Topic: BUILD should block RUN until done and cancel it if unsuccessful

When one clicks BUILD, followed swiftly by clicking RUN, Simply Fortran ought to hold and block execution of the RUN command until the BUILD has completed.
Put differently, a click on RUN should cause the program to check that there is not an unfinished BUILD operation in progress before actually RUNning.

At present I keep making a slightly irritating error when re-compiling a long program.  In my present project, compiling, linking and loading takes 10 or 15 seconds.  When I hover the mouse over RUN I seem often to click it a split second too soon, when BUILD has not quite finished.  The result is that the BUILD function is unable to save the executable (because it is in use), and my 10 or 15 second BUILD fails and has to be repeated!  This, in software development, is an interminable delay!

Making BUILD block any RUN command would prevent this.

Ideally, one should be able to click BUILD then RUN in quick succession and have the run wait for the BUILD to finish.  A further suggestion:  When RUN is clicked within, say, one second after BUILD, the program could recognise that the user is trying to test a new build, and actually CANCEL the RUN if the BUILD fails.
This would be another tiny time-saver, adding more slickness to the program.
----
John

  • jeff
  • Administrator
  • Offline

Re: BUILD should block RUN until done and cancel it if unsuccessful

John,

These are logical suggestions.  The current "checks" to make sure builds are complete before running are admittedly rudimentary.  I think it makes sense to add some more logical backend functionality to these operations.

Jeff Armstrong
Approximatrix, LLC