Program Code not saved (Bug)

Forums ProRealTime English forum ProRealTime platform support Program Code not saved (Bug)

Viewing 5 posts - 31 through 35 (of 35 total)
  • #170716

    You are correct with your procedure (it is described extensively in this topic). But what you missed is the explicit action PRT involves “under way” and which is not applied when first the platform is in trouble (which is very easy to let happen with larger back tests).

    Mind you, my last post even describes a NEW strategy, copied from an other one. The fact that I have done it is proven by the fact that the thing runs in AutoTrading. Still the whole program is not to be found anywhere. And please, I also described I know about this procedure; again, once the platform is in trouble, nothing helps.

    I have never seen an editor programme without the SAVE button or functionality.

    Me neither, and I know a few. If someone from PRT only listens to this. How the heck can I code with confidence, knowing that each other minute I could lose my work, not even having an explicit means to secure that. @Emanuele PRT,   are you still there ?

    #171457

    Apologies Peter, I am not super techy and I cannot understand from your answer if you still have a problem or not in saving your strategies.

    What I want to add to my previous comment is: on top of doing Ctrl-S multiple times during the day, I also export the strategy in my PC by time to time (in particular if I reach a final stage). I think with these two actions I am bullet-proof against losing data. Exporting is priceless, as you can always delete a good strategy by mistake… there is when we value having back-ups.

    I hope these ideas help your trading modeling too. Gaby

     

    #171636

    I also export the strategy in my PC by time to time (in particular if I reach a final stage).

    Gaby – This is of course sufficiently bullet proof. But it is not about the possibility of losing your codes, it is about the possibility of losing your changes just made. Example:

    I am changing many things in a code and it takes me 30 minutes. Quitting the Editor to so-called save the changes will NOT be done when eventually your system or PRT runs stuck and and PRT can not normally exit. Remember, only then your code changes are saved for real (with the notice that chart changes etc. are just saved on Ctrl-s etc. – so no problem there).
    The only way with guaranteed save (as it seems) is running a backtest of the code you just changed. Thus, would I want to be secure then each 5 or 10 minutes of code changes, I must briefly start the backtest (running is 1 second is sufficient).

    Outside the backtest “solution”, the problem only occurs when during my backtesting the system is not able to normally exit any more (you can’t see that unless really 100% stuck), because at that moment only the code is saved to the cloud. This means that to be really safe you must exit PRT once per 5-10 minutes or whatever time you not want to lose. But hey, since it takes 5-10 minutes easily to re-setup your backtest environment (with the proper time periods and history etc.) plus you would lose your idea of what you were doing, this is not an option, obviously.

    Exporting your code will always work. So if your workflow would consist of doing that each 5-10 minutes (etc.) you would be good. But it’s a strange act, isn’t it ?
    All it requires – especially for the good feeling as a coder – is that save button which in addition does not save locally but to the cloud right away.

    For context, maybe it helps if I tell that for each backtesting session except for the last of the day, PRT will get stuck eventually by guarantee. I am not sure where I showed this (could be in this topic), how the memory will eventually be eaten fast by the editor. All you can do is recognize that it starts to happen (editor becomes unresponsive a first occasion) and quickly Save and Exit before all gets stuck and you lose your last changes. The message “The last action you performed required more memory than available” (similar) is the most clear indication, but it is not said that prior to that all will run stuck already. Memory problems (indeed) can not be dealt with by any (PRT) programmer. One needs to prevent them.

    Peter

     

    PS: Memory Leak or something – SLOW (Bug)

    #185097

    Hi All,

    The past week I have numerous times been explicitly quitting the platform by means of crossing away, and answer Save and Exit to the message presented – with that not bothering about changes in the Editor.
    I can inform you that this now works fine.

    So you can type something in the Editor, don’t do anything further (no brief start of a Backtest required) and just quit (and Save).


    This is not related to the platform hanging or dying during a backtest, having done Ctrl-s (Save) in advance (I just did not test that, thus no new judgment about this either).

    #185235

    The past week I have numerous times been explicitly quitting the platform by means of crossing away, and answer Save and Exit to the message presented – with that not bothering about changes in the Editor. I can inform you that this now works fine.

    I must have been dreaming/sleeping  for a week in a row because today there is no way that it works like that (again). And you can bet that I checked for a last time before I wrote this yesterday.

    What works today is crossing away the Editor (without doing anything further) and Save at Exit.

    Yesterday I marked this topic as “Resolved”, but I should not have done that; There is still no way to save the changed code in a normal fashion all coding platforms in the world are used to – in case the platform crashes. This characterises best in this way :

    • Type a lot in the Editor and all the time you feel not inclined to backtest because it can’t work (code incomplete);
    • Save as many times as you want (Ctrl-S – Main Menu);
    • Platform crash.

    Now all is lost.

    What should work is either

    • regularly quit the Editor and Save (Ctrl-S – Main Menu) or
    • regularly press the ProBacktest button, never mind errors occurring because of incomplete code and Save (Ctrl-S – Main Menu).

    The latter method takes the least of time.

    Something else : it is clear that ProRealtime development just refuses to solve this issue. Please give us a normal Save button and save that to the cloud. Not to my internal memory only !

    2 users thanked author for this post.
Viewing 5 posts - 31 through 35 (of 35 total)

Create your free account now and post your request to benefit from the help of the community
Register or Login