe!Leaf::eval+0x1ed Core\DataFilter.cpp:3057

GoldenCheetah.exe!Leaf::eval+0x139e0 Core\DataFilter.cpp:6413

GoldenCheetah.exe!Leaf::eval+0x1ed Core\DataFilter.cpp:3057

GoldenCheetah.exe!Leaf::eval+0x13ac6 Core\DataFilter.cpp:6418

GoldenCheetah.exe!Leaf::eval+0x1ed Core\DataFilter.cpp:3057

GoldenCheetah.exe!Leaf::eval+0x139e0 Core\DataFilter.cpp:6413

GoldenCheetah.exe!Leaf::eval+0x1ed Core\DataFilter.cpp:3057

GoldenCheetah.exe!Leaf::eval+0x13ac6 Core\DataFilter.cpp:6418

GoldenCheetah.exe!Leaf::eval+0x1ed Core\DataFilter.cpp:3057

GoldenCheetah.exe!Leaf::eval+0x139e0 Core\DataFilter.cpp:6413

GoldenCheetah.exe!Leaf::eval+0x1ed Core\DataFilter.cpp:3057

GoldenCheetah.exe!Leaf::eval+0x13ac6 Core\DataFilter.cpp:6418

GoldenCheetah.exe!Leaf::eval+0x139e0 Core\DataFilter.cpp:6413

GoldenCheetah.exe!Leaf::eval+0x1ed Core\DataFilter.cpp:3057

GoldenCheetah.exe!Leaf::eval+0x13ac6 Core\DataFilter.cpp:6418

GoldenCheetah.exe!Leaf::eval+0x1ed Core\DataFilter.cpp:3057

GoldenCheetah.exe!Leaf::eval+0x13ac6 Core\DataFilter.cpp:6418

GoldenCheetah.exe!Leaf::eval+0x139e0 Core\DataFilter.cpp:6413

GoldenCheetah.exe!DataFilter::evaluate+0x3f5 Core\DataFilter.cpp:2777

GoldenCheetah.exe!UserData::setRideItem+0x2c9 Core\UserData.cpp:362

GoldenCheetah.exe!AllPlotWindow::rideSelected+0x1e1 Charts\AllPlotWindow.cpp:1837

It doesn't look like something directly related to this build, likely stack requirements are higher now and that expression was near the previous limits which are now exceeded, but I am not sure. It would be nice to have a less abrupt termination for expressions we cannot evaluate, perhaps with a recursion limit on evaluator, but for now it is what it is. 

You can avoid the crash commenting out the formula, to access the chart without a crash you can create an empty manual activity, select it and then edit the Ride chart.

[Mark Liversedge]

I wonder if making Result x in the parameters for Leaf::eval() a reference will help?

Don't think we every update it so no need to have a copy?

Mark

[Manuel Oberti]

Thank you very much Ale for your help....I'll try and I'll inform you 

Thanks

[Rui_B]

suddenly Train mode crashes the app every time (macOS11.4)

[Ale Martinez]

El sábado, 12 de junio de 2021 a la(s) 12:45:40 UTC-3, Mark Liversedge escribió:

I wonder if making Result x in the parameters for Leaf::eval() a reference will help?

Bingo, that fixes the problem!

 

Don't think we every update it so no need to have a copy?

Declaring it as const& reveals there are some cases where it is, but they are only a few and we can make a copy there.

A more subtle one is accessors (number, string) returning a reference, I opted for making a copy of the whole Result since they are a few too, but perhaps to add a const accessor would be better.

[Rui_B]

I managed to build a w/o this morning using the wizard... the w/o displayed nicely. once I change the ant+ device the app crashes every time I click on the Train tab

[Rui_B]

seems to be the Ant+ configurator... I through out the configglobal-trainmode.ini file and relaunched the app... as soon as choose to add a new device and choose Ant+, the app crashes
... none of the other choices crash the app.

[Mark Liversedge]

Hi Ale

Yes I took the exact same approach, are you going to merge or should I (I need to finish mine) ?

Mark

[Ale Martinez]

El domingo, 13 de junio de 2021 a la(s) 06:53:34 UTC-3, Mark Liversedge escribió:

Hi Ale

Yes I took the exact same approach, are you going to merge or should I (I need to finish mine) ?

No problem Mark, I will do. 

[Ale Martinez]

This doesn't look like an issue specific to this build since ANT+ support has not changed, you could confirm installing the previous one.

[Rui_B]

Thx Ale, I reinstalled the January Build and the problem went away.

[Rui_B]

I can again select different Ant+ Configurations (different SRM configs) and create new ones... with the June build... a crash happens when switching ant+ configurations or trying to add new ones.

[Rui_B]

Will test further, but I discovered the new Garmin USB Ant+ Stick is faulty... intermittently works. January build suddenly behaved like the earlier experience with June build. swapping back to the old g-stick... the problem went away.

will test further with June build later

[Rui_B]

OK I can confirm both Jan & Jun 3.6 dev versions on MacOS11 crash when selecting or trying to create an Ant+ device and there is no Ant+ USB stick (data reader) available. — my new Garmin USB Ant+ stick is intermittently failing ... with a working Ant+ data sensor, both versions run perfectly.

[Manuel Oberti]

Fantastic Ale e Mark......with your suggestion GC works correctly.

Then I installed the version ci.3875 and GC also works with my estpower formula

Thans to all for support

[Manuel Oberti]

I installed latest build but it seems that in the raw data section, in the new workouts, the L / R power balance from my workout column of my pm favero bepro has disappeared. The data efficiency, roundness are present, but L / R no

thanks

[Ale Martinez]

El viernes, 18 de junio de 2021 a la(s) 12:21:55 UTC-3, manuel...@gmail.com escribió:

I installed latest build but it seems that in the raw data section, in the new workouts, the L / R power balance from my workout column of my pm favero bepro has disappeared. The data efficiency, roundness are present, but L / R no

This is a know issue with Wahoo Elemnt Bolt  head units, see https://github.com/GoldenCheetah/GoldenCheetah/issues/3819.

If that is not your case, please attach the zipped FIT file to diagnose.

[Manuel Oberti]

Ale, This is the .fit file

Thanks

[Ale Martinez]

It is the same problem reported before: L/R Balance in that file doesn't comply the FIT standard, that's the reason it is not imported, the referenced issue for details.

You can confirm this uploading the file to Garmin Connect: no L/R data there either.

[Manuel Oberti]

Thanks Ale,

 this is an ipbike .fit file. and the developer confirms that nothing has changed in his code, but this lack of L / R Balance has been missing for 1 month. So has anything changed in the .fit standard since 1 month?
Thanks

[Ale Martinez]

El sábado, 19 de junio de 2021 a la(s) 01:22:47 UTC-3, manuel...@gmail.com escribió:

Thanks Ale,

 this is an ipbike .fit file. and the developer confirms that nothing has changed in his code, but this lack of L / R Balance has been missing for 1 month. So has anything changed in the .fit standard since 1 month?

Not that I know.

[Inti Loaiza]

Hi,

Sorry, previously i posted  this message on a wrong thread.

After this update i cannot set or see previous values of my running CP, for that reason running metrics are wrong because app is taking bike CP.

Additional,

I set to use metric Units,

but, now is showing imperial units:

Thank you.

[Ale Martinez]

El jueves, 24 de junio de 2021 a la(s) 18:47:31 UTC-3, Inti Loaiza escribió:

Hi,

Sorry, previously i posted  this message on a wrong thread.

After this update i cannot set or see previous values of my running CP, for that reason running metrics are wrong because app is taking bike CP.

You need to add Run to Sport field values in https://github.com/GoldenCheetah/GoldenCheetah/wiki/UG_Preferences_Data-Fields#fields

[Marcus Zurhorst]

Hello all,

I have a couple of issues in the training editor. 

But they are so obvious and non-subtle that I am wondering whether I might have either a broken installation or the wrong expectation how this should work.

Issue 1:

When I only click once, there is a diagonal curve drawn.  But in the settings, this is specified as a steady-state section.

When I click a second time, the settings now specify this as a transient.

There seems to be a missmatch between what is drawn and how this is described in the settings. 

When I press enter in the settings, the curves actually turn into what would expect.

Issue 2:

I realize that the selection of an intervall in the settings is incomplete.

As depicted, the last recovery cycle is not highlighted in dark-blue, and consequently, the duration is also wrongly calculated to 23 minutes instead of 24 minutes  (=8x (2+1)).

Further Issues:

I had some more trouble, but I cannot cleanly reproduce those. At some point, my intervall gets resolved into a longer list of individual sections in the settings.

Whilst still correct, this is quite cumbersome when I want to edit something on the pattern.

And finally, I also had missmatches, where the highlighted section does not match what I select in the settings. 

Can anybody reproduce those issues?

Thanks a lot.

Kind regards,

   Marcus

[Ale Martinez]

El viernes, 2 de julio de 2021 a la(s) 05:30:14 UTC-3, marcusz...@gmail.com escribió:

Hello all,

I have a couple of issues in the training editor. 

But they are so obvious and non-subtle that I am wondering whether I might have either a broken installation or the wrong expectation how this should work.

Issue 1:

When I only click once, there is a diagonal curve drawn.  But in the settings, this is specified as a steady-state section.

When I click a second time, the settings now specify this as a transient.

There seems to be a missmatch between what is drawn and how this is described in the settings. 

When I press enter in the settings, the curves actually turn into what would expect.

I did some tests and found 2 problems related to this:

a) the drawn curve should be represented as 898s@0-270 in this case, instead of 898@270

b) manually changing the qwkcode to 898s@0-270 it doesn't work since 0 is not well handled as starting watts

I will push a fix for this to avoid the inconsistency, but in practice it is not very useful to start from 0 watts and anything higher works.

OTOH, if you want to start with a block (the second picture), you can do it keeping the click a little longer before release.

 

Issue 2:

I realize that the selection of an intervall in the settings is incomplete.

As depicted, the last recovery cycle is not highlighted in dark-blue, and consequently, the duration is also wrongly calculated to 23 minutes instead of 24 minutes  (=8x (2+1)).

I cannot reproduce this

[Ale Martinez]

El sábado, 3 de julio de 2021 a la(s) 18:46:08 UTC-3, Ale Martinez escribió:

El viernes, 2 de julio de 2021 a la(s) 05:30:14 UTC-3, marcusz...@gmail.com escribió:

Hello all,

I have a couple of issues in the training editor. 

But they are so obvious and non-subtle that I am wondering whether I might have either a broken installation or the wrong expectation how this should work.

Issue 1:

When I only click once, there is a diagonal curve drawn.  But in the settings, this is specified as a steady-state section.

When I click a second time, the settings now specify this as a transient.

There seems to be a missmatch between what is drawn and how this is described in the settings. 

When I press enter in the settings, the curves actually turn into what would expect.

I did some tests and found 2 problems related to this:

a) the drawn curve should be represented as 898s@0-270 in this case, instead of 898@270

b) manually changing the qwkcode to 898s@0-270 it doesn't work since 0 is not well handled as starting watts

I will push a fix for this to avoid the inconsistency, but in practice it is not very useful to start from 0 watts and anything higher works.

OTOH, if you want to start with a block (the second picture), you can do it keeping the click a little longer before release.

https://ci.appveyor.com/project/Joern-R/goldencheetah-knhd8/builds/39858143/artifacts 

[Marcus Zurhorst]

Hello,

I may have found another issue right now.

I wanted to syncronize my new Withings scale, but I am failing at the OAuth window.

1)   I cannot enter an @ character in the email adress field

2)   The "continue with Google" link is not working.  Nothing happens.

("Continue with Apple" works fine, but that does not help me)

Kind regards,

   Marcus

[Mark Liversedge]

The dialog is delivered by Withings not GC so there isn't much we can do about that.

Weird you can't use an '@" though.

Mark

[Marcus Zurhorst]

I went to their website and had basically the same login form in my browser. Here, the Google button worked fine.

Can you reproduce this? I am on a recent Win10 btw.

At least the @ thing was mentioned last year here in the list. Missed to test this in their website.

Regards, 

  Marcus

--
_______________________________________________
Golden-Cheetah-Users mailing list
golden-che...@googlegroups.com
http://groups.google.com/group/golden-cheetah-users?hl=en
---
You received this message because you are subscribed to the Google Groups "golden-cheetah-users" group.
To unsubscribe from this group and stop receiving emails from it, send an email to golden-cheetah-u...@googlegroups.com.
To view this discussion on the web visit https://groups.google.com/d/msgid/golden-cheetah-users/752fa89d-7d37-43c5-a899-af5b1b9eddd3n%40googlegroups.com.

[Marcus Zurhorst]

Ok, I have now repeated the test with the @ char.  On their website, it is working fine.

Could this be related to the rendering component for this site?

Best Regards,

  Marcus

--

Marcus Zurhorst

Zum Baerler Busch 6

47199 Duisburg

+49 (2841) 94 91 95 9
+49 (176) 70 52 92 75

[Ale Martinez]

El martes, 6 de julio de 2021 a la(s) 03:17:30 UTC-3, marcusz...@gmail.com escribió:

Ok, I have now repeated the test with the @ char.  On their website, it is working fine.

Could this be related to the rendering component for this site?

I cannot reproduce this on Windows 10, what OS and locale are you using?

 

Best Regards,

  Marcus

Am Di., 6. Juli 2021 um 07:00 Uhr schrieb Marcus Zurhorst <marcusz...@gmail.com>:

I went to their website and had basically the same login form in my browser. Here, the Google button worked fine.

Using your system browser or GC web chart? 

[Marcus Zurhorst]

I am on a recent Win10 Pro, 20H2. Locale is german.

I was using Firefox when the google button and @ char were working, within GC3.6 it does not.

Where inside GC is specified which URL is called for the Withings connection? - Could not find this in Github.

--

_______________________________________________
Golden-Cheetah-Users mailing list
golden-che...@googlegroups.com
http://groups.google.com/group/golden-cheetah-users?hl=en
---
You received this message because you are subscribed to the Google Groups "golden-cheetah-users" group.
To unsubscribe from this group and stop receiving emails from it, send an email to golden-cheetah-u...@googlegroups.com.

To view this discussion on the web visit https://groups.google.com/d/msgid/golden-cheetah-users/82f9aeb2-d970-4bb7-bbed-33ec0dae07e8n%40googlegroups.com.

[pepe]

The  issue with @ character was  mentioned in " https://groups.google.com/g/golden-cheetah-users/c/6NIf5bCv5AY/m/dmgCveR6BwAJ "

The copy paste workaround still works.

I am using WIN 8.1 Pro Portuguese with Firefox

[Ale Martinez]

El miércoles, 7 de julio de 2021 a la(s) 02:08:40 UTC-3, marcusz...@gmail.com escribió:

I am on a recent Win10 Pro, 20H2. Locale is german.

I was using Firefox when the google button and @ char were working, within GC3.6 it does not.

Where inside GC is specified which URL is called for the Withings connection? - Could not find this in Github.

https://github.com/GoldenCheetah/GoldenCheetah/blob/55e9cdb444035647a2623a23a5562b41a40afe22/src/Cloud/OAuthDialog.cpp#L182 

[Ale Martinez]

El martes, 6 de julio de 2021 a la(s) 02:00:38 UTC-3, marcusz...@gmail.com escribió:

I went to their website and had basically the same login form in my browser. Here, the Google button worked fine.

Can you reproduce this? I am on a recent Win10 btw.

Google authentication policies has become increasingly restrictive, current wording from https://developers.google.com/identity/protocols/oauth2/policies preclude the use in our current framework:

Use secure browsers

A developer must not direct a Google OAuth 2.0 authorization request to an embedded user-agent under the developer's control. Embedded user-agents include, but are not limited to, software libraries that allow a developer to insert arbitrary scripts, alter the default routing of a request to the Google OAuth server, or access session cookies. All browsing environments must allow the user to verify the current connection to the Google OAuth server including the requested URI and connection security information.

[Ale Martinez]

It looks like a non english keyboard mapping issue. I can reproduce it switching to an ESP Latin America keyboard layout while it works with ENG US or INTL layouts.