However this does not work well with the IDE. For instance the log of a submodel if called by
$call =gams submodel.gms
will be invisible and lost. There are few thing we can do to improve the situation. The settings I use are as follows:
* flags to run submodels better under the IDE
$setglobal ide "ide=%gams.ide% lo=%gams.lo% errorlog=%gams.errorlog% errmsg=1"
$call =gams submodel.gms %ide%
$if errorlevel 1 $abort submodel.gms has issues
This will solve most problems of calling gams via $call when using the IDE, and also works good when not running under the IDE.
The one remaining and somewhat annoying problem (bug) is that clicking on black lines of the submodel log, will cause the IDE to go to listing file of the main model. I.e. the wrong listing file is loaded. Clicking on the red lines work fine: the IDE will go the correct source file.
I think it would be an improvement if GAMS has better built-in support for executing submodels (compile time and execution time) where we don't have to worry about all these parameters and with better interaction with the IDE.