bugs.txt

Mask | File 2021-06-20 FIXED

## 2024-11-01 locked up when region.N with wrong region (GSM) cf line 1438

## 2024-11-03 secr help not working?

## 2024-11-03 check dual use of suggestbufferlink (Main and Habitat mask tabs) wrt enable/disable

# 2024-11-21 The calculation of expected time for fitting a model is based on a 
single call to secr.fit() with the details argument LLonly = TRUE. Prior to 2024-11-21 this was done twice, 
once for display purposes and once at the point of model fitting to compare to the warning and total limits. 
Strangely, the second expected time was typically many times larger.

This could be reproduced in the browser simply by calling secr.fit twice, even with identical 
arguments, sessionInfo and environment variables (checked with Sys.getenv). The effect only happens reliably in a fresh R session.

The spurious overestimate has been largely avoided by making the calculation only once and saving the value in timerv$expected.