Problems installing EGSnrc on Ubuntu 25.04

[MaF-MPJ]

After updating Ubuntu to version 25.04, I followed the usual procedure to install EGSnrc.
I have versions 14.2.0 of gfortran, gcc, and g++ compilers installed on my system.
When compiling the EGSnrc environment, the following warnings appear:

---

Compiling mortran3.f /home/mario/EGSnrc/HEN_HOUSE/lib/linux64/machine.f

mortran3.f:2278:72:

2278 | 99 O(J+M-1)=ICHAR(CH(M))
| 1
Warning: Fortran 2018 deleted feature: DO termination statement which is not END DO or CONTINUE with label 99 at (1)

mortran3.f:2287:72:
2287 | 5181 CH(L)=CHAR( HC(L) )
| 1
Warning: Fortran 2018 deleted feature: DO termination statement which is not END DO or CONTINUE with label 5181 at (1)
Making mortran3.dat
Compilation succeeded!

---

Mortran compilation ...
OK

Compiling pegs4_linux64.f with flags '-fPIC -O2 -mtune=native -mcmodel=medium'

pegs4_linux64.f:3610:28:
3610 | real4 FUNCTION PAIRFR(E)
| 1
......
3685 | REAL FUNCTION PAIRFR(X)
| 2
Warning: INTENT mismatch in argument 'x' between (1) and (2)
pegs4_linux64.f:3639:28:
3639 | real4 FUNCTION PAIRFZ(E)
| 1
......
3699 | REAL FUNCTION PAIRFZ(X)
| 2
Warning: INTENT mismatch in argument 'x' between (1) and (2)
Compilation succeeded!

---

=> Compiling BEAMnrc tool dosxyz_show ...
gcc -O2 -I/usr/X11R6/include -o /home/mario/EGSnrc/HEN_HOUSE/bin/linux64/dosxyz_show -L/usr/X11R6/lib dosxyz_show.c -lXm -lXt -lX11 -lm

dosxyz_show.c: In function 'get_data':

dosxyz_show.c:1268:3: warning: ignoring return value of 'fscanf' declared with attribute 'warn_unused_result' [-Wunused-result]
1268 | fscanf(fp,"%d\n",&n_material);
| ^~~~~~~~~~~~~~~~~~~~~~~~~~~~~
dosxyz_show.c:1271:5: warning: ignoring return value of 'fscanf' declared with attribute 'warn_unused_result' [-Wunused-result]
1271 | fscanf(fp,"%s\n",material);
| ^~~~~~~~~~~~~~~~~~~~~~~~~~
dosxyz_show.c:1274:31: warning: ignoring return value of 'fscanf' declared with attribute 'warn_unused_result' [-Wunused-result]
1274 | for(i=0; i<n_material; i++) fscanf(fp,"%f",&aux);
.......

---

The result of compiling a simulation code is the following:

$ myapp -i point_source.egsinp
.........
.........
At line 21316 of file egsnrc_linux64.F (unit = 6, file = 'stdout')
Fortran runtime error: Missing comma between descriptors
(/a$)

The used code worked on previous Ubuntu version and actually works on Windows 11.

I've already installed EGSnrc several times on previous versions of Ubuntu and Windows 11, so I'm familiar with the procedure.
Can you help me?

[rtownson]

This is an issue with the newer compilers, if you switch to an older gcc compiler it should work fine.

Just keep in mind that you should really start fresh and clone into a new folder when switching compilers.

[rtownson]

Yes, it should work with gcc-11, that's what I'm using.

[MaF-MPJ]

I switched to gcc-11 but without success. Everything is as it was before. I also tried installing gcc-8 and getting the necessary libraries, but the installation failed:

Unsatisfied dependencies:
libgcc-8-dev : Depends: libasan5 (>= 8.4.0-3ubuntu2) but it is not installable
Error: Unable to correct problems, you have held broken packages.
Error: The following information from --solver 3.0 may provide additional context:
Unable to satisfy dependencies. Reached two conflicting decisions:

1. libgcc-8-dev:amd64=8.4.0-3ubuntu2 is selected for install
2. libgcc-8-dev:amd64 Depends libasan5 (>= 8.4.0-3ubuntu2)
   but none of the choices are installable:
   [no choices]

I really don't have an idea how to resolve the problem.....

[rtownson]

Did you fully delete the old EGSnrc directory, clone fresh, and ensure the new compilers were in use? I would double check in the installation log that it was indeed using the gcc11 compilers, it's easy to make the mistake of environment variables not being updated.

[MaF-MPJ]

Yes, I cloned a fresh one and the compiler was gfortran-11.

[MaF-MPJ]

Unfortunately, there is no gcc distribution prior to gcc-11; at least not one that be easily installable.

[crcrewso]

I looked over my notes, the flag for newer gcc's is -std=legacy. if you add that to your standard Fortran flags you should be good to go.

1. Reinstall gcc/gfortran 14
2. Edit: Run configure/Egs Installer GUI again
3. when prompted in configure
   Input standard compilation flags: [-fPIC]
   replace with -fPIC -std=legacy

[MaF-MPJ]

I was very confident in this solution, but unfortunately it only solved part of the problem.
Let me summarize the situation.
I performed the GUI installation using gfortran-11 and the flags '-fPIC -std=legacy'.
First, I noticed that opening the GUI produces the following message in the terminal window:

Mario@Mubuntu:~$ ./EGSnrc-configure-linux
Case-insensitive sorting unsupported in the Posix collation implementation
Numeric mode unsupported in the Posix collation implementation
Case-insensitive sorting unsupported in the Posix collation implementation
Numeric mode unsupported in the Posix collation implementation

........etc

Comparing the logs produced without and with the addition of '-std=legacy', the following warning message is no longer present in the latter case:

Compiling mortran3.f /home/mario/EGSnrc/HEN_HOUSE/lib/linux64/machine.f

mortran3.f:2278:72:

2278 | 99 O(J+M-1)=ICHAR(CH(M))
| 1
Warning: Fortran 2018 deleted feature: DO termination statement which is not END DO or CONTINUE with label 99 at (1)

mortran3.f:2287:72: 

2287 | 5181 CH(L)=CHAR( HC(L) ) 
| 1
Warning: Fortran 2018 deleted feature: DO termination statement which is not END DO or CONTINUE with label 5181 at (1)

Making mortran3.dat


Compilation succeeded!

However, all other warnings remained unchanged:

Compiling pegs4_linux64.f with flags '-fPIC -std=legacy -O2 -mtune=native -mcmodel=medium'

pegs4_linux64.f:3610:28: 

3610 | real*4 FUNCTION PAIRFR(E) 
| 1
...... 
3685 | REAL FUNCTION PAIRFR(X) 
| 2
Warning: INTENT mismatch in argument 'x' between (1) and (2)
pegs4_linux64.f:3639:28: 

3639 | real*4 FUNCTION PAIRFZ(E) 
| 1
...... 
3699 | REAL FUNCTION PAIRFZ(X) 
| 2
Warning: INTENT mismatch in argument 'x' between (1) and (2)

*****************************************

g++ -I../lib/linux64 -fPIC -DBUILD_IAEA_DLL -O2 -mtune=native -mcmodel=medium -std=c++14 -c -o ../egs++/dso/linux64/utilities.o utilities.cpp

utilities.cpp: In function 'FILE* open_file(char*, const char*, const char*)':
utilities.cpp:227:12: warning: ignoring return value of 'char* fgets(char*, int, FILE*)' declared with attribute 'warn_unused_result' [-Wunused-result] 
227 | fgets(string,MAX_STR_LEN,stdin); 
| ~~~~~^~~~~~~~~~~~~~~~~~~~~~~~~~

utilities.cpp: In function 'int ok_check()':
utilities.cpp:266:9: warning: ignoring return value of 'char* fgets(char*, int, FILE*)' declared with attribute 'warn_unused_result' [-Wunused-result] 
266 | fgets(reply,MAX_STR_LEN,stdin); 
| ~~~~~^~~~~~~~~~~~~~~~~~~~~~~~~

---

=> Compiling BEAMnrc tool dosxyz_show ...


gcc -O2 -I/usr/X11R6/include -o /home/mario/EGSnrc/HEN_HOUSE/bin/linux64/dosxyz_show -L/usr/X11R6/lib dosxyz_show.c -lXm -lXt -lX11 -lm

dosxyz_show.c: In function 'get_data':

dosxyz_show.c:1268:3: warning: ignoring return value of 'fscanf' declared with attribute 'warn_unused_result' [-Wunused-result] 
1268 | fscanf(fp,"%d\n",&n_material); 
| ^~~~~~~~~~~~~~~~~~~~~~~~~~~~~
dosxyz_show.c:1271:5: warning: ignoring return value of 'fscanf' declared with attribute 'warn_unused_result' [-Wunused-result] 
1271 | fscanf(fp,"%s\n",material); 
| ^~~~~~~~~~~~~~~~~~~~~~~~~~
dosxyz_show.c:1274:31: warning: ignoring return value of 'fscanf' declared with attribute 'warn_unused_result' [-Wunused-result] 
1274 | for(i=0; i<n_material; i++) fscanf(fp,"%f",&aux); 

.........etc.

I don't know what to do anymore... I also tried with gfortran-14 and setting the flags as suggested, with the same result....

[MaF-MPJ]

I tried configuring EGSnrc without using the GUI. Everything seemed to go smoothly, with no warnings or error messages during the process.
Even running make in the myapp folder didn't report any errors.
However, trying to run a well-tested simulation file, resulted in the following:

At line 21316 of file egsnrc_linux.F (unit = 6, file = 'stdout')
Fortran runtime error: Missing comma between descriptors
(/a$)
^

Any ideas?

[rtownson]

Those warnings are fine to ignore. I expected the compiler flag to remove this last issue, but as a temporary workaround what you can do is just open HEN_HOUSE/src/egsnrc.mortran, and do a search replace of (/a$) with (/a), then recompile the application.

I'll submit this fix soon, but it won't be in the official release for quite some time.

[MaF-MPJ]

I made the indicated substitution in the HEN_HOUSE/src/egsnrc.mortran file, but I still get the same error:

At line 22111 of file egsnrc_linux.F (unit = 6, file = 'stdout')
Fortran runtime error: Missing comma between descriptors
(a$)
^

I ran a grep search for other occurrences of /a$, but found none in the entire /EGSnrc folder and sub-folders.

[rtownson]

Did you recompile the application? E.g. in egs_home/myapp do make clean; make.

[bilal-medphy]

I got the same error. Again searched for a$ in egsnrc.mortran file, replaced it with a. Recompiled the code. Now it is running myapp code. But I don't know this substitution is correct, but it works

…

On Wed, 22 Oct 2025, 11:42 pm MaF-MPJ, ***@***.***> wrote: I made the indicated substitution in the HEN_HOUSE/src/egsnrc.mortran file, but I still get the same error: At line 22111 of file egsnrc_linux.F (unit = 6, file = 'stdout') Fortran runtime error: Missing comma between descriptors (a$) ^ I ran a grep search for other occurrences of /a$, but found none in the entire /EGSnrc folder and sub-folders. — Reply to this email directly, view it on GitHub <#1335 (reply in thread)>, or unsubscribe <https://github.com/notifications/unsubscribe-auth/AQK3RGUNHHT36DFW6RU7CV33Y7CH7AVCNFSM6AAAAACJTOROIOVHI2DSMVQWIX3LMV43URDJONRXK43TNFXW4Q3PNVWWK3TUHMYTINZVGQYDIOI> . You are receiving this because you are subscribed to this thread.Message ID: ***@***.***>

[bilal-medphy]

Yes I did ‘make clean’ and then ‘make’ I modified two instances of a$ in the egsnrc.mortran file and replaced with a. It compiled with 8 warnings. But successfully runs my input files. Is this correct. Or does it change the getinp. I think these takes data of atomic relaxation and Rayleigh interaction.

…

On Thu, 23 Oct 2025 at 12:06 AM, Reid Townson ***@***.***> wrote: Did you recompile the application? E.g. in egs_home/myapp do make clean; make. — Reply to this email directly, view it on GitHub <#1335 (reply in thread)>, or unsubscribe <https://github.com/notifications/unsubscribe-auth/AQK3RGSLKPSZAIVAEWI53AD3Y7FDHAVCNFSM6AAAAACJTOROIOVHI2DSMVQWIX3LMV43URDJONRXK43TNFXW4Q3PNVWWK3TUHMYTINZVGQZDCNY> . You are receiving this because you commented.Message ID: ***@***.***>

[MaF-MPJ]

I finally did it!
Procedure:

1. Run the configuration by inserting the "-iPIC -std=legacy" flags for gfortran (it doesn't matter whether you use the GUI or not);
2. Open the HEN_HOUSE/src/egsnrc.mortran file and replace "/a$" with "/a" (5 substitutions);
3. In the appropriate folder, run the "make clean" and then "make" commands;
4. Open the egsnrc_linux64.F file created by make in the same folder and replace "a$" with "a" (2 substitutions);
5. Run make again (without first running the make clean command!).
   This worked for the myapp folder, created as instructed in the Getting Started tutorial.

Last note
Does the message "library needed for dlopen -ldl" imply a problem?

Thanks everyone for the great help!
However, I believe a more definitive solution to the problem is needed.

[rtownson]

If you mean the message it prints during installation that -ldl is the library needed, that's normal and not a problem.

Indeed I plan to properly address these issues for the 2026 release.

[MaF-MPJ]

Yes, that's what I meant.. Thanks!