View previous topic :: View next topic |
Author |
Message |
stfark1
Joined: 02 Sep 2008 Posts: 224
|
Posted: Tue Apr 05, 2022 11:44 pm Post subject: Strange Entry |
|
|
Using 64 bit debug, have an entry point in one of my subroutines called XENTER. When I debug and "go to" the routine that has this entry point, the commands execute down to the entry point then abort, return to the calling routine. strange issue, has this been reported as a bug in the 64 bit debug function or? please let me know, Sid Kraft |
|
Back to top |
|
|
PaulLaidler Site Admin
Joined: 21 Feb 2005 Posts: 8037 Location: Salford, UK
|
Posted: Wed Apr 06, 2022 6:33 am Post subject: |
|
|
Sid
ENTRY points are "depreciated" in modern Fortran. In other words it is considered to be bad practice to use ENTRY points. They are supported because they appear in legacy Fortran code.
If you need to use ENTRY points then please post a simple sample program that illustrates the failure that you are getting. |
|
Back to top |
|
|
mecej4
Joined: 31 Oct 2006 Posts: 1897
|
Posted: Wed Apr 06, 2022 12:53 pm Post subject: |
|
|
Sid, it is usually fruitless to ask if an problem that you are encountering with a program now was addressed in the past.
Unless you have provided a reproducer for the problem or given a precise and complete description, searching old posts for something similar is not worth the effort. Even if we find an old bug report, such as http://forums.silverfrost.com/viewtopic.php?t=3743 for your ENTRY bug, it is unlikely that you are using the old version of FTN95 that contained the bug, or that the old bug is pertinent. That report pertained to the 8.3 version (and earlier versions, perhaps) of the compiler, and that bug is no longer present in the current 8.83 compiler.
As Paul advised, please provide a reproducer. When you report any bug, please state the compiler version and compiler options used. |
|
Back to top |
|
|
stfark1
Joined: 02 Sep 2008 Posts: 224
|
Posted: Thu Apr 07, 2022 2:03 am Post subject: Strange Entry |
|
|
OK, used to work but now does not, will re-structure the code to eliminate the entry function. Not sure why it was eliminated as it is an easy way to provide an entry to "go around" existing code. Thanks anyway, Sid Kraft |
|
Back to top |
|
|
stfark1
Joined: 02 Sep 2008 Posts: 224
|
Posted: Thu Apr 07, 2022 2:09 am Post subject: Strange Entry |
|
|
Mecej4:
Would be almost impossible to present the sample that is causing the error, my project is a mathematical, surface "fitting" program, 32 Fortran subroutines, average 30,000+ Fortran statements each. Not sure that I have the latest version 64 bit program, version 8.83 but will check. Where can I download the latest version from? Sid Kraft |
|
Back to top |
|
|
PaulLaidler Site Admin
Joined: 21 Feb 2005 Posts: 8037 Location: Salford, UK
|
|
Back to top |
|
|
JohnCampbell
Joined: 16 Feb 2006 Posts: 2593 Location: Sydney
|
Posted: Sat Apr 09, 2022 6:50 am Post subject: |
|
|
Subroutines that used ENTRY typically required that all local variables were SAVE (static). This was the default with old "F66" compilers.
The present (F95+) default is dynamic allocation of local variabes/arrays to the stack, so when the routine is re-entered all previous values were lost.
This is a likely cause of your problem, but not definately.
There are compiler options for local variables to be saved and not placed on the stack.
See /Save and also /LOcal_zero, /OLd_arrays and /DO1 which retain some old FORTRAN behaviour. |
|
Back to top |
|
|
|