Anywhere on decode window crashes it.

I also restart jtalert every time it crashes just decode window.

Laurie before I go and mess anything else up, do I delete that via the software or is it a file I need to located in a folder?

All of a sudden the bottom half of my JTALert disappeared.

I turned OFF CHECK BEFORE and now it logs successfully.

If I'm not mistaken you don't get very good caching on a remote network file by default.

Mike

1. When a callsign is FIRST decoded to perform a B4 check for the current Band/Mode.
   
   
2. When confirming a QSO was logged by reading the log. Past bad experiences with other loggers has shown that the status reported by the logger (via its api) after logging may not be correct and cannot be relied upon. HRD was notorious for reporting logging success when the QSO was never logged. Direct reading of the log avoids that.
   
   
3. When performing the Alert database rebuild.

de Laurie VK3AMA

Besides I have already figured out how to solve my problem with JTALert NOT logging to ACLog. It seems that when JTALert does the check for Worked Before that is when ACLog pops the error message and the Log entry is deleted. I turned OFF CHECK BEFORE and now it logs successfully. I never understand why JTALert needed to remind you that you have already worked a station before when WSIT-X already does by turning the contact GREEN.

   I suspect the timing issue is ONLY with the Before Check. I am running ACLog on a separate computer thus have to use the Manual Mode to access ACLog. Also do not understand why JTALert needs a PHYSICAL Path to the database when the N3FJP API gives the Client all the access it needs.

So put this thread asleep. I have found a flaw in how the BeFor Check is done when ACLog is on a separate computer. I found a work around for the flaw. If the Developer wants to correct this, I am more than willing to tests an update.

JoAnn Donaldson - AB8YZ 

• First, there is nothing to fix, there is no B4 flaw to correct.

• What is this error message that ACLog is showing? Post an image of this ACLog generated message.

• WSJT-X reporting of B4 is independent of your ACLog log file and can be grossly out-of-sync with ACLog. JTAlert does its B4 checks against your primary log, ACLog, not the WSJTX log file which at any time could be empty, truncated or missing.
  
  
• The Timing adjustments your referring to have NOTHING to do with B4 checks. It used used purely for the QSO logged success checks. This is why the settings entry is under the "Logging" section (funny that).
  
  
• The ACLog API is very inefficient when performing many Callsign lookups (for the B4 checks) at the end of a busy FT8 period where there may be 20, 30, 40 or more callsigns to check. Direct querying of the Log file is the quickest technique.
  
  
• The ACLog API doesn't allow for custom sql queries that JTAlert uses, so direct log access is needed. This applies to all the supported loggers, not just ACLog.
  
  
• You have slowed things down further by having the ACLog file on a different PC than JTAlert so network latency is introduced further slowing log file reads.  JTAlert does cache (in memory) the results of the slow disk file lookups, but there will always be at least one physical log lookup when a callsign is first decoded.
  
  
• The Alert database (used to determine what alerts a callsigns/decode generates) requires direct access to the Log file in order to run the many hundreds of sql queries needed  to rebuild the database. The Alert database is used for high-speed determination of what alerts are triggered, it is impractical to run multiple queries (if even possible) via the ACLog API to check if a Callsign is a needed DXCC, State, Grid, Prefix, etc, at the end of a FT8 period. This in-memory database is used.