[GNC] Auto commit with auto save?

[John Layman]

One of the remarkably slick features of Microsoft Money was that it kept track of the last date on which it downloaded a day's adjusted close, and worked forward from that to the present whenever there was a gap.

-----Original Message-----
From: gnucash-user <gnucash-user-bounces+john.layman=ieee.org at gnucash.org> On Behalf Of David T. via gnucash-user
Sent: Monday, May 8, 2023 3:00 PM
To: David G. Pickett <dgpickett at aol.com>
Cc: gnucash-user at gnucash.org
Subject: Re: [GNC] Auto commit with auto save?

Yes, one can imagine that. I can imagine that it would be possible to add something like that into the get quotes function that already exists, but I'm not a programmer, so I accept that if I want to track commodity values in GnuCash, I have to click the get quotes button. The get quotes button is as close to a listening socket as I'll get. 

I'm curious-- does auto save complete its process if a transaction is open? I'd imagine that it doesn't. You might want to check on that. It would be a drag to discover after the fact that a series of edits were lost because auto save didn't complete-- especially if you're going to keep using an external cron job script that relies on an unrelated function (auto save) to preserve your financial data. 

⁣David T. ​

On May 8, 2023, 5:55 PM, at 5:55 PM, "David G. Pickett" <dgpickett at aol.com> wrote:
>I use auto save, and if the app is running, the cron kills it and 
>removes the lock before running the quote fetch.
>
>One can imagine the app having a listening socket to let you ask it to 
>get quotes from a cron script.  Or having it maintain a schedule where 
>it gets all quotes, like an internal crontab.
>
>
>-----Original Message-----
>From: David T. <sunfish62 at yahoo.com>
>To: David G. Pickett <dgpickett at aol.com>
>Cc: David G. Pickett via gnucash-user <gnucash-user at gnucash.org>
>Sent: Sun, May 7, 2023 11:34 pm
>Subject: Re: [GNC] Auto commit with auto save?
>
>Nope and nope. Sorry. 
>
>It seems to me that leaving GnuCash open and running a con job against 
>the open app is a recipe for troubles just like the ones you have 
>encountered. You could tell the script to abort if it found the lock 
>file, but that would require you to close the app every night, which 
>you're not doing now.
>
>David T. On May 7, 2023, at 9:25 PM, "David G. Pickett"
><dgpickett at aol.com> wrote:
>Any suggestions on a) how a shell script tells that the auto save is 
>incomplete, b) even if it knew, what it could do about it?
> 
> 
>  -----Original Message-----
> From: David T. <sunfish62 at yahoo.com>
> To: David G. Pickett <dgpickett at aol.com>
> Cc: gnucash-user at gnucash.org
> Sent: Sun, May 7, 2023 1:28 am
> Subject: Re: [GNC] Auto commit with auto save?
> 
>I agree that it would be nice to have some visual cue that a 
>transaction has been edited but not saved; that would be useful. I'm 
>honestly not sure why that hasn't been implemented.
> 
>The app does throw a dialog onscreen when a user tries to save with an 
>open transaction. Unfortunately, the message is generic, and a user is 
>forced to look through the open tabs and try to figure out which 
>register holds this transaction. Others have commented on this in the 
>past.
> 
>If I recall correctly, you were having trouble because you have a cron 
>job set up to retrieve quotes at a specified time each day, and this 
>job causes the file to close dirty if there is an open transaction. The 
>problem in this case, is that this cron job doesn't have necessary 
>fault testing and tolerance. I'd suggest working on ensuring that this 
>cron job was properly set up to handle your specific situation.
> 
>David T.  On May 6, 2023, at 9:02 PM, "David G. Pickett via 
>gnucash-user" < gnucash-user at gnucash.org> wrote:
>"Don't do that!" does not prevent data loss from human error, which for 
>this app behavior is too easy to create and not realize.
>
>         
> 
>If the commit was automatic every time you modified a transaction when 
>the new state was valid, then I would not leave the tab in an 
>uncommitted state, but that is not how it was devised.  I leave because 
>it looks fine.  There is no indication of an uncommitted transaction I 
>can see.  In terms of human factors, one might go to another tab for 
>information to complete a transaction, so we do not want to prevent the 
>user leaving a tab with an uncommitted, possibly invalid transaction.
>Maybe we should color the tab red while in this state?  Or pop up a 
>dialog if it persists a bit too long, or if auto save fires on its 
>timer?  But the user may have left, trusting in auto save, so I suggest 
>an auto commit if valid on auto save.
> 
>gnucash-user mailing list
>gnucash-user at gnucash.org
>To update your subscription preferences or to unsubscribe: 
> https://lists.gnucash.org/mailman/listinfo/gnucash-user
>-----
>Please remember to CC this list on all your replies. 
>You can do this by using Reply-To-List or Reply-All.  
> 
> 
_______________________________________________
gnucash-user mailing list
gnucash-user at gnucash.org
To update your subscription preferences or to unsubscribe:
https://lists.gnucash.org/mailman/listinfo/gnucash-user
-----
Please remember to CC this list on all your replies.
You can do this by using Reply-To-List or Reply-All.