Re: [PATCH, v2] Fortran: detect blanks within literal constants in free-form mode [PR92805]

2022-07-29 Thread Mikael Morin 

Le 29/07/2022 à 21:59, Harald Anlauf via Fortran a écrit :


Am 29.07.22 um 13:11 schrieb Mikael Morin:

 > and use gfc_next_char instead of gfc_match_char

in get_kind.


There is one important functionality in gfc_match_char(): it manages
the locus.  We would need then to add this explicitly to get_kind,
which does not look to me like a big improvement over the present
solution.  Otherwise I get test regressions.

Indeed, I overlooked that, but my opinion remains that we shouldn’t play 
with fixed vs free form considerations here.

So the options I can see are:
 - handle the locus in get_kind; we do it a lot already in matching 
functions, so it wouldn’t be different here.

 - implement a variant of gfc_match_char without space gobbling.
 - use gfc_match(...), which is a bit heavy weight to match a single 
char string, but otherwise would keep things concise.


My preference goes to the third option, but I’m fine with either of them 
if you have a different one.

[PATCH, v2] Fortran: detect blanks within literal constants in free-form mode [PR92805]

2022-07-29 Thread Harald Anlauf via Gcc-patches 

Hi Mikael,

Am 29.07.22 um 13:11 schrieb Mikael Morin:

Hello,

Le 28/07/2022 à 22:11, Harald Anlauf via Fortran a écrit :

Dear all,

in free-form mode, blanks are significant, so they cannot appear
in literal constants, especially not before or after the "_" that
separates the literal and the kind specifier.

The initial patch from Steve addressed numerical literals, which
I completed by adjusting the parsing of string literals.

Regtested on x86_64-pc-linux-gnu.  OK for mainline?


It looks correct, but I think we should continue to have the free vs
fixed form abstracted away from the parsing code.


yes, that makes sense.


So, I suggest instead to remove the calls to gfc_gobble_whitespace in
match_string_constant,


Indeed, removing these simplifies the function and indeed works!

> and use gfc_next_char instead of gfc_match_char

in get_kind.


There is one important functionality in gfc_match_char(): it manages
the locus.  We would need then to add this explicitly to get_kind,
which does not look to me like a big improvement over the present
solution.  Otherwise I get test regressions.


Mikael



I've attached a revised version with improved match_string_constant().
What do you think?

Thanks,
Harald
From f8e7c297b7c9e5a2b22185c7e0d638764c33aa71 Mon Sep 17 00:00:00 2001
From: Harald Anlauf 
Date: Thu, 28 Jul 2022 22:07:02 +0200
Subject: [PATCH] Fortran: detect blanks within literal constants in free-form
 mode [PR92805]

gcc/fortran/ChangeLog:

	PR fortran/92805
	* primary.cc (get_kind): Do not skip over blanks in free-form mode.
	(match_string_constant): Likewise.

gcc/testsuite/ChangeLog:

	PR fortran/92805
	* gfortran.dg/literal_constants.f: New test.
	* gfortran.dg/literal_constants.f90: New test.

Co-authored-by: Steven G. Kargl 
---
 gcc/fortran/primary.cc| 19 +++
 gcc/testsuite/gfortran.dg/literal_constants.f | 20 
 .../gfortran.dg/literal_constants.f90 | 24 +++
 3 files changed, 53 insertions(+), 10 deletions(-)
 create mode 100644 gcc/testsuite/gfortran.dg/literal_constants.f
 create mode 100644 gcc/testsuite/gfortran.dg/literal_constants.f90

diff --git a/gcc/fortran/primary.cc b/gcc/fortran/primary.cc
index 3f01f67cd49..604f98a8dd9 100644
--- a/gcc/fortran/primary.cc
+++ b/gcc/fortran/primary.cc
@@ -92,14 +92,21 @@ get_kind (int *is_iso_c)
 {
   int kind;
   match m;
+  char c;
 
   *is_iso_c = 0;
 
+  c = gfc_peek_ascii_char ();
+  if (gfc_current_form == FORM_FREE && gfc_is_whitespace (c))
+return -2;
+
   if (gfc_match_char ('_') != MATCH_YES)
 return -2;
 
-  m = match_kind_param (, is_iso_c);
-  if (m == MATCH_NO)
+  m = MATCH_NO;
+  c = gfc_peek_ascii_char ();
+  if ((gfc_current_form == FORM_FREE && gfc_is_whitespace (c))
+  || (m = match_kind_param (, is_iso_c)) == MATCH_NO)
 gfc_error ("Missing kind-parameter at %C");
 
   return (m == MATCH_YES) ? kind : -1;
@@ -1074,17 +1081,9 @@ match_string_constant (gfc_expr **result)
   c = gfc_next_char ();
 }
 
-  if (c == ' ')
-{
-  gfc_gobble_whitespace ();
-  c = gfc_next_char ();
-}
-
   if (c != '_')
 goto no_match;
 
-  gfc_gobble_whitespace ();
-
   c = gfc_next_char ();
   if (c != '\'' && c != '"')
 goto no_match;
diff --git a/gcc/testsuite/gfortran.dg/literal_constants.f b/gcc/testsuite/gfortran.dg/literal_constants.f
new file mode 100644
index 000..4d1f1b7eb4c
--- /dev/null
+++ b/gcc/testsuite/gfortran.dg/literal_constants.f
@@ -0,0 +1,20 @@
+! { dg-do compile }
+! { dg-options "-ffixed-form" }
+! PR fortran/92805 - blanks within literal constants in fixed-form mode
+
+  implicit none
+  integer, parameter :: ck = kind ("a")  ! default character kind
+  integer, parameter :: rk = kind (1.0)  ! default real kind
+  print *, 1_"abc"
+  print *, 1 _"abc"
+  print *, 1_ "abc"
+  print *, ck_"a"
+  print *, ck _"ab"
+  print *, ck_ "ab"
+  print *, 3.1415_4
+  print *, 3.1415 _4
+  print *, 3.1415_ 4
+  print *, 3.1415_rk
+  print *, 3.1415 _rk
+  print *, 3.1415_ rk
+  end
diff --git a/gcc/testsuite/gfortran.dg/literal_constants.f90 b/gcc/testsuite/gfortran.dg/literal_constants.f90
new file mode 100644
index 000..f8908f9ad76
--- /dev/null
+++ b/gcc/testsuite/gfortran.dg/literal_constants.f90
@@ -0,0 +1,24 @@
+! { dg-do compile }
+! { dg-options "-ffree-form" }
+! PR fortran/92805 - blanks within literal constants in free-form mode
+
+  implicit none
+  integer, parameter :: ck = kind ("a")  ! default character kind
+  integer, parameter :: rk = kind (1.0)  ! default real kind
+  print *, 1_"abc"
+  print *, 1 _"abc"   ! { dg-error "Syntax error" }
+  print *, 1_ "abc"   ! { dg-error "Missing kind-parameter" }
+  print *, 1 _ "abc"  ! { dg-error "Syntax error" }
+  print *, ck_"a"
+  print *, ck _"ab"   ! { dg-error "Syntax error" }
+  print *, ck_ "ab"   ! { dg-error "Syntax error" }
+