[patch, fortran] PR93486 - ICE on valid with nested submodules and long submodule names

[Andrew Benson]

I opened PR93486 for this problem:

The following causes an ICE with revision 
ad690d79cfbb905c5546c9333c5fd089d906505b:

module ivs
  interface l
     module procedure l_
  end interface l
contains
  function l_()
  end function l_
end module ivs

module aModeratleyLongModuleName
  use ivs
  interface
     module subroutine cmo()
     end subroutine cmo
  end interface
end module aModeratleyLongModuleName

submodule (aModeratleyLongModuleName) 
aNameForASubmoduleThatIsVeryLongButWhichIsLegalStill
contains
  module procedure cmo
  end procedure cmo
end submodule aNameForASubmoduleThatIsVeryLongButWhichIsLegalStill

submodule 
(aModeratleyLongModuleName:aNameForASubmoduleThatIsVeryLongButWhichIsLegalStill)
 
sb
end submodule sb

submodule (aModeratleyLongModuleName:sb) sc
end submodule sc


$ gfortran -v
Using built-in specs.
COLLECT_GCC=gfortran
COLLECT_LTO_WRAPPER=/data001/abenson/Galacticus/Tools_Devel_Install/bin/../
libexec/gcc/x86_64-pc-linux-gnu/10.0.1/lto-wrapper
Target: x86_64-pc-linux-gnu
Configured with: ../gcc-git/configure --prefix=/home/abenson/Galacticus/
Tools_Devel --enable-languages=c,c++,fortran --disable-multilib
Thread model: posix
Supported LTO compression algorithms: zlib
gcc version 10.0.1 20200128 (experimental) (GCC) 


$ gfortran -c test.F90 -o test.o
f951: internal compiler error: Segmentation fault
0xe1bc9f crash_signal
        ../../gcc-git/gcc/toplev.c:328
0x7f98119c71ef ???
        /data001/abenson/Galacticus/Tools/glibc-2.12.1/signal/../sysdeps/unix/
sysv/linux/x86_64/sigaction.c:0
0x84b3f0 gfc_use_modules()
        ../../gcc-git/gcc/fortran/module.c:7315
0x85acc8 use_modules
        ../../gcc-git/gcc/fortran/parse.c:114
0x866daa parse_module
        ../../gcc-git/gcc/fortran/parse.c:6111
0x86733c gfc_parse_file()
        ../../gcc-git/gcc/fortran/parse.c:6428
0x8b780f gfc_be_parse_file
        ../../gcc-git/gcc/fortran/f95-lang.c:210
Please submit a full bug report,
with preprocessed source if appropriate.
Please include the complete backtrace with any bug report.
See <https://gcc.gnu.org/bugs/> for instructions.


The ICE goes away if the module and/or submodule names are made shorter.

The problem is caused when loading (generic or operator) interfaces from a 
module file. The module name can include the submodule name, so a size of 
2*GFC_MAX_SYMBOL_LEN+2 is required to read this in.

The attached patch fixes this problem and regression tests cleanly *if* the 
patch for PR87103 is applied (otherwise that problem gets triggered by the new 
test case).

PR87103 has been a long-standing issue - there's a patch that works (either my 
original ugly fix, or the better fix proposed by Bernhard at https://
gcc.gnu.org/ml/fortran/2018-09/msg00044.html ). It would be very nice to get 
one of these patches committed.

-Andrew

-- 

* Andrew Benson: http://users.obs.carnegiescience.edu/abenson/contact.html

* Galacticus: https://github.com/galacticusorg/galacticus

diff --git a/gcc/fortran/module.c b/gcc/fortran/module.c
index 4487f65..b6a4e87 100644
--- a/gcc/fortran/module.c
+++ b/gcc/fortran/module.c
@@ -4568,7 +4568,9 @@ static void
 load_operator_interfaces (void)
 {
   const char *p;
-  char name[GFC_MAX_SYMBOL_LEN + 1], module[GFC_MAX_SYMBOL_LEN + 1];
+  /* "module" must be large enough for the case of submodules in which the name
+     has the form module.submodule */
+  char name[GFC_MAX_SYMBOL_LEN + 1], module[2 * GFC_MAX_SYMBOL_LEN + 2];
   gfc_user_op *uop;
   pointer_info *pi = NULL;
   int n, i;
@@ -4624,7 +4626,9 @@ static void
 load_generic_interfaces (void)
 {
   const char *p;
-  char name[GFC_MAX_SYMBOL_LEN + 1], module[GFC_MAX_SYMBOL_LEN + 1];
+  /* "module" must be large enough for the case of submodules in which the name
+     has the form module.submodule */
+  char name[GFC_MAX_SYMBOL_LEN + 1], module[2 * GFC_MAX_SYMBOL_LEN + 2];
   gfc_symbol *sym;
   gfc_interface *generic = NULL, *gen = NULL;
   int n, i, renamed;
diff --git a/gcc/testsuite/gfortran.dg/pr93486.f90 b/gcc/testsuite/gfortran.dg/pr93486.f90
new file mode 100644
index 0000000..5037d40
--- /dev/null
+++ b/gcc/testsuite/gfortran.dg/pr93486.f90
@@ -0,0 +1,30 @@
+! { dg-do compile }
+! PR fortran/93486
+module ivs
+  interface l
+     module procedure l_
+  end interface l
+contains
+  function l_()
+  end function l_
+end module ivs
+
+module aModeratleyLongModuleName
+  use ivs
+  interface
+     module subroutine cmo()
+     end subroutine cmo
+  end interface
+end module aModeratleyLongModuleName
+
+submodule (aModeratleyLongModuleName) aNameForASubmoduleThatIsVeryLongButWhichIsLegalStill
+contains
+  module procedure cmo
+  end procedure cmo
+end submodule aNameForASubmoduleThatIsVeryLongButWhichIsLegalStill
+
+submodule (aModeratleyLongModuleName:aNameForASubmoduleThatIsVeryLongButWhichIsLegalStill) sb
+end submodule sb
+
+submodule (aModeratleyLongModuleName:sb) sc
+end submodule sc

2020-01-28  Andrew Benson  <abenso...@gmail.com>

	PR fortran/93486
	* module.c: Increase size of variables used to read module names
	when loading interfaces from module files to permit cases where
	the name is the concatenation of a module and submodule name.
	* gfortran.dg/pr93486.f90: New test.