dfGUI v3.3 beta2 (Need Testers)

[Digital Parasite]

I had some time to add some of the features you guys requested and now have 3.3 beta2 out. I haven't had much time to test the new features so please download the latest version and let me know how things work. I will be going away on vacation tomorrow so won't be able to respond for a couple of weeks. After I am back I will see what I can do to add some of the other requests as well.

You can download beta 2 here:
http://gilchrist.ca/jeff/dfGUI/dfGUIv33beta.zip

Please post all bug reports in this thread so I can track and fix the problems without getting too many duplicates.


What's New in beta2/Needs Testing
- Fixed problem where dfGUI was not reporting proper # of buffered generations if DF client was stopped and # was greater than 250.

- Fixed average generation time not resetting when a new protein is detected.

- Increased Title Bar text size from 8 characters to 25.

- Added Up/Down arrows to warn and restart inactive client configuration settings. The restart value turns red if you choose less than 5 minutes since on longer proteins the DF client might not have enough time to finish the set number of proteins before progress.txt is updated and dfGUI may unnecessarily restart the DF client too often. User beware.

- Added Start, Stop, and Upload pop-up menu items to tray icon.

- Added button in Config2 tab to open DF error.log file using default Windows .log viewer.

- Added button in Config2 tab to launch DF client with -purgeuploadlist xxx command to flush buffered gens after offline "update"

- Added config option in Config2 tab to upload buffered files after each 250 gen set has been completed.


Items added in beta1
- Removed all possibilities of dfGUI getting into an inifite loop and using up 100% of the CPU.

- Changed default # of structures per generation to 100

- dfGUI will now igore the gen 0 energy score for "Best Energy So Far" rating as it uses a different system than the energy for the rest of the generations.

- dfGUI will now reset the benchmak and graph information if it detects that a protein of a different size is being used.

- Refresh button status after clicking on STOP so that Start button becomes active when DF client stops and can be used again without having to wait for next refresh period.

- Renamed "Graph Config" to "Config 2" and added a few more options to it.

- dfGUI will now report if the progress.txt file has not been updated in a specified amount of time (default: 20 minutes) and then warn the user that the client may have crashed.


Thanks,
Jeff.

[-=N0N@ME420=-]

*Downloads and tests

[IronBits]

Seems to be working just fine

[Xelas]

OK here, so far.

[HaloJones]

Looking Good. Love the Display of the error log!

[Grumpy]

So far so good, and I love the error log view also

[Xelas]

Running HT on WinXP PRO (2 services). I'm on dialup, so I run offline. DFGui has not been restarted, and the system does not shut down, sleep, or hibernate. The system has been running non-stop for over 24 hours, then I decide to upload (after connecting). I have a script that:

1) stops the services
2) copies the directories to another place
3) starts the services
4) starts the copied dirs with -u t

With the DFGui that tracks service one, everything is fine. DFGui#2, however, seems to skip or miss gens badly. The previous version, 3.2, never gave me such problems. I cleared the DFGui directories before copying over the new versions, and both DFGuis have been tracking their respective directories since the beginning of the set.

[Xelas]

A bit after the previous post, I see this: (both views are taken within seconds of each other - the two last gens are missing from the time graph)

[Ned]

Works great on all my Windows boxen. It in PRODUCTION now for me.

Keep up the great work...

Ned

[pfb]

Similar to Xelas, when I reboot the PC dfGui loses all the info upto the gen it was on - previous versions never did this

[-=N0N@ME420=-]

Remove this!! Its REALLY annoying!

*Other than that GREAT WORK!*

[Digital Parasite]

Originally posted by -=N0N@ME420=-
Remove this!! Its REALLY annoying!

The image is not showing up for me (broken link) so I'm not sure what you find annoying.

Jeff.

[Digital Parasite]

Ok I'm back from my vacation so I will start looking into the problems that you have found.

Xelas, no idea what is going on there. So it works fine with the Service version but not monitoring a non-service client? Very strange...

Jeff.

[-=N0N@ME420=-]

Originally posted by Digital Parasite
Ok I'm back from my vacation so I will start looking into the problems that you have found.

Xelas, no idea what is going on there. So it works fine with the Service version but not monitoring a non-service client? Very strange...

Jeff.

didn't know you were on vaction...

*re-uploads*

[Digital Parasite]

Originally posted by -=N0N@ME420=-
didn't know you were on vaction...
*re-uploads*

Ok, now I can see. There is no way I can "remove" this. That is the behaviour of the input box provided by Borland. If you learn how to input the numbers you should never see that error.

If the box has room for 3 digits and you just want to change one of the digits, click on the one you want to change and then type the digit you want it changed to.

If you are changing multiple numbers like you want to change the value 050 to 125, then click on the first 0 in the box and type 1 then 2 then 5. You shouldn't get any errors from that.

The code requires a valid integer in order to work properly so the value of ___ you have in the box is not and thus warning you to change the result otherwise the program would crash.

Jeff.

[pfb]

the only way to get around that is either to use a normal text edit and do some post-editing checking or use a spin edit - I've never used the mask edit for anything other than passwords and wouldn't consider it user-friendly for number entry anyway (just my opinion )

[HaloJones]

One thing about the error.log viewer: it needs to work like the UNIX "tail" command and open at the end of the log, not the beginning.

[-=N0N@ME420=-]

Originally posted by Digital Parasite
Ok, now I can see. There is no way I can "remove" this. That is the behaviour of the input box provided by Borland. If you learn how to input the numbers you should never see that error.

I Donwana learn to put numbers in the correct way, if you just make it so that no number = null therefore no error message should be displayed, or make the error msg null, do something!! If I highlight all numbers and start typing them in, bing! eror!

bah...

[Digital Parasite]

Originally posted by -=N0N@ME420=-
I Donwana learn to put numbers in the correct way, if you just make it so that no number = null therefore no error message should be displayed, or make the error msg null, do something!! If I highlight all numbers and start typing them in, bing! eror!

bah...

If you look at the reply I made above, you will see that this behaviour is built-in to the Borland component so I can't change how it works.

Jeff.

[Grumpy]

I am having some issues where I stop the Client with DFGUI and it looks as tho DFGUI freezes, but after 15 seconds it comes back to life..this is on 4 of my 7 Boxen..all Win2K

[Digital Parasite]

Originally posted by Grumpy
I am having some issues where I stop the Client with DFGUI and it looks as tho DFGUI freezes, but after 15 seconds it comes back to life..this is on 4 of my 7 Boxen..all Win2K

That is expected. When you click on "Stop" the GUI tries to stop the DF client and does some checking to make sure it actually stops. So you can't do anything else with the GUI until the DF client itself stops. The client sometimes takes 10 or 15 seconds to quit if it is in the middle of working on a structure.

Jeff.

[-=N0N@ME420=-]

Originally posted by Digital Parasite
If you look at the reply I made above, you will see that this behaviour is built-in to the Borland component so I can't change how it works.

Jeff.

Does that mean borland is a sucky language?

[bwkaz]

Borland is not a language.

Borland is a company that sells a C++ compiler with a GUI framework. (One thing they did do well, though, is use the same framework DLLs with both their C++ and Delphi (aka Object Pascal) compilers. That's not really relevant to this discussion, but it's cool nonetheless. )

The GUI framework -- specifically, the masked-edit component -- is what's giving you that message.

[Ned]

Jeff, I'd like to add to your wish list...

1/ When there is a protein change, please push out the requested Graphs with their current values... Currently they seem to be tossed, so you can't see where the last protein ended up...

2/ When there is a protein change, I currently delete the status files for dfGUI in order to see only the stats for the new protein... Could you add an option to clear the statistics so that only the new protein is reported, not stats for all proteins that you've processed since dfGUI was started? It helps me validate what the project reports that I've uploaded.

Ned

[Digital Parasite]

Ned,

What version are you using, v3.3beta2 or an older one?

#1) I will check into that but I thought I had it set up to do that already.

#2) If for some reason things the stats don't reset automatically, you can currently already do it manually. Go to the "Advanced" pane and click on "Reset" beside the "Benchmark Running TIme" to clear the benchmark stats, then go to the "Energy Graph" pane and click on "Reset" to reset those stats, then do the same thing on the "Time Graph" pane. Everything should be reset at that point. This should be done automatically but maybe my code isn't working properly.

Jeff.

[Ned]

"What version are you using, v3.3beta2 or an older one?"
-- v3.3 B2

#1) "I will check into that but I thought I had it set up to do that already."

--Sorry, trust me to not check this version... I had been starting with new folders for each protein lately and missed seeing your upgrade work. This last protein change I was away, so it happened automatically. Review showed me that it puts out the partial set PLUS and empty set graph (with almost the same time stamp).

#2) "If for some reason things the stats don't reset automatically, you can currently already do it manually. Go to the "Advanced" pane and click on "Reset" beside the "Benchmark Running TIme" to clear the benchmark stats, then go to the "Energy Graph" pane and click on "Reset" to reset those stats, then do the same thing on the "Time Graph" pane. Everything should be reset at that point. This should be done automatically but maybe my code isn't working properly."

--Benchmark, Energy, and Time seem to clear as advertised... The value that doesn't clear is displayed on the "Points" pane called "Estimated Client Total Points". That is the value I'd like an option to clear at protein change. I realize that not everyone would want that behaviour. I should have been more clear in my initial request.

Ned

[Digital Parasite]

Originally posted by Ned
"What version are you using, v3.3beta2 or an older one?"
--Sorry, trust me to not check this version... I had been starting with new folders for each protein lately and missed seeing your upgrade work. This last protein change I was away, so it happened automatically. Review showed me that it puts out the partial set PLUS and empty set graph (with almost the same time stamp).

Ok, so the latest version does save the proper graph and then reset itself properly? But also saves an "extra" blank graph too?

Originally posted by Ned
--Benchmark, Energy, and Time seem to clear as advertised... The value that doesn't clear is displayed on the "Points" pane called "Estimated Client Total Points". That is the value I'd like an option to clear at protein change. I realize that not everyone would want that behaviour. I should have been more clear in my initial request.

Hmmm that might be more difficult as it actually reads the information from the DF client to figure it out. The DF client shows how many sets of 250 generations the client has completed and from that dfGUI calculates the esimated total points. That value doesn't get reset by the DF client so I don't think there is anything I can do for that.

What you want is more like a "Points for this protein" counter right?

Jeff.

[Ned]

Jeff,

Ok, so the latest version does save the proper graph and then reset itself properly? But also saves an "extra" blank graph too?

Yes, on three machines running windows (xp, w2k, winme), I got two graphs per type requested during protein changeover.

What you want is more like a "Points for this protein" counter right?

Yes, that is exactly what I would like....

Thanks for listening...

Ned

[Digital Parasite]

Hi Ned, what about this protein changeover? I just checked my machines and when the protein changed happened, it wrote out the graphs at its current location like expected, but never wrote any blank graphs on my machine. Did you get blank graphs this time on any of your v3.3b2 GUIs?

Jeff.

[Ned]

Jeff,

Did you get blank graphs this time on any of your v3.3b2 GUIs?

Yes, the two machines (winme, xp) did switch over and produce both the requested graph and the blank graph. Since no one else has chimed in with the same complaint, perhaps we could take this offline. I'd be willing to package up my beta two folders (I have them separate from the distribfold folder) and FTP them to you if you'd find it useful.

Ned

[tpdooley]

Ned: Just because you're the only one posting a problem doesn't mean that you're the only one experiencing the problem. You may be the only one verifying things properly or paying attention or using the dfGUI beta that way. (I use it to configure the df service, or to check how many buffered gens a client has when the internet connection has been lost..)

As an example: in the beta test for the folding client, one person mentioned having to remove the -s parameter. All the other -s using beta testers followed suit; and the problem was never mentioned again until after the release when everyone that had -s parameters in the foldit.bat file suddenly had non working clients. Whoops..

[Digital Parasite]

Ok well as you all have heard by now, the DF project is closing at the end of this month so I will not be adding/fixing any more features to dfGUI.

I just want to thank all the people who did a great job at translating dfGUI into many different languages and the people who made some great suggestions. And of coure Bryan Kadzban for porting dfGUI to Linux, making it available to all our Unix friends.

It was fun while it lasted and I hope you all got some use out of the software.

Take care,
Jeff.

[tpdooley]

Thanks, Jeff, for the app that allowed us to benchmark our systems in phase I; and all the other non trivial purposes dfGUI was used for over the years.

[Nofinger]

You did a fine job in making Dfgui, Thanks for writing the program and for your support.

[-=N0N@ME420=-]

Originally posted by Nofinger
You did a fine job in making Dfgui, Thanks for writing the program and for your support.

Ditto

[HaloJones]

DFHQ also owe you a massive thank-you for making DF accessible.

[PS2pcGAMER]

Yep, dfgui has made my life easier. Thanks.

[Ned]

Jeff,

Thank you for the dfGUI tool that allowed me to track my DF progress on several flavours of Windows over the years. I didn't use all the features by any means, but as a visually oriented person, I found it great for following my progress and diagnosing problems I encountered... Again... Thanks.

Ned

[gOhAsE]

Oh no,

dfGUI was THE tool for convincing the people to calculate for df.

Thank you Jeff for programming this great tool.

This version would have been the best one by far.

10,000,000,000 is not far away anymore.


Again, a BIG THANK YOU.

[Digital Parasite]

Thanks guys for all your nice comments. Good luck with whatever project you decide to jump to next.

You never know, you must just see another GUI out there for another project some day.

Jeff.