[PATCH v2] libcpp: Implement -Wbidi-chars for CVE-2021-42574 [PR103026]

Marek Polacek polacek@redhat.com
Tue Nov 16 19:50:44 GMT 2021


On Mon, Nov 15, 2021 at 06:15:40PM -0500, David Malcolm wrote:
> > On Mon, Nov 08, 2021 at 04:33:43PM -0500, Marek Polacek wrote:
> > > Ping, can we conclude on the name?   IMHO, -Wbidirectional is just fine,
> > > but changing the name is a trivial operation. 
> > 
> > Here's a patch with a better name (suggested by Jonathan W.).  Otherwise no
> > changes.
> 
> Thanks for implementing this.
> 
> > 
> > Bootstrapped/regtested on x86_64-pc-linux-gnu, ok for trunk?
> > 
> > -- >8 --
> > From a link below:
> > "An issue was discovered in the Bidirectional Algorithm in the Unicode
> > Specification through 14.0. It permits the visual reordering of
> > characters via control sequences, which can be used to craft source code
> > that renders different logic than the logical ordering of tokens
> > ingested by compilers and interpreters. Adversaries can leverage this to
> > encode source code for compilers accepting Unicode such that targeted
> > vulnerabilities are introduced invisibly to human reviewers."
> > 
> > More info:
> > https://nvd.nist.gov/vuln/detail/CVE-2021-42574
> > https://trojansource.codes/
> > 
> > This is not a compiler bug.  However, to mitigate the problem, this patch
> > implements -Wbidi-chars=[none|unpaired|any] to warn about possibly
> > misleading Unicode bidirectional characters the preprocessor may encounter.
> > 
> > The default is =unpaired, which warns about improperly terminated
> > bidirectional characters; e.g. a LRE without its appertaining PDF.  The
> 
> I like the default.

Great.

> Wording nit: maybe use "corresponding" rather than "appertaining"; I
> believe the latter has a sense that one is part of the other, when they
> are more like peers.

OK, fixed.

> > level =any warns about any use of bidirectional characters.
> 
> Terminology nit:
> The patch is referring to "bidirectional characters", but I think the
> term "bidirectional control characters" would be better.

Adjusted.
 
> For example, a passage of text containing both numbers and characters
> in a right-to-left script could be considered "bidirectional", since
> the numbers are written from left-to-right.
> 
> Specifically, the patch looks for these specific characters:
>   * U+202A LEFT-TO-RIGHT EMBEDDING
>   * U+202B RIGHT-TO-LEFT EMBEDDING
>   * U+202C POP DIRECTIONAL FORMATTING
>   * U+202D LEFT-TO-RIGHT OVERRIDE
>   * U+202E RIGHT-TO-LEFT OVERRIDE
>   * U+2066 LEFT-TO-RIGHT ISOLATE
>   * U+2067 RIGHT-TO-LEFT ISOLATE
>   * U+2068 FIRST STRONG ISOLATE
>   * U+2069 POP DIRECTIONAL ISOLATE
> 
> However, the following characters could also be considered as
> "bidirectional control characters":
>   * U+200E ‎LEFT-TO-RIGHT MARK (UTF-8: E2 80 8E)
>   * U+200F ‎RIGHT-TO-LEFT MARK (UTF-8: E2 80 8F)
> but aren't checked for in the patch.  Should they be?  I can imagine
> ways in which they could be abused, so I think so.

I'd only intended to check the bidi chars described in the original
trojan source pdf, but I added checking for U+200E/U+200F too, since
it was easy enough.  AFAIK they aren't popped by a PDF/PDI like the
rest, so don't need to go on the vec, and so we only warn with =any.
Tests: Wbidi-chars-16.c + Wbidi-chars-17.c
  
> [...snip...]
> 
> > diff --git a/gcc/c-family/c.opt b/gcc/c-family/c.opt
> > index 06457ac739e..b047df0f125 100644
> > --- a/gcc/c-family/c.opt
> > +++ b/gcc/c-family/c.opt
> > @@ -374,6 +374,30 @@ Wbad-function-cast
> >  C ObjC Var(warn_bad_function_cast) Warning
> >  Warn about casting functions to incompatible types.
> >  
> > +Wbidi-chars
> > +C ObjC C++ ObjC++ Warning Alias(Wbidi-chars=,any,none)
> > +;
> > +
> > +Wbidi-chars=
> > +C ObjC C++ ObjC++ RejectNegative Joined Warning CPP(cpp_warn_bidirectional) CppReason(CPP_W_BIDIRECTIONAL) Var(warn_bidirectional) Init(bidirectional_unpaired) Enum(cpp_bidirectional_level)
> > +-Wbidi-chars=[none|unpaired|any] Warn about UTF-8 bidirectional characters.
> 
> "control characters"
 
Fixed.

> [...snip...]
> 
> >  
> > +@item -Wbidi-chars=@r{[}none@r{|}unpaired@r{|}any@r{]}
> > +@opindex Wbidi-chars=
> > +@opindex Wbidi-chars
> > +@opindex Wno-bidi-chars
> > +Warn about possibly misleading UTF-8 bidirectional characters in comments,
> 
> (and here again)
 
Fixed.

> > +string literals, character constants, and identifiers.  Such characters can
> > +change left-to-right writing direction into right-to-left (and vice versa),
> > +which can cause confusion between the logical order and visual order.  This
> > +may be dangerous; for instance, it may seem that a piece of code is not
> > +commented out, whereas it in fact is.
> > +
> > +There are three levels of warning supported by GCC@.  The default is
> > +@option{-Wbidi-chars=unpaired}, which warns about improperly terminated
> > +bidi contexts.  @option{-Wbidi-chars=none} turns the warning off.
> > +@option{-Wbidi-chars=any} warns about any use of bidirectional characters.
> 
> (and again)

Fixed.

> [...snip...]
> 
> 
> > diff --git a/gcc/testsuite/c-c++-common/Wbidi-chars-4.c b/gcc/testsuite/c-c++-common/Wbidi-chars-4.c
> > new file mode 100644
> > index 00000000000..9fd4bc535ca
> > --- /dev/null
> > +++ b/gcc/testsuite/c-c++-common/Wbidi-chars-4.c
> > @@ -0,0 +1,166 @@
> > +/* PR preprocessor/103026 */
> > +/* { dg-do compile } */
> > +/* { dg-options "-Wbidi-chars=any -Wno-multichar -Wno-overflow" } */
> > +/* Test all bidi chars in various contexts (identifiers, comments,
> > +   string literals, character constants), both UCN and UTF-8.  The bidi
> > +   chars here are properly terminated, except for the character constants.  */
> > +
> > +/* a b c LRE‪ 1 2 3 PDF‬ x y z */
> > +/* { dg-warning "U\\+202A" "" { target *-*-* } .-1 } */
> > +/* a b c RLE‫ 1 2 3 PDF‬ x y z */
> > +/* { dg-warning "U\\+202B" "" { target *-*-* } .-1 } */
> > +/* a b c LRO‭ 1 2 3 PDF‬ x y z */
> > +/* { dg-warning "U\\+202D" "" { target *-*-* } .-1 } */
> > +/* a b c RLO‮ 1 2 3 PDF‬ x y z */
> > +/* { dg-warning "U\\+202E" "" { target *-*-* } .-1 } */
> > +/* a b c LRI⁦ 1 2 3 PDI⁩ x y z */
> > +/* { dg-warning "U\\+2066" "" { target *-*-* } .-1 } */
> > +/* a b c RLI⁧ 1 2 3 PDI⁩ x y */
> > +/* { dg-warning "U\\+2067" "" { target *-*-* } .-1 } */
> > +/* a b c FSI⁨ 1 2 3 PDI⁩ x y z */
> > +/* { dg-warning "U\\+2068" "" { target *-*-* } .-1 } */
> 
> AIUI the Unicode bidirectionality algorithm works at the line level,
> and so each line in a block comment should be checked individually for
> unclossed bidi control chars, rather than a block comment as a whole. 
> Hence I think the test case needs to have block comment test coverage
> for:
> - single line blocks
> - first line of a multiline block comment
> - middle line of a multiline block comment
> - final line of a multiline block comment
> but I think the patch as it stands is only checking for the first of
> these four cases.

The patch handles all of them, because of:
1534           if (warn_bidi_p)
1535             maybe_warn_bidi_on_close (pfile, cur);
in _cpp_skip_block_comment, but I was lacking some more testing, so I've
added some testing, and included a new test: Wbidi-chars-15.c.

> > diff --git a/gcc/testsuite/c-c++-common/Wbidi-chars-5.c b/gcc/testsuite/c-c++-common/Wbidi-chars-5.c
> > new file mode 100644
> > index 00000000000..efb26309b68
> > --- /dev/null
> > +++ b/gcc/testsuite/c-c++-common/Wbidi-chars-5.c
> > @@ -0,0 +1,166 @@
> > +/* PR preprocessor/103026 */
> > +/* { dg-do compile } */
> > +/* { dg-options "-Wbidi-chars=unpaired -Wno-multichar -Wno-overflow" } */
> > +/* Test all bidi chars in various contexts (identifiers, comments,
> > +   string literals, character constants), both UCN and UTF-8.  The bidi
> > +   chars here are properly terminated, except for the character constants.  */
> 
> Similar comment as above re block comments

I've extended the test too.
 
> [...snip...]
> 
> > diff --git a/libcpp/include/cpplib.h b/libcpp/include/cpplib.h
> > index 176f8c5bbce..60cf08ddd35 100644
> > --- a/libcpp/include/cpplib.h
> > +++ b/libcpp/include/cpplib.h
> > @@ -319,6 +319,17 @@ enum cpp_main_search
> >    CMS_system,  /* Search the system INCLUDE path.  */
> >  };
> >  
> > +/* The possible bidirectional characters checking levels, from least
> > +   restrictive to most.  */
> > +enum cpp_bidirectional_level {
> > +  /* No checking.  */
> > +  bidirectional_none,
> > +  /* Only detect unpaired uses of bidirectional characters.  */
> > +  bidirectional_unpaired,
> > +  /* Detect any use of bidirectional characters.  */
> > +  bidirectional_any
> > +};
> 
> As before, "control characters".

Fixed.
 
> [...snip...]
> 
> > @@ -539,6 +550,10 @@ struct cpp_options
> >    /* True if warn about differences between C++98 and C++11.  */
> >    bool cpp_warn_cxx11_compat;
> >  
> > +  /* Nonzero of bidirectional characters checking is on.  See enum
> 
> s/of/if/
> 
> and usual nit about "control characters".

Fixed.
 
> > +     cpp_bidirectional_level.  */
> > +  unsigned char cpp_warn_bidirectional;
> > +
> >    /* Dependency generation.  */
> >    struct
> >    {
> 
> [...snip...]
> 
> > diff --git a/libcpp/lex.c b/libcpp/lex.c
> > index fa2253d41c3..3fb518e202b 100644
> > --- a/libcpp/lex.c
> > +++ b/libcpp/lex.c
> > @@ -1164,6 +1164,300 @@ _cpp_process_line_notes (cpp_reader *pfile, int in_comment)
> >      }
> >  }
> >  
> > +namespace bidi {
> > +  enum class kind {
> > +    NONE, LRE, RLE, LRO, RLO, LRI, RLI, FSI, PDF, PDI
> > +  };
> > +
> > +  /* All the UTF-8 encodings of bidi characters start with E2.  */
> > +  constexpr uchar utf8_start = 0xe2;
> 
> Is there a difference between "constexpr" vs "const" here?  (sorry for
> my ignorance)

I just wanted to make sure that utf8_start will be usable in contexts where
an integral constant expression is required.  'const' objects need not be
initialized with compile-time constants, but 'constexpr' objects do.
 
> > +
> > +  /* A vector holding currently open bidi contexts.  We use a char for
> > +     each context, its LSB is 1 if it represents a PDF context, 0 if it
> > +     represents a PDI context.  The next bit is 1 if this context was open
> > +     by a bidi character written as a UCN, and 0 when it was UTF-8.  */
> > +  semi_embedded_vec <unsigned char, 16> vec;
> > +
> > +  /* Close the whole comment/identifier/string literal/character constant
> > +     context.  */
> > +  void on_close ()
> > +  {
> > +    vec.truncate (0);
> > +  }
> > +
> > +  /* Pop the last element in the vector.  */
> > +  void pop ()
> > +  {
> > +    unsigned int len = vec.count ();
> > +    gcc_checking_assert (len > 0);
> > +    vec.truncate (len - 1);
> > +  }
> > +
> > +  /* Return the context of the Ith element.  */
> > +  kind ctx_at (unsigned int i)
> > +  {
> > +    return (vec[i] & 1) ? kind::PDF : kind::PDI;
> > +  }
> > +
> > +  /* Return which context is currently opened.  */
> > +  kind current_ctx ()
> > +  {
> > +    unsigned int len = vec.count ();
> > +    if (len == 0)
> > +      return kind::NONE;
> > +    return ctx_at (len - 1);
> > +  }
> > +
> > +  /* Return true if the current context comes from a UCN origin, that is,
> > +     the bidi char which started this bidi context was written as a UCN.  */
> > +  bool current_ctx_ucn_p ()
> > +  {
> > +    unsigned int len = vec.count ();
> > +    gcc_checking_assert (len > 0);
> > +    return (vec[len - 1] >> 1) & 1;
> > +  }
> > +
> > +  /* We've read a bidi char, update the current vector as necessary.  */
> > +  void on_char (kind k, bool ucn_p)
> > +  {
> > +    switch (k)
> > +      {
> > +      case kind::LRE:
> > +      case kind::RLE:
> > +      case kind::LRO:
> > +      case kind::RLO:
> > +	vec.push (ucn_p ? 3u : 1u);
> > +	break;
> > +      case kind::LRI:
> > +      case kind::RLI:
> > +      case kind::FSI:
> > +	vec.push (ucn_p ? 2u : 0u);
> > +	break;
> 
> I don't like the hand-coded bit fields here, where bit 1 and bit 2 in
> the above have special meaning, but aren't clearly labelled as such.
> 
> Please can you at least use some kind of constant/define to make clear
> the meaning of the bits.  Even clearer would be bitfields; is there a
> performance reason for not using them?  (though this code is only
> called on bidi control chars, which presumably is a rare occurrence). 
> My patch here:
>   "[PATCH 2/2] Capture locations of bidi chars and underline ranges"
>     https://gcc.gnu.org/pipermail/gcc-patches/2021-November/583160.html
> did some refactoring of this patch, replacing hand-coded bit
> manipulation with bitfields in a struct (as well as then using that as
> a good place to stach location_t values, and then using these
> locations).
> Would it be helpful if I split that part of my patch out?
 
I think they are just fine here, given they are used only in bidi:: and
not outside of it.  And I could just use a simple unsigned char in
semi_embedded_vec instead of inventing a new struct.

Your diagnostic patch changes it because you need to remember a location
too, so we're changing it anyway, so I left it be so that you have fewer
conflicts.
 
> [...snip...]
> 
> > +/* Parse a sequence of 3 bytes starting with P and return its bidi code.  */
> > +
> > +static bidi::kind
> > +get_bidi_utf8 (const unsigned char *const p)
> > +{
> > +  gcc_checking_assert (p[0] == bidi::utf8_start);
> > +
> > +  if (p[1] == 0x80)
> > +    switch (p[2])
> 
> get_bidi_utf8 accesss up to 2 bytes beyond "p"...
> 
> [...snip...]
> 
> ...and is called in various places such as...
> 
> > @@ -1218,6 +1519,13 @@ _cpp_skip_block_comment (cpp_reader *pfile)
> >  
> >  	  cur = buffer->cur;
> >  	}
> > +      /* If this is a beginning of a UTF-8 encoding, it might be
> > +	 a bidirectional character.  */
> > +      else if (__builtin_expect (c == bidi::utf8_start, 0) && warn_bidi_p)
> > +	{
> > +	  bidi::kind kind = get_bidi_utf8 (cur - 1);
> > +	  maybe_warn_bidi_on_char (pfile, cur, kind, /*ucn_p=*/false);
> > +	}
> 
> Are we guaranteed to have a '\n' at the end of the buffer?  (even for
> the final line of the file)  That would ensure that we don't read past
> the end of the buffer.

We've discussed this in our internal thread; I think I said that you
will always get a '\n' so this is not going to read past the end of
the buffer.  To be sure...
 
> Can we have testcases involving malformed UTF-8, in which, say:
> - the final byte of the input file is 0xe2
> - the final two bytes of the input file are 0xe2 0x80
> for each of block comment, C++-style comment, string-literal,
> identifier, etc?
> (or is that overkill?)

...I'd crafted a malformed text file using hexedit but couldn't get
it to crash.  I'd rather not include it in the testsuite though.

> [...snip...]
> 
> > @@ -1505,13 +1855,17 @@ lex_identifier (cpp_reader *pfile, const uchar *base, bool starts_ucn,
> >      {
> >        /* Slower version for identifiers containing UCNs
> >  	 or extended chars (including $).  */
> > -      do {
> > -	while (ISIDNUM (*pfile->buffer->cur))
> > -	  {
> > -	    NORMALIZE_STATE_UPDATE_IDNUM (nst, *pfile->buffer->cur);
> > -	    pfile->buffer->cur++;
> > -	  }
> > -      } while (forms_identifier_p (pfile, false, nst));
> > +      do
> > +	{
> > +	  while (ISIDNUM (*pfile->buffer->cur))
> > +	    {
> > +	      NORMALIZE_STATE_UPDATE_IDNUM (nst, *pfile->buffer->cur);
> > +	      pfile->buffer->cur++;
> > +	    }
> > +	}
> > +      while (forms_identifier_p (pfile, false, nst));
> 
> Is the above purely a whitespace change?

Yes.
 
> > +      if (warn_bidi_p)
> > +	maybe_warn_bidi_on_close (pfile, pfile->buffer->cur);
> >        result = _cpp_interpret_identifier (pfile, base,
> >  					  pfile->buffer->cur - base);
> >        *spelling = cpp_lookup (pfile, base, pfile->buffer->cur - base);
> > @@ -1758,6 +2112,8 @@ static void
> >  lex_raw_string (cpp_reader *pfile, cpp_token *token, const uchar *base)
> >  {
> >    const uchar *pos = base;
> > +  const bool warn_bidi_p = (CPP_OPTION (pfile, cpp_warn_bidirectional)
> > +			    != bidirectional_none);
> 
> There are lots of places where the patch uses:
> 
> const bool warn_bidi_p = (CPP_OPTION (pfile, cpp_warn_bidirectional)
> > +			    != bidirectional_none);
> 
> Maybe make it an inline member function:
> 
>   bool warn_bidi_p () const
>   {
>     return CPP_OPTION (this, cpp_warn_bidirectional) != bidirectional_none;
>   }
> 
> so that these can all be:
> 
>   const bool warn_bidi_p = pfile->warn_bidi_p ();
> 
> ?

OK, I guess pfile->warn_bidi_p () is visually more appealing.  Since adding
a member function won't change the PODness of cpp_reader, I went with
your suggestion and added warn_bidi_p.

> Hope this is constructive; thanks again for the patch

Thanks a lot for taking a look!

Bootstrapped/regtested on x86_64-pc-linux-gnu, ok for trunk?

-- >8 --
>From a link below:
"An issue was discovered in the Bidirectional Algorithm in the Unicode
Specification through 14.0. It permits the visual reordering of
characters via control sequences, which can be used to craft source code
that renders different logic than the logical ordering of tokens
ingested by compilers and interpreters. Adversaries can leverage this to
encode source code for compilers accepting Unicode such that targeted
vulnerabilities are introduced invisibly to human reviewers."

More info:
https://nvd.nist.gov/vuln/detail/CVE-2021-42574
https://trojansource.codes/

This is not a compiler bug.  However, to mitigate the problem, this patch
implements -Wbidi-chars=[none|unpaired|any] to warn about possibly
misleading Unicode bidirectional control characters the preprocessor may
encounter.

The default is =unpaired, which warns about improperly terminated
bidirectional control characters; e.g. a LRE without its corresponding PDF.
The level =any warns about any use of bidirectional control characters.

This patch handles both UCNs and UTF-8 characters.  UCNs designating
bidi characters in identifiers are accepted since r204886.  Then r217144
enabled -fextended-identifiers by default.  Extended characters in C/C++
identifiers have been accepted since r275979.  However, this patch still
warns about mixing UTF-8 and UCN bidi characters; there seems to be no
good reason to allow mixing them.

We warn in different contexts: comments (both C and C++-style), string
literals, character constants, and identifiers.  Expectedly, UCNs are ignored
in comments and raw string literals.  The bidirectional control characters
can nest so this patch handles that as well.

I have not included nor tested this at all with Fortran (which also has
string literals and line comments).

Dave M. posted patches improving diagnostic involving Unicode characters.
This patch does not make use of this new infrastructure yet.

	PR preprocessor/103026

gcc/c-family/ChangeLog:

	* c.opt (Wbidi-chars, Wbidi-chars=): New option.

gcc/ChangeLog:

	* doc/invoke.texi: Document -Wbidi-chars.

libcpp/ChangeLog:

	* include/cpplib.h (enum cpp_bidirectional_level): New.
	(struct cpp_options): Add cpp_warn_bidirectional.
	(enum cpp_warning_reason): Add CPP_W_BIDIRECTIONAL.
	* internal.h (struct cpp_reader): Add warn_bidi_p member
	function.
	* init.c (cpp_create_reader): Set cpp_warn_bidirectional.
	* lex.c (bidi): New namespace.
	(get_bidi_utf8): New function.
	(get_bidi_ucn): Likewise.
	(maybe_warn_bidi_on_close): Likewise.
	(maybe_warn_bidi_on_char): Likewise.
	(_cpp_skip_block_comment): Implement warning about bidirectional
	control characters.
	(skip_line_comment): Likewise.
	(forms_identifier_p): Likewise.
	(lex_identifier): Likewise.
	(lex_string): Likewise.
	(lex_raw_string): Likewise.

gcc/testsuite/ChangeLog:

	* c-c++-common/Wbidi-chars-1.c: New test.
	* c-c++-common/Wbidi-chars-2.c: New test.
	* c-c++-common/Wbidi-chars-3.c: New test.
	* c-c++-common/Wbidi-chars-4.c: New test.
	* c-c++-common/Wbidi-chars-5.c: New test.
	* c-c++-common/Wbidi-chars-6.c: New test.
	* c-c++-common/Wbidi-chars-7.c: New test.
	* c-c++-common/Wbidi-chars-8.c: New test.
	* c-c++-common/Wbidi-chars-9.c: New test.
	* c-c++-common/Wbidi-chars-10.c: New test.
	* c-c++-common/Wbidi-chars-11.c: New test.
	* c-c++-common/Wbidi-chars-12.c: New test.
	* c-c++-common/Wbidi-chars-13.c: New test.
	* c-c++-common/Wbidi-chars-14.c: New test.
	* c-c++-common/Wbidi-chars-15.c: New test.
	* c-c++-common/Wbidi-chars-16.c: New test.
	* c-c++-common/Wbidi-chars-17.c: New test.
---
 gcc/c-family/c.opt                          |  24 ++
 gcc/doc/invoke.texi                         |  21 +-
 gcc/testsuite/c-c++-common/Wbidi-chars-1.c  |  12 +
 gcc/testsuite/c-c++-common/Wbidi-chars-10.c |  27 ++
 gcc/testsuite/c-c++-common/Wbidi-chars-11.c |  13 +
 gcc/testsuite/c-c++-common/Wbidi-chars-12.c |  19 +
 gcc/testsuite/c-c++-common/Wbidi-chars-13.c |  17 +
 gcc/testsuite/c-c++-common/Wbidi-chars-14.c |  38 ++
 gcc/testsuite/c-c++-common/Wbidi-chars-15.c |  33 ++
 gcc/testsuite/c-c++-common/Wbidi-chars-16.c |  26 ++
 gcc/testsuite/c-c++-common/Wbidi-chars-17.c |  30 ++
 gcc/testsuite/c-c++-common/Wbidi-chars-2.c  |   9 +
 gcc/testsuite/c-c++-common/Wbidi-chars-3.c  |  11 +
 gcc/testsuite/c-c++-common/Wbidi-chars-4.c  | 188 +++++++++
 gcc/testsuite/c-c++-common/Wbidi-chars-5.c  | 188 +++++++++
 gcc/testsuite/c-c++-common/Wbidi-chars-6.c  | 155 +++++++
 gcc/testsuite/c-c++-common/Wbidi-chars-7.c  |   9 +
 gcc/testsuite/c-c++-common/Wbidi-chars-8.c  |  13 +
 gcc/testsuite/c-c++-common/Wbidi-chars-9.c  |  29 ++
 libcpp/include/cpplib.h                     |  18 +-
 libcpp/init.c                               |   1 +
 libcpp/internal.h                           |   7 +
 libcpp/lex.c                                | 424 +++++++++++++++++++-
 23 files changed, 1298 insertions(+), 14 deletions(-)
 create mode 100644 gcc/testsuite/c-c++-common/Wbidi-chars-1.c
 create mode 100644 gcc/testsuite/c-c++-common/Wbidi-chars-10.c
 create mode 100644 gcc/testsuite/c-c++-common/Wbidi-chars-11.c
 create mode 100644 gcc/testsuite/c-c++-common/Wbidi-chars-12.c
 create mode 100644 gcc/testsuite/c-c++-common/Wbidi-chars-13.c
 create mode 100644 gcc/testsuite/c-c++-common/Wbidi-chars-14.c
 create mode 100644 gcc/testsuite/c-c++-common/Wbidi-chars-15.c
 create mode 100644 gcc/testsuite/c-c++-common/Wbidi-chars-16.c
 create mode 100644 gcc/testsuite/c-c++-common/Wbidi-chars-17.c
 create mode 100644 gcc/testsuite/c-c++-common/Wbidi-chars-2.c
 create mode 100644 gcc/testsuite/c-c++-common/Wbidi-chars-3.c
 create mode 100644 gcc/testsuite/c-c++-common/Wbidi-chars-4.c
 create mode 100644 gcc/testsuite/c-c++-common/Wbidi-chars-5.c
 create mode 100644 gcc/testsuite/c-c++-common/Wbidi-chars-6.c
 create mode 100644 gcc/testsuite/c-c++-common/Wbidi-chars-7.c
 create mode 100644 gcc/testsuite/c-c++-common/Wbidi-chars-8.c
 create mode 100644 gcc/testsuite/c-c++-common/Wbidi-chars-9.c

diff --git a/gcc/c-family/c.opt b/gcc/c-family/c.opt
index 8a4cd634f77..3976fc368db 100644
--- a/gcc/c-family/c.opt
+++ b/gcc/c-family/c.opt
@@ -374,6 +374,30 @@ Wbad-function-cast
 C ObjC Var(warn_bad_function_cast) Warning
 Warn about casting functions to incompatible types.
 
+Wbidi-chars
+C ObjC C++ ObjC++ Warning Alias(Wbidi-chars=,any,none)
+;
+
+Wbidi-chars=
+C ObjC C++ ObjC++ RejectNegative Joined Warning CPP(cpp_warn_bidirectional) CppReason(CPP_W_BIDIRECTIONAL) Var(warn_bidirectional) Init(bidirectional_unpaired) Enum(cpp_bidirectional_level)
+-Wbidi-chars=[none|unpaired|any] Warn about UTF-8 bidirectional control characters.
+
+; Required for these enum values.
+SourceInclude
+cpplib.h
+
+Enum
+Name(cpp_bidirectional_level) Type(int) UnknownError(argument %qs to %<-Wbidi-chars%> not recognized)
+
+EnumValue
+Enum(cpp_bidirectional_level) String(none) Value(bidirectional_none)
+
+EnumValue
+Enum(cpp_bidirectional_level) String(unpaired) Value(bidirectional_unpaired)
+
+EnumValue
+Enum(cpp_bidirectional_level) String(any) Value(bidirectional_any)
+
 Wbool-compare
 C ObjC C++ ObjC++ Var(warn_bool_compare) Warning LangEnabledBy(C ObjC C++ ObjC++,Wall)
 Warn about boolean expression compared with an integer value different from true/false.
diff --git a/gcc/doc/invoke.texi b/gcc/doc/invoke.texi
index 6070288856c..a22758d18ee 100644
--- a/gcc/doc/invoke.texi
+++ b/gcc/doc/invoke.texi
@@ -327,7 +327,9 @@ Objective-C and Objective-C++ Dialects}.
 -Warith-conversion @gol
 -Warray-bounds  -Warray-bounds=@var{n}  -Warray-compare @gol
 -Wno-attributes  -Wattribute-alias=@var{n} -Wno-attribute-alias @gol
--Wno-attribute-warning  -Wbool-compare  -Wbool-operation @gol
+-Wno-attribute-warning  @gol
+-Wbidi-chars=@r{[}none@r{|}unpaired@r{|}any@r{]} @gol
+-Wbool-compare  -Wbool-operation @gol
 -Wno-builtin-declaration-mismatch @gol
 -Wno-builtin-macro-redefined  -Wc90-c99-compat  -Wc99-c11-compat @gol
 -Wc11-c2x-compat @gol
@@ -7678,6 +7680,23 @@ Attributes considered include @code{alloc_align}, @code{alloc_size},
 This is the default.  You can disable these warnings with either
 @option{-Wno-attribute-alias} or @option{-Wattribute-alias=0}.
 
+@item -Wbidi-chars=@r{[}none@r{|}unpaired@r{|}any@r{]}
+@opindex Wbidi-chars=
+@opindex Wbidi-chars
+@opindex Wno-bidi-chars
+Warn about possibly misleading UTF-8 bidirectional control characters in
+comments, string literals, character constants, and identifiers.  Such
+characters can change left-to-right writing direction into right-to-left
+(and vice versa), which can cause confusion between the logical order and
+visual order.  This may be dangerous; for instance, it may seem that a piece
+of code is not commented out, whereas it in fact is.
+
+There are three levels of warning supported by GCC@.  The default is
+@option{-Wbidi-chars=unpaired}, which warns about improperly terminated
+bidi contexts.  @option{-Wbidi-chars=none} turns the warning off.
+@option{-Wbidi-chars=any} warns about any use of bidirectional control
+characters.
+
 @item -Wbool-compare
 @opindex Wno-bool-compare
 @opindex Wbool-compare
diff --git a/gcc/testsuite/c-c++-common/Wbidi-chars-1.c b/gcc/testsuite/c-c++-common/Wbidi-chars-1.c
new file mode 100644
index 00000000000..34f5ac19271
--- /dev/null
+++ b/gcc/testsuite/c-c++-common/Wbidi-chars-1.c
@@ -0,0 +1,12 @@
+/* PR preprocessor/103026 */
+/* { dg-do compile } */
+
+int main() {
+    int isAdmin = 0;
+    /*‮ } ⁦if (isAdmin)⁩ ⁦ begin admins only */
+/* { dg-warning "bidirectional" "" { target *-*-* } .-1 } */
+        __builtin_printf("You are an admin.\n");
+    /* end admins only ‮ { ⁦*/
+/* { dg-warning "bidirectional" "" { target *-*-* } .-1 } */
+    return 0;
+}
diff --git a/gcc/testsuite/c-c++-common/Wbidi-chars-10.c b/gcc/testsuite/c-c++-common/Wbidi-chars-10.c
new file mode 100644
index 00000000000..3f851b69e65
--- /dev/null
+++ b/gcc/testsuite/c-c++-common/Wbidi-chars-10.c
@@ -0,0 +1,27 @@
+/* PR preprocessor/103026 */
+/* { dg-do compile } */
+/* { dg-options "-Wbidi-chars=unpaired" } */
+/* More nesting testing.  */
+
+/* RLE‫ LRI⁦ PDF‬ PDI⁩*/
+/* { dg-warning "unpaired" "" { target *-*-* } .-1 } */
+int LRE_\u202a_PDF_\u202c;
+int LRE_\u202a_PDF_\u202c_LRE_\u202a_PDF_\u202c;
+int LRE_\u202a_LRI_\u2066_PDF_\u202c_PDI_\u2069;
+/* { dg-warning "unpaired" "" { target *-*-* } .-1 } */
+int RLE_\u202b_RLI_\u2067_PDF_\u202c_PDI_\u2069;
+/* { dg-warning "unpaired" "" { target *-*-* } .-1 } */
+int RLE_\u202b_RLI_\u2067_PDI_\u2069_PDF_\u202c;
+int FSI_\u2068_LRO_\u202d_PDI_\u2069_PDF_\u202c;
+int FSI_\u2068;
+/* { dg-warning "unpaired" "" { target *-*-* } .-1 } */
+int FSI_\u2068_PDI_\u2069;
+int FSI_\u2068_FSI_\u2068_PDI_\u2069;
+/* { dg-warning "unpaired" "" { target *-*-* } .-1 } */
+int RLI_\u2067_RLI_\u2067_RLI_\u2067_RLI_\u2067_RLI_\u2067_RLI_\u2067_RLI_\u2067_PDI_\u2069_PDI_\u2069_PDI_\u2069_PDI_\u2069_PDI_\u2069_PDI_\u2069_PDI_\u2069;
+int RLI_\u2067_RLI_\u2067_RLI_\u2067_RLI_\u2067_RLI_\u2067_RLI_\u2067_RLI_\u2067_PDI_\u2069_PDI_\u2069_PDI_\u2069_PDI_\u2069_PDI_\u2069_PDI_\u2069;
+/* { dg-warning "unpaired" "" { target *-*-* } .-1 } */
+int RLI_\u2067_RLI_\u2067_RLI_\u2067_RLI_\u2067_RLI_\u2067_RLI_\u2067_RLI_\u2067_PDI_\u2069_PDI_\u2069_PDI_\u2069_PDI_\u2069_PDI_\u2069_PDI_\u2069_PDF_\u202c;
+/* { dg-warning "unpaired" "" { target *-*-* } .-1 } */
+int RLI_\u2067_RLI_\u2067_RLI_\u2067_RLI_\u2067_FSI_\u2068_PDI_\u2069_PDI_\u2069_PDI_\u2069_PDI_\u2069;
+/* { dg-warning "unpaired" "" { target *-*-* } .-1 } */
diff --git a/gcc/testsuite/c-c++-common/Wbidi-chars-11.c b/gcc/testsuite/c-c++-common/Wbidi-chars-11.c
new file mode 100644
index 00000000000..270ce2368a9
--- /dev/null
+++ b/gcc/testsuite/c-c++-common/Wbidi-chars-11.c
@@ -0,0 +1,13 @@
+/* PR preprocessor/103026 */
+/* { dg-do compile } */
+/* { dg-options "-Wbidi-chars=unpaired" } */
+/* Test that we warn when mixing UCN and UTF-8.  */
+
+int LRE_‪_PDF_\u202c;
+/* { dg-warning "mismatch" "" { target *-*-* } .-1 } */
+int LRE_\u202a_PDF_‬_;
+/* { dg-warning "mismatch" "" { target *-*-* } .-1 } */
+const char *s1 = "LRE_‪_PDF_\u202c";
+/* { dg-warning "mismatch" "" { target *-*-* } .-1 } */
+const char *s2 = "LRE_\u202a_PDF_‬";
+/* { dg-warning "mismatch" "" { target *-*-* } .-1 } */
diff --git a/gcc/testsuite/c-c++-common/Wbidi-chars-12.c b/gcc/testsuite/c-c++-common/Wbidi-chars-12.c
new file mode 100644
index 00000000000..b07eec1da91
--- /dev/null
+++ b/gcc/testsuite/c-c++-common/Wbidi-chars-12.c
@@ -0,0 +1,19 @@
+/* PR preprocessor/103026 */
+/* { dg-do compile { target { c || c++11 } } } */
+/* { dg-options "-Wbidi-chars=any" } */
+/* Test raw strings.  */
+
+const char *s1 = R"(a b c LRE‪ 1 2 3 PDF‬ x y z)";
+/* { dg-warning "U\\+202A" "" { target *-*-* } .-1 } */
+const char *s2 = R"(a b c RLE‫ 1 2 3 PDF‬ x y z)";
+/* { dg-warning "U\\+202B" "" { target *-*-* } .-1 } */
+const char *s3 = R"(a b c LRO‭ 1 2 3 PDF‬ x y z)";
+/* { dg-warning "U\\+202D" "" { target *-*-* } .-1 } */
+const char *s4 = R"(a b c RLO‮ 1 2 3 PDF‬ x y z)";
+/* { dg-warning "U\\+202E" "" { target *-*-* } .-1 } */
+const char *s7 = R"(a b c FSI⁨ 1 2 3 PDI⁩ x y) z";
+/* { dg-warning "U\\+2068" "" { target *-*-* } .-1 } */
+const char *s8 = R"(a b c PDI⁩ x y )z";
+/* { dg-warning "U\\+2069" "" { target *-*-* } .-1 } */
+const char *s9 = R"(a b c PDF‬ x y z)";
+/* { dg-warning "U\\+202C" "" { target *-*-* } .-1 } */
diff --git a/gcc/testsuite/c-c++-common/Wbidi-chars-13.c b/gcc/testsuite/c-c++-common/Wbidi-chars-13.c
new file mode 100644
index 00000000000..b2dd9fde752
--- /dev/null
+++ b/gcc/testsuite/c-c++-common/Wbidi-chars-13.c
@@ -0,0 +1,17 @@
+/* PR preprocessor/103026 */
+/* { dg-do compile { target { c || c++11 } } } */
+/* { dg-options "-Wbidi-chars=unpaired" } */
+/* Test raw strings.  */
+
+const char *s1 = R"(a b c LRE‪ 1 2 3)";
+/* { dg-warning "unpaired" "" { target *-*-* } .-1 } */
+const char *s2 = R"(a b c RLE‫ 1 2 3)";
+/* { dg-warning "unpaired" "" { target *-*-* } .-1 } */
+const char *s3 = R"(a b c LRO‭ 1 2 3)";
+/* { dg-warning "unpaired" "" { target *-*-* } .-1 } */
+const char *s4 = R"(a b c FSI⁨ 1 2 3)";
+/* { dg-warning "unpaired" "" { target *-*-* } .-1 } */
+const char *s5 = R"(a b c LRI⁦ 1 2 3)";
+/* { dg-warning "unpaired" "" { target *-*-* } .-1 } */
+const char *s6 = R"(a b c RLI⁧ 1 2 3)";
+/* { dg-warning "unpaired" "" { target *-*-* } .-1 } */
diff --git a/gcc/testsuite/c-c++-common/Wbidi-chars-14.c b/gcc/testsuite/c-c++-common/Wbidi-chars-14.c
new file mode 100644
index 00000000000..ba5f75d9553
--- /dev/null
+++ b/gcc/testsuite/c-c++-common/Wbidi-chars-14.c
@@ -0,0 +1,38 @@
+/* PR preprocessor/103026 */
+/* { dg-do compile } */
+/* { dg-options "-Wbidi-chars=unpaired" } */
+/* Test PDI handling, which also pops any subsequent LREs, RLEs, LROs,
+   or RLOs.  */
+
+/* LRI_⁦_LRI_⁦_RLE_‫_RLE_‫_RLE_‫_PDI_⁩*/
+/* { dg-warning "unpaired" "" { target *-*-* } .-1 } */
+// LRI_⁦_RLE_‫_RLE_‫_RLE_‫_PDI_⁩
+// LRI_⁦_RLO_‮_RLE_‫_RLE_‫_PDI_⁩
+// LRI_⁦_RLO_‮_RLE_‫_PDI_⁩
+// FSI_⁨_RLO_‮_PDI_⁩
+// FSI_⁨_FSI_⁨_RLO_‮_PDI_⁩
+/* { dg-warning "unpaired" "" { target *-*-* } .-1 } */
+
+int LRI_\u2066_LRI_\u2066_LRE_\u202a_LRE_\u202a_LRE_\u202a_PDI_\u2069;
+/* { dg-warning "unpaired" "" { target *-*-* } .-1 } */
+int LRI_\u2066_LRI_\u2066_LRE_\u202a_LRE_\u202a_LRE_\u202a_PDI_\u2069_PDI_\u2069;
+int LRI_\u2066_LRI_\u2066_LRI_\u2066_LRE_\u202a_LRE_\u202a_LRE_\u202a_PDI_\u2069_PDI_\u2069;
+/* { dg-warning "unpaired" "" { target *-*-* } .-1 } */
+int PDI_\u2069;
+int LRI_\u2066_PDI_\u2069;
+int RLI_\u2067_PDI_\u2069;
+int LRE_\u202a_LRI_\u2066_PDI_\u2069;
+/* { dg-warning "unpaired" "" { target *-*-* } .-1 } */
+int LRI_\u2066_LRE_\u202a_PDF_\u202c_PDI_\u2069;
+int LRI_\u2066_LRE_\u202a_LRE_\u202a_PDF_\u202c_PDI_\u2069;
+int RLI_\u2067_LRI_\u2066_LRE_\u202a_LRE_\u202a_PDF_\u202c_PDI_\u2069;
+/* { dg-warning "unpaired" "" { target *-*-* } .-1 } */
+int FSI_\u2068_LRI_\u2066_LRE_\u202a_LRE_\u202a_PDF_\u202c_PDI_\u2069;
+/* { dg-warning "unpaired" "" { target *-*-* } .-1 } */
+int RLO_\u202e_PDI_\u2069;
+/* { dg-warning "unpaired" "" { target *-*-* } .-1 } */
+int RLI_\u2067_PDI_\u2069_RLI_\u2067;
+/* { dg-warning "unpaired" "" { target *-*-* } .-1 } */
+int FSI_\u2068_PDF_\u202c_PDI_\u2069;
+int FSI_\u2068_FSI_\u2068_PDF_\u202c_PDI_\u2069;
+/* { dg-warning "unpaired" "" { target *-*-* } .-1 } */
diff --git a/gcc/testsuite/c-c++-common/Wbidi-chars-15.c b/gcc/testsuite/c-c++-common/Wbidi-chars-15.c
new file mode 100644
index 00000000000..68e65001a01
--- /dev/null
+++ b/gcc/testsuite/c-c++-common/Wbidi-chars-15.c
@@ -0,0 +1,33 @@
+/* PR preprocessor/103026 */
+/* { dg-do compile } */
+/* { dg-options "-Wbidi-chars=unpaired" } */
+/* Test unpaired bidi control chars in multiline comments.  */
+
+/*
+ * LRE‪ end
+ */
+/* { dg-warning "unpaired" "" { target *-*-* } .-2 } */
+/*
+ * RLE‫ end
+ */
+/* { dg-warning "unpaired" "" { target *-*-* } .-2 } */
+/*
+ * LRO‭ end
+ */
+/* { dg-warning "unpaired" "" { target *-*-* } .-2 } */
+/*
+ * RLO‮ end
+ */
+/* { dg-warning "unpaired" "" { target *-*-* } .-2 } */
+/*
+ * LRI⁦ end
+ */
+/* { dg-warning "unpaired" "" { target *-*-* } .-2 } */
+/*
+ * RLI⁧ end
+ */
+/* { dg-warning "unpaired" "" { target *-*-* } .-2 } */
+/*
+ * FSI⁨ end
+ */
+/* { dg-warning "unpaired" "" { target *-*-* } .-2 } */
diff --git a/gcc/testsuite/c-c++-common/Wbidi-chars-16.c b/gcc/testsuite/c-c++-common/Wbidi-chars-16.c
new file mode 100644
index 00000000000..baa0159861c
--- /dev/null
+++ b/gcc/testsuite/c-c++-common/Wbidi-chars-16.c
@@ -0,0 +1,26 @@
+/* PR preprocessor/103026 */
+/* { dg-do compile } */
+/* { dg-options "-Wbidi-chars=any" } */
+/* Test LTR/RTL chars.  */
+
+/* LTR<‎> */
+/* { dg-warning "U\\+200E" "" { target *-*-* } .-1 } */
+// LTR<‎>
+/* { dg-warning "U\\+200E" "" { target *-*-* } .-1 } */
+/* RTL<‏> */
+/* { dg-warning "U\\+200F" "" { target *-*-* } .-1 } */
+// RTL<‏>
+/* { dg-warning "U\\+200F" "" { target *-*-* } .-1 } */
+
+const char *s1 = "LTR<‎>";
+/* { dg-warning "U\\+200E" "" { target *-*-* } .-1 } */
+const char *s2 = "LTR\u200e";
+/* { dg-warning "U\\+200E" "" { target *-*-* } .-1 } */
+const char *s3 = "LTR\u200E";
+/* { dg-warning "U\\+200E" "" { target *-*-* } .-1 } */
+const char *s4 = "RTL<‏>";
+/* { dg-warning "U\\+200F" "" { target *-*-* } .-1 } */
+const char *s5 = "RTL\u200f";
+/* { dg-warning "U\\+200F" "" { target *-*-* } .-1 } */
+const char *s6 = "RTL\u200F";
+/* { dg-warning "U\\+200F" "" { target *-*-* } .-1 } */
diff --git a/gcc/testsuite/c-c++-common/Wbidi-chars-17.c b/gcc/testsuite/c-c++-common/Wbidi-chars-17.c
new file mode 100644
index 00000000000..07cb4321f96
--- /dev/null
+++ b/gcc/testsuite/c-c++-common/Wbidi-chars-17.c
@@ -0,0 +1,30 @@
+/* PR preprocessor/103026 */
+/* { dg-do compile } */
+/* { dg-options "-Wbidi-chars=unpaired" } */
+/* Test LTR/RTL chars.  */
+
+/* LTR<‎> */
+/* { dg-bogus "unpaired" "" { target *-*-* } .-1 } */
+// LTR<‎>
+/* { dg-bogus "unpaired" "" { target *-*-* } .-1 } */
+/* RTL<‏> */
+/* { dg-bogus "unpaired" "" { target *-*-* } .-1 } */
+// RTL<‏>
+/* { dg-bogus "unpaired" "" { target *-*-* } .-1 } */
+int ltr_\u200e;
+/* { dg-error "universal character " "" { target *-*-* } .-1 } */
+int rtl_\u200f;
+/* { dg-error "universal character " "" { target *-*-* } .-1 } */
+
+const char *s1 = "LTR<‎>";
+/* { dg-bogus "unpaired" "" { target *-*-* } .-1 } */
+const char *s2 = "LTR\u200e";
+/* { dg-bogus "unpaired" "" { target *-*-* } .-1 } */
+const char *s3 = "LTR\u200E";
+/* { dg-bogus "unpaired" "" { target *-*-* } .-1 } */
+const char *s4 = "RTL<‏>";
+/* { dg-bogus "unpaired" "" { target *-*-* } .-1 } */
+const char *s5 = "RTL\u200f";
+/* { dg-bogus "unpaired" "" { target *-*-* } .-1 } */
+const char *s6 = "RTL\u200F";
+/* { dg-bogus "unpaired" "" { target *-*-* } .-1 } */
diff --git a/gcc/testsuite/c-c++-common/Wbidi-chars-2.c b/gcc/testsuite/c-c++-common/Wbidi-chars-2.c
new file mode 100644
index 00000000000..2340374f276
--- /dev/null
+++ b/gcc/testsuite/c-c++-common/Wbidi-chars-2.c
@@ -0,0 +1,9 @@
+/* PR preprocessor/103026 */
+/* { dg-do compile } */
+
+int main() {
+    /* Say hello; newline⁧/*/ return 0 ;
+/* { dg-warning "bidirectional" "" { target *-*-* } .-1 } */
+    __builtin_printf("Hello world.\n");
+    return 0;
+}
diff --git a/gcc/testsuite/c-c++-common/Wbidi-chars-3.c b/gcc/testsuite/c-c++-common/Wbidi-chars-3.c
new file mode 100644
index 00000000000..9dc7edb6e64
--- /dev/null
+++ b/gcc/testsuite/c-c++-common/Wbidi-chars-3.c
@@ -0,0 +1,11 @@
+/* PR preprocessor/103026 */
+/* { dg-do compile } */
+
+int main() {
+    const char* access_level = "user";
+    if (__builtin_strcmp(access_level, "user‮ ⁦// Check if admin⁩ ⁦")) {
+/* { dg-warning "bidirectional" "" { target *-*-* } .-1 } */
+        __builtin_printf("You are an admin.\n");
+    }
+    return 0;
+}
diff --git a/gcc/testsuite/c-c++-common/Wbidi-chars-4.c b/gcc/testsuite/c-c++-common/Wbidi-chars-4.c
new file mode 100644
index 00000000000..639e5c62e88
--- /dev/null
+++ b/gcc/testsuite/c-c++-common/Wbidi-chars-4.c
@@ -0,0 +1,188 @@
+/* PR preprocessor/103026 */
+/* { dg-do compile } */
+/* { dg-options "-Wbidi-chars=any -Wno-multichar -Wno-overflow" } */
+/* Test all bidi chars in various contexts (identifiers, comments,
+   string literals, character constants), both UCN and UTF-8.  The bidi
+   chars here are properly terminated, except for the character constants.  */
+
+/* a b c LRE‪ 1 2 3 PDF‬ x y z */
+/* { dg-warning "U\\+202A" "" { target *-*-* } .-1 } */
+/* a b c RLE‫ 1 2 3 PDF‬ x y z */
+/* { dg-warning "U\\+202B" "" { target *-*-* } .-1 } */
+/* a b c LRO‭ 1 2 3 PDF‬ x y z */
+/* { dg-warning "U\\+202D" "" { target *-*-* } .-1 } */
+/* a b c RLO‮ 1 2 3 PDF‬ x y z */
+/* { dg-warning "U\\+202E" "" { target *-*-* } .-1 } */
+/* a b c LRI⁦ 1 2 3 PDI⁩ x y z */
+/* { dg-warning "U\\+2066" "" { target *-*-* } .-1 } */
+/* a b c RLI⁧ 1 2 3 PDI⁩ x y */
+/* { dg-warning "U\\+2067" "" { target *-*-* } .-1 } */
+/* a b c FSI⁨ 1 2 3 PDI⁩ x y z */
+/* { dg-warning "U\\+2068" "" { target *-*-* } .-1 } */
+
+/* Same but C++ comments instead.  */
+// a b c LRE‪ 1 2 3 PDF‬ x y z
+/* { dg-warning "U\\+202A" "" { target *-*-* } .-1 } */
+// a b c RLE‫ 1 2 3 PDF‬ x y z
+/* { dg-warning "U\\+202B" "" { target *-*-* } .-1 } */
+// a b c LRO‭ 1 2 3 PDF‬ x y z
+/* { dg-warning "U\\+202D" "" { target *-*-* } .-1 } */
+// a b c RLO‮ 1 2 3 PDF‬ x y z
+/* { dg-warning "U\\+202E" "" { target *-*-* } .-1 } */
+// a b c LRI⁦ 1 2 3 PDI⁩ x y z
+/* { dg-warning "U\\+2066" "" { target *-*-* } .-1 } */
+// a b c RLI⁧ 1 2 3 PDI⁩ x y
+/* { dg-warning "U\\+2067" "" { target *-*-* } .-1 } */
+// a b c FSI⁨ 1 2 3 PDI⁩ x y z
+/* { dg-warning "U\\+2068" "" { target *-*-* } .-1 } */
+
+/* Here we're closing an unopened context, warn when =any.  */
+/* a b c PDI⁩ x y z */
+/* { dg-warning "U\\+2069" "" { target *-*-* } .-1 } */
+/* a b c PDF‬ x y z */
+/* { dg-warning "U\\+202C" "" { target *-*-* } .-1 } */
+// a b c PDI⁩ x y z
+/* { dg-warning "U\\+2069" "" { target *-*-* } .-1 } */
+// a b c PDF‬ x y z
+/* { dg-warning "U\\+202C" "" { target *-*-* } .-1 } */
+
+/* Multiline comments.  */
+/* a b c PDI⁩ x y z
+   */
+/* { dg-warning "U\\+2069" "" { target *-*-* } .-2 } */
+/* a b c PDF‬ x y z
+   */
+/* { dg-warning "U\\+202C" "" { target *-*-* } .-2 } */
+/* first
+   a b c PDI⁩ x y z
+   */
+/* { dg-warning "U\\+2069" "" { target *-*-* } .-2 } */
+/* first
+   a b c PDF‬ x y z
+   */
+/* { dg-warning "U\\+202C" "" { target *-*-* } .-2 } */
+/* first
+   a b c PDI⁩ x y z */
+/* { dg-warning "U\\+2069" "" { target *-*-* } .-1 } */
+/* first
+   a b c PDF‬ x y z */
+/* { dg-warning "U\\+202C" "" { target *-*-* } .-1 } */
+
+void
+g1 ()
+{
+  const char *s1 = "a b c LRE‪ 1 2 3 PDF‬ x y z";
+/* { dg-warning "U\\+202A" "" { target *-*-* } .-1 } */
+  const char *s2 = "a b c RLE‫ 1 2 3 PDF‬ x y z";
+/* { dg-warning "U\\+202B" "" { target *-*-* } .-1 } */
+  const char *s3 = "a b c LRO‭ 1 2 3 PDF‬ x y z";
+/* { dg-warning "U\\+202D" "" { target *-*-* } .-1 } */
+  const char *s4 = "a b c RLO‮ 1 2 3 PDF‬ x y z";
+/* { dg-warning "U\\+202E" "" { target *-*-* } .-1 } */
+  const char *s5 = "a b c LRI⁦ 1 2 3 PDI⁩ x y z";
+/* { dg-warning "U\\+2066" "" { target *-*-* } .-1 } */
+  const char *s6 = "a b c RLI⁧ 1 2 3 PDI⁩ x y z";
+/* { dg-warning "U\\+2067" "" { target *-*-* } .-1 } */
+  const char *s7 = "a b c FSI⁨ 1 2 3 PDI⁩ x y z";
+/* { dg-warning "U\\+2068" "" { target *-*-* } .-1 } */
+  const char *s8 = "a b c PDI⁩ x y z";
+/* { dg-warning "U\\+2069" "" { target *-*-* } .-1 } */
+  const char *s9 = "a b c PDF‬ x y z";
+/* { dg-warning "U\\+202C" "" { target *-*-* } .-1 } */
+
+  const char *s10 = "a b c LRE\u202a 1 2 3 PDF\u202c x y z";
+/* { dg-warning "U\\+202A" "" { target *-*-* } .-1 } */
+  const char *s11 = "a b c LRE\u202A 1 2 3 PDF\u202c x y z";
+/* { dg-warning "U\\+202A" "" { target *-*-* } .-1 } */
+  const char *s12 = "a b c RLE\u202b 1 2 3 PDF\u202c x y z";
+/* { dg-warning "U\\+202B" "" { target *-*-* } .-1 } */
+  const char *s13 = "a b c RLE\u202B 1 2 3 PDF\u202c x y z";
+/* { dg-warning "U\\+202B" "" { target *-*-* } .-1 } */
+  const char *s14 = "a b c LRO\u202d 1 2 3 PDF\u202c x y z";
+/* { dg-warning "U\\+202D" "" { target *-*-* } .-1 } */
+  const char *s15 = "a b c LRO\u202D 1 2 3 PDF\u202c x y z";
+/* { dg-warning "U\\+202D" "" { target *-*-* } .-1 } */
+  const char *s16 = "a b c RLO\u202e 1 2 3 PDF\u202c x y z";
+/* { dg-warning "U\\+202E" "" { target *-*-* } .-1 } */
+  const char *s17 = "a b c RLO\u202E 1 2 3 PDF\u202c x y z";
+/* { dg-warning "U\\+202E" "" { target *-*-* } .-1 } */
+  const char *s18 = "a b c LRI\u2066 1 2 3 PDI\u2069 x y z";
+/* { dg-warning "U\\+2066" "" { target *-*-* } .-1 } */
+  const char *s19 = "a b c RLI\u2067 1 2 3 PDI\u2069 x y z";
+/* { dg-warning "U\\+2067" "" { target *-*-* } .-1 } */
+  const char *s20 = "a b c FSI\u2068 1 2 3 PDI\u2069 x y z";
+/* { dg-warning "U\\+2068" "" { target *-*-* } .-1 } */
+}
+
+void
+g2 ()
+{
+  const char c1 = '\u202a';
+/* { dg-warning "U\\+202A" "" { target *-*-* } .-1 } */
+  const char c2 = '\u202A';
+/* { dg-warning "U\\+202A" "" { target *-*-* } .-1 } */
+  const char c3 = '\u202b';
+/* { dg-warning "U\\+202B" "" { target *-*-* } .-1 } */
+  const char c4 = '\u202B';
+/* { dg-warning "U\\+202B" "" { target *-*-* } .-1 } */
+  const char c5 = '\u202d';
+/* { dg-warning "U\\+202D" "" { target *-*-* } .-1 } */
+  const char c6 = '\u202D';
+/* { dg-warning "U\\+202D" "" { target *-*-* } .-1 } */
+  const char c7 = '\u202e';
+/* { dg-warning "U\\+202E" "" { target *-*-* } .-1 } */
+  const char c8 = '\u202E';
+/* { dg-warning "U\\+202E" "" { target *-*-* } .-1 } */
+  const char c9 = '\u2066';
+/* { dg-warning "U\\+2066" "" { target *-*-* } .-1 } */
+  const char c10 = '\u2067';
+/* { dg-warning "U\\+2067" "" { target *-*-* } .-1 } */
+  const char c11 = '\u2068';
+/* { dg-warning "U\\+2068" "" { target *-*-* } .-1 } */
+}
+
+int a‪b‬c;
+/* { dg-warning "U\\+202A" "" { target *-*-* } .-1 } */
+int a‫b‬c;
+/* { dg-warning "U\\+202B" "" { target *-*-* } .-1 } */
+int a‭b‬c;
+/* { dg-warning "U\\+202D" "" { target *-*-* } .-1 } */
+int a‮b‬c;
+/* { dg-warning "U\\+202E" "" { target *-*-* } .-1 } */
+int a⁦b⁩c;
+/* { dg-warning "U\\+2066" "" { target *-*-* } .-1 } */
+int a⁧b⁩c;
+/* { dg-warning "U\\+2067" "" { target *-*-* } .-1 } */
+int a⁨b⁩c;
+/* { dg-warning "U\\+2068" "" { target *-*-* } .-1 } */
+int A‬X;
+/* { dg-warning "U\\+202C" "" { target *-*-* } .-1 } */
+int A\u202cY;
+/* { dg-warning "U\\+202C" "" { target *-*-* } .-1 } */
+int A\u202CY2;
+/* { dg-warning "U\\+202C" "" { target *-*-* } .-1 } */
+
+int d\u202ae\u202cf;
+/* { dg-warning "U\\+202A" "" { target *-*-* } .-1 } */
+int d\u202Ae\u202cf2;
+/* { dg-warning "U\\+202A" "" { target *-*-* } .-1 } */
+int d\u202be\u202cf;
+/* { dg-warning "U\\+202B" "" { target *-*-* } .-1 } */
+int d\u202Be\u202cf2;
+/* { dg-warning "U\\+202B" "" { target *-*-* } .-1 } */
+int d\u202de\u202cf;
+/* { dg-warning "U\\+202D" "" { target *-*-* } .-1 } */
+int d\u202De\u202cf2;
+/* { dg-warning "U\\+202D" "" { target *-*-* } .-1 } */
+int d\u202ee\u202cf;
+/* { dg-warning "U\\+202E" "" { target *-*-* } .-1 } */
+int d\u202Ee\u202cf2;
+/* { dg-warning "U\\+202E" "" { target *-*-* } .-1 } */
+int d\u2066e\u2069f;
+/* { dg-warning "U\\+2066" "" { target *-*-* } .-1 } */
+int d\u2067e\u2069f;
+/* { dg-warning "U\\+2067" "" { target *-*-* } .-1 } */
+int d\u2068e\u2069f;
+/* { dg-warning "U\\+2068" "" { target *-*-* } .-1 } */
+int X\u2069;
+/* { dg-warning "U\\+2069" "" { target *-*-* } .-1 } */
diff --git a/gcc/testsuite/c-c++-common/Wbidi-chars-5.c b/gcc/testsuite/c-c++-common/Wbidi-chars-5.c
new file mode 100644
index 00000000000..68cb053144b
--- /dev/null
+++ b/gcc/testsuite/c-c++-common/Wbidi-chars-5.c
@@ -0,0 +1,188 @@
+/* PR preprocessor/103026 */
+/* { dg-do compile } */
+/* { dg-options "-Wbidi-chars=unpaired -Wno-multichar -Wno-overflow" } */
+/* Test all bidi chars in various contexts (identifiers, comments,
+   string literals, character constants), both UCN and UTF-8.  The bidi
+   chars here are properly terminated, except for the character constants.  */
+
+/* a b c LRE‪ 1 2 3 PDF‬ x y z */
+/* { dg-bogus "unpaired" "" { target *-*-* } .-1 } */
+/* a b c RLE‫ 1 2 3 PDF‬ x y z */
+/* { dg-bogus "unpaired" "" { target *-*-* } .-1 } */
+/* a b c LRO‭ 1 2 3 PDF‬ x y z */
+/* { dg-bogus "unpaired" "" { target *-*-* } .-1 } */
+/* a b c RLO‮ 1 2 3 PDF‬ x y z */
+/* { dg-bogus "unpaired" "" { target *-*-* } .-1 } */
+/* a b c LRI⁦ 1 2 3 PDI⁩ x y z */
+/* { dg-bogus "unpaired" "" { target *-*-* } .-1 } */
+/* a b c RLI⁧ 1 2 3 PDI⁩ x y */
+/* { dg-bogus "unpaired" "" { target *-*-* } .-1 } */
+/* a b c FSI⁨ 1 2 3 PDI⁩ x y z */
+/* { dg-bogus "unpaired" "" { target *-*-* } .-1 } */
+
+/* Same but C++ comments instead.  */
+// a b c LRE‪ 1 2 3 PDF‬ x y z
+/* { dg-bogus "unpaired" "" { target *-*-* } .-1 } */
+// a b c RLE‫ 1 2 3 PDF‬ x y z
+/* { dg-bogus "unpaired" "" { target *-*-* } .-1 } */
+// a b c LRO‭ 1 2 3 PDF‬ x y z
+/* { dg-bogus "unpaired" "" { target *-*-* } .-1 } */
+// a b c RLO‮ 1 2 3 PDF‬ x y z
+/* { dg-bogus "unpaired" "" { target *-*-* } .-1 } */
+// a b c LRI⁦ 1 2 3 PDI⁩ x y z
+/* { dg-bogus "unpaired" "" { target *-*-* } .-1 } */
+// a b c RLI⁧ 1 2 3 PDI⁩ x y
+/* { dg-bogus "unpaired" "" { target *-*-* } .-1 } */
+// a b c FSI⁨ 1 2 3 PDI⁩ x y z
+/* { dg-bogus "unpaired" "" { target *-*-* } .-1 } */
+
+/* Here we're closing an unopened context, warn when =any.  */
+/* a b c PDI⁩ x y z */
+/* { dg-bogus "unpaired" "" { target *-*-* } .-1 } */
+/* a b c PDF‬ x y z */
+/* { dg-bogus "unpaired" "" { target *-*-* } .-1 } */
+// a b c PDI⁩ x y z
+/* { dg-bogus "unpaired" "" { target *-*-* } .-1 } */
+// a b c PDF‬ x y z
+/* { dg-bogus "unpaired" "" { target *-*-* } .-1 } */
+
+/* Multiline comments.  */
+/* a b c PDI⁩ x y z
+   */
+/* { dg-bogus "unpaired" "" { target *-*-* } .-2 } */
+/* a b c PDF‬ x y z
+   */
+/* { dg-bogus "unpaired" "" { target *-*-* } .-2 } */
+/* first
+   a b c PDI⁩ x y z
+   */
+/* { dg-bogus "unpaired" "" { target *-*-* } .-2 } */
+/* first
+   a b c PDF‬ x y z
+   */
+/* { dg-bogus "unpaired" "" { target *-*-* } .-2 } */
+/* first
+   a b c PDI⁩ x y z */
+/* { dg-bogus "unpaired" "" { target *-*-* } .-1 } */
+/* first
+   a b c PDF‬ x y z */
+/* { dg-bogus "unpaired" "" { target *-*-* } .-1 } */
+
+void
+g1 ()
+{
+  const char *s1 = "a b c LRE‪ 1 2 3 PDF‬ x y z";
+/* { dg-bogus "unpaired" "" { target *-*-* } .-1 } */
+  const char *s2 = "a b c RLE‫ 1 2 3 PDF‬ x y z";
+/* { dg-bogus "unpaired" "" { target *-*-* } .-1 } */
+  const char *s3 = "a b c LRO‭ 1 2 3 PDF‬ x y z";
+/* { dg-bogus "unpaired" "" { target *-*-* } .-1 } */
+  const char *s4 = "a b c RLO‮ 1 2 3 PDF‬ x y z";
+/* { dg-bogus "unpaired" "" { target *-*-* } .-1 } */
+  const char *s5 = "a b c LRI⁦ 1 2 3 PDI⁩ x y z";
+/* { dg-bogus "unpaired" "" { target *-*-* } .-1 } */
+  const char *s6 = "a b c RLI⁧ 1 2 3 PDI⁩ x y z";
+/* { dg-bogus "unpaired" "" { target *-*-* } .-1 } */
+  const char *s7 = "a b c FSI⁨ 1 2 3 PDI⁩ x y z";
+/* { dg-bogus "unpaired" "" { target *-*-* } .-1 } */
+  const char *s8 = "a b c PDI⁩ x y z";
+/* { dg-bogus "unpaired" "" { target *-*-* } .-1 } */
+  const char *s9 = "a b c PDF‬ x y z";
+/* { dg-bogus "unpaired" "" { target *-*-* } .-1 } */
+
+  const char *s10 = "a b c LRE\u202a 1 2 3 PDF\u202c x y z";
+/* { dg-bogus "unpaired" "" { target *-*-* } .-1 } */
+  const char *s11 = "a b c LRE\u202A 1 2 3 PDF\u202c x y z";
+/* { dg-bogus "unpaired" "" { target *-*-* } .-1 } */
+  const char *s12 = "a b c RLE\u202b 1 2 3 PDF\u202c x y z";
+/* { dg-bogus "unpaired" "" { target *-*-* } .-1 } */
+  const char *s13 = "a b c RLE\u202B 1 2 3 PDF\u202c x y z";
+/* { dg-bogus "unpaired" "" { target *-*-* } .-1 } */
+  const char *s14 = "a b c LRO\u202d 1 2 3 PDF\u202c x y z";
+/* { dg-bogus "unpaired" "" { target *-*-* } .-1 } */
+  const char *s15 = "a b c LRO\u202D 1 2 3 PDF\u202c x y z";
+/* { dg-bogus "unpaired" "" { target *-*-* } .-1 } */
+  const char *s16 = "a b c RLO\u202e 1 2 3 PDF\u202c x y z";
+/* { dg-bogus "unpaired" "" { target *-*-* } .-1 } */
+  const char *s17 = "a b c RLO\u202E 1 2 3 PDF\u202c x y z";
+/* { dg-bogus "unpaired" "" { target *-*-* } .-1 } */
+  const char *s18 = "a b c LRI\u2066 1 2 3 PDI\u2069 x y z";
+/* { dg-bogus "unpaired" "" { target *-*-* } .-1 } */
+  const char *s19 = "a b c RLI\u2067 1 2 3 PDI\u2069 x y z";
+/* { dg-bogus "unpaired" "" { target *-*-* } .-1 } */
+  const char *s20 = "a b c FSI\u2068 1 2 3 PDI\u2069 x y z";
+/* { dg-bogus "unpaired" "" { target *-*-* } .-1 } */
+}
+
+void
+g2 ()
+{
+  const char c1 = '\u202a';
+/* { dg-warning "unpaired" "" { target *-*-* } .-1 } */
+  const char c2 = '\u202A';
+/* { dg-warning "unpaired" "" { target *-*-* } .-1 } */
+  const char c3 = '\u202b';
+/* { dg-warning "unpaired" "" { target *-*-* } .-1 } */
+  const char c4 = '\u202B';
+/* { dg-warning "unpaired" "" { target *-*-* } .-1 } */
+  const char c5 = '\u202d';
+/* { dg-warning "unpaired" "" { target *-*-* } .-1 } */
+  const char c6 = '\u202D';
+/* { dg-warning "unpaired" "" { target *-*-* } .-1 } */
+  const char c7 = '\u202e';
+/* { dg-warning "unpaired" "" { target *-*-* } .-1 } */
+  const char c8 = '\u202E';
+/* { dg-warning "unpaired" "" { target *-*-* } .-1 } */
+  const char c9 = '\u2066';
+/* { dg-warning "unpaired" "" { target *-*-* } .-1 } */
+  const char c10 = '\u2067';
+/* { dg-warning "unpaired" "" { target *-*-* } .-1 } */
+  const char c11 = '\u2068';
+/* { dg-warning "unpaired" "" { target *-*-* } .-1 } */
+}
+
+int a‪b‬c;
+/* { dg-bogus "unpaired" "" { target *-*-* } .-1 } */
+int a‫b‬c;
+/* { dg-bogus "unpaired" "" { target *-*-* } .-1 } */
+int a‭b‬c;
+/* { dg-bogus "unpaired" "" { target *-*-* } .-1 } */
+int a‮b‬c;
+/* { dg-bogus "unpaired" "" { target *-*-* } .-1 } */
+int a⁦b⁩c;
+/* { dg-bogus "unpaired" "" { target *-*-* } .-1 } */
+int a⁧b⁩c;
+/* { dg-bogus "unpaired" "" { target *-*-* } .-1 } */
+int a⁨b⁩c;
+/* { dg-bogus "unpaired" "" { target *-*-* } .-1 } */
+int A‬X;
+/* { dg-bogus "unpaired" "" { target *-*-* } .-1 } */
+int A\u202cY;
+/* { dg-bogus "unpaired" "" { target *-*-* } .-1 } */
+int A\u202CY2;
+/* { dg-bogus "unpaired" "" { target *-*-* } .-1 } */
+
+int d\u202ae\u202cf;
+/* { dg-bogus "unpaired" "" { target *-*-* } .-1 } */
+int d\u202Ae\u202cf2;
+/* { dg-bogus "unpaired" "" { target *-*-* } .-1 } */
+int d\u202be\u202cf;
+/* { dg-bogus "unpaired" "" { target *-*-* } .-1 } */
+int d\u202Be\u202cf2;
+/* { dg-bogus "unpaired" "" { target *-*-* } .-1 } */
+int d\u202de\u202cf;
+/* { dg-bogus "unpaired" "" { target *-*-* } .-1 } */
+int d\u202De\u202cf2;
+/* { dg-bogus "unpaired" "" { target *-*-* } .-1 } */
+int d\u202ee\u202cf;
+/* { dg-bogus "unpaired" "" { target *-*-* } .-1 } */
+int d\u202Ee\u202cf2;
+/* { dg-bogus "unpaired" "" { target *-*-* } .-1 } */
+int d\u2066e\u2069f;
+/* { dg-bogus "unpaired" "" { target *-*-* } .-1 } */
+int d\u2067e\u2069f;
+/* { dg-bogus "unpaired" "" { target *-*-* } .-1 } */
+int d\u2068e\u2069f;
+/* { dg-bogus "unpaired" "" { target *-*-* } .-1 } */
+int X\u2069;
+/* { dg-bogus "unpaired" "" { target *-*-* } .-1 } */
diff --git a/gcc/testsuite/c-c++-common/Wbidi-chars-6.c b/gcc/testsuite/c-c++-common/Wbidi-chars-6.c
new file mode 100644
index 00000000000..0ce6fff2dee
--- /dev/null
+++ b/gcc/testsuite/c-c++-common/Wbidi-chars-6.c
@@ -0,0 +1,155 @@
+/* PR preprocessor/103026 */
+/* { dg-do compile } */
+/* { dg-options "-Wbidi-chars=unpaired" } */
+/* Test nesting of bidi chars in various contexts.  */
+
+/* Terminated by the wrong char:  */
+/* a b c LRE‪ 1 2 3 PDI⁩ x y z */
+/* { dg-warning "unpaired" "" { target *-*-* } .-1 } */
+/* a b c RLE‫ 1 2 3 PDI⁩ x y  z*/
+/* { dg-warning "unpaired" "" { target *-*-* } .-1 } */
+/* a b c LRO‭ 1 2 3 PDI⁩ x y z */
+/* { dg-warning "unpaired" "" { target *-*-* } .-1 } */
+/* a b c RLO‮ 1 2 3 PDI⁩ x y z */
+/* { dg-warning "unpaired" "" { target *-*-* } .-1 } */
+/* a b c LRI⁦ 1 2 3 PDF‬ x y z */
+/* { dg-warning "unpaired" "" { target *-*-* } .-1 } */
+/* a b c RLI⁧ 1 2 3 PDF‬ x y z */
+/* { dg-warning "unpaired" "" { target *-*-* } .-1 } */
+/* a b c FSI⁨ 1 2 3 PDF‬ x y  z*/
+/* { dg-warning "unpaired" "" { target *-*-* } .-1 } */
+
+/* LRE‪ PDF‬ */
+/* LRE‪ LRE‪ PDF‬ PDF‬ */
+/* PDF‬ LRE‪ PDF‬ */
+/* LRE‪ PDF‬ LRE‪ PDF‬ */
+/* LRE‪ LRE‪ PDF‬ */
+/* { dg-warning "unpaired" "" { target *-*-* } .-1 } */
+/* PDF‬ LRE‪ */
+/* { dg-warning "unpaired" "" { target *-*-* } .-1 } */
+
+// a b c LRE‪ 1 2 3 PDI⁩ x y z
+/* { dg-warning "unpaired" "" { target *-*-* } .-1 } */
+// a b c RLE‫ 1 2 3 PDI⁩ x y  z*/
+/* { dg-warning "unpaired" "" { target *-*-* } .-1 } */
+// a b c LRO‭ 1 2 3 PDI⁩ x y z 
+/* { dg-warning "unpaired" "" { target *-*-* } .-1 } */
+// a b c RLO‮ 1 2 3 PDI⁩ x y z 
+/* { dg-warning "unpaired" "" { target *-*-* } .-1 } */
+// a b c LRI⁦ 1 2 3 PDF‬ x y z 
+/* { dg-warning "unpaired" "" { target *-*-* } .-1 } */
+// a b c RLI⁧ 1 2 3 PDF‬ x y z 
+/* { dg-warning "unpaired" "" { target *-*-* } .-1 } */
+// a b c FSI⁨ 1 2 3 PDF‬ x y  z
+/* { dg-warning "unpaired" "" { target *-*-* } .-1 } */
+
+// LRE‪ PDF‬ 
+// LRE‪ LRE‪ PDF‬ PDF‬
+// PDF‬ LRE‪ PDF‬
+// LRE‪ PDF‬ LRE‪ PDF‬
+// LRE‪ LRE‪ PDF‬
+/* { dg-warning "unpaired" "" { target *-*-* } .-1 } */
+// PDF‬ LRE‪
+/* { dg-warning "unpaired" "" { target *-*-* } .-1 } */
+
+void
+g1 ()
+{
+  const char *s1 = "a b c LRE‪ 1 2 3 PDI⁩ x y z";
+/* { dg-warning "unpaired" "" { target *-*-* } .-1 } */
+  const char *s2 = "a b c LRE\u202a 1 2 3 PDI\u2069 x y z";
+/* { dg-warning "unpaired" "" { target *-*-* } .-1 } */
+  const char *s3 = "a b c RLE‫ 1 2 3 PDI⁩ x y ";
+/* { dg-warning "unpaired" "" { target *-*-* } .-1 } */
+  const char *s4 = "a b c RLE\u202b 1 2 3 PDI\u2069 x y z";
+/* { dg-warning "unpaired" "" { target *-*-* } .-1 } */
+  const char *s5 = "a b c LRO‭ 1 2 3 PDI⁩ x y z";
+/* { dg-warning "unpaired" "" { target *-*-* } .-1 } */
+  const char *s6 = "a b c LRO\u202d 1 2 3 PDI\u2069 x y z";
+/* { dg-warning "unpaired" "" { target *-*-* } .-1 } */
+  const char *s7 = "a b c RLO‮ 1 2 3 PDI⁩ x y z";
+/* { dg-warning "unpaired" "" { target *-*-* } .-1 } */
+  const char *s8 = "a b c RLO\u202e 1 2 3 PDI\u2069 x y z";
+/* { dg-warning "unpaired" "" { target *-*-* } .-1 } */
+  const char *s9 = "a b c LRI⁦ 1 2 3 PDF‬ x y z";
+/* { dg-warning "unpaired" "" { target *-*-* } .-1 } */
+  const char *s10 = "a b c LRI\u2066 1 2 3 PDF\u202c x y z";
+/* { dg-warning "unpaired" "" { target *-*-* } .-1 } */
+  const char *s11 = "a b c RLI⁧ 1 2 3 PDF‬ x y z\
+    ";
+/* { dg-warning "unpaired" "" { target *-*-* } .-2 } */
+  const char *s12 = "a b c RLI\u2067 1 2 3 PDF\u202c x y z";
+/* { dg-warning "unpaired" "" { target *-*-* } .-1 } */
+  const char *s13 = "a b c FSI⁨ 1 2 3 PDF‬ x y z";
+/* { dg-warning "unpaired" "" { target *-*-* } .-1 } */
+  const char *s14 = "a b c FSI\u2068 1 2 3 PDF\u202c x y z";
+/* { dg-warning "unpaired" "" { target *-*-* } .-1 } */
+  const char *s15 = "PDF‬ LRE‪";
+/* { dg-warning "unpaired" "" { target *-*-* } .-1 } */
+  const char *s16 = "PDF\u202c LRE\u202a";
+/* { dg-warning "unpaired" "" { target *-*-* } .-1 } */
+  const char *s17 = "LRE‪ PDF‬";
+  const char *s18 = "LRE\u202a PDF\u202c";
+  const char *s19 = "LRE‪ LRE‪ PDF‬ PDF‬";
+  const char *s20 = "LRE\u202a LRE\u202a PDF\u202c PDF\u202c";
+  const char *s21 = "PDF‬ LRE‪ PDF‬";
+  const char *s22 = "PDF\u202c LRE\u202a PDF\u202c";
+  const char *s23 = "LRE‪ LRE‪ PDF‬";
+/* { dg-warning "unpaired" "" { target *-*-* } .-1 } */
+  const char *s24 = "LRE\u202a LRE\u202a PDF\u202c";
+/* { dg-warning "unpaired" "" { target *-*-* } .-1 } */
+  const char *s25 = "PDF‬ LRE‪";
+/* { dg-warning "unpaired" "" { target *-*-* } .-1 } */
+  const char *s26 = "PDF\u202c LRE\u202a";
+/* { dg-warning "unpaired" "" { target *-*-* } .-1 } */
+  const char *s27 = "PDF‬ LRE\u202a";
+/* { dg-warning "unpaired" "" { target *-*-* } .-1 } */
+  const char *s28 = "PDF\u202c LRE‪";
+/* { dg-warning "unpaired" "" { target *-*-* } .-1 } */
+}
+
+int aLRE‪bPDI⁩;
+/* { dg-warning "unpaired" "" { target *-*-* } .-1 } */
+int A\u202aB\u2069C;
+/* { dg-warning "unpaired" "" { target *-*-* } .-1 } */
+int aRLE‫bPDI⁩;
+/* { dg-warning "unpaired" "" { target *-*-* } .-1 } */
+int a\u202bB\u2069c;
+/* { dg-warning "unpaired" "" { target *-*-* } .-1 } */
+int aLRO‭bPDI⁩;
+/* { dg-warning "unpaired" "" { target *-*-* } .-1 } */
+int a\u202db\u2069c2;
+/* { dg-warning "unpaired" "" { target *-*-* } .-1 } */
+int aRLO‮bPDI⁩;
+/* { dg-warning "unpaired" "" { target *-*-* } .-1 } */
+int a\u202eb\u2069;
+/* { dg-warning "unpaired" "" { target *-*-* } .-1 } */
+int aLRI⁦bPDF‬;
+/* { dg-warning "unpaired" "" { target *-*-* } .-1 } */
+int a\u2066b\u202c;
+/* { dg-warning "unpaired" "" { target *-*-* } .-1 } */
+int aRLI⁧bPDF‬c
+;
+/* { dg-warning "unpaired" "" { target *-*-* } .-2 } */
+int a\u2067b\u202c;
+/* { dg-warning "unpaired" "" { target *-*-* } .-1 } */
+int aFSI⁨bPDF‬;
+/* { dg-warning "unpaired" "" { target *-*-* } .-1 } */
+int a\u2068b\u202c;
+/* { dg-warning "unpaired" "" { target *-*-* } .-1 } */
+int aFSI⁨bPD\u202C;
+/* { dg-warning "unpaired" "" { target *-*-* } .-1 } */
+int aFSI\u2068bPDF‬_;
+/* { dg-warning "unpaired" "" { target *-*-* } .-1 } */
+int aLRE‪bPDF‬b; 
+int A\u202aB\u202c;
+int a_LRE‪_LRE‪_b_PDF‬_PDF‬;
+int A\u202aA\u202aB\u202cB\u202c;
+int aPDF‬bLREadPDF‬;
+int a_\u202C_\u202a_\u202c;
+int a_LRE‪_b_PDF‬_c_LRE‪_PDF‬;
+int a_\u202a_\u202c_\u202a_\u202c_;
+int a_LRE‪_b_PDF‬_c_LRE‪;
+/* { dg-warning "unpaired" "" { target *-*-* } .-1 } */
+int a_\u202a_\u202c_\u202a_;
+/* { dg-warning "unpaired" "" { target *-*-* } .-1 } */
diff --git a/gcc/testsuite/c-c++-common/Wbidi-chars-7.c b/gcc/testsuite/c-c++-common/Wbidi-chars-7.c
new file mode 100644
index 00000000000..d012d420ec0
--- /dev/null
+++ b/gcc/testsuite/c-c++-common/Wbidi-chars-7.c
@@ -0,0 +1,9 @@
+/* PR preprocessor/103026 */
+/* { dg-do compile } */
+/* { dg-options "-Wbidi-chars=any" } */
+/* Test we ignore UCNs in comments.  */
+
+// a b c \u202a 1 2 3
+// a b c \u202A 1 2 3
+/* a b c \u202a 1 2 3 */
+/* a b c \u202A 1 2 3 */
diff --git a/gcc/testsuite/c-c++-common/Wbidi-chars-8.c b/gcc/testsuite/c-c++-common/Wbidi-chars-8.c
new file mode 100644
index 00000000000..4f54c5092ec
--- /dev/null
+++ b/gcc/testsuite/c-c++-common/Wbidi-chars-8.c
@@ -0,0 +1,13 @@
+/* PR preprocessor/103026 */
+/* { dg-do compile } */
+/* { dg-options "-Wbidi-chars=any" } */
+/* Test \u vs \U.  */
+
+int a_\u202A;
+/* { dg-warning "U\\+202A" "" { target *-*-* } .-1 } */
+int a_\u202a_2;
+/* { dg-warning "U\\+202A" "" { target *-*-* } .-1 } */
+int a_\U0000202A_3;
+/* { dg-warning "U\\+202A" "" { target *-*-* } .-1 } */
+int a_\U0000202a_4;
+/* { dg-warning "U\\+202A" "" { target *-*-* } .-1 } */
diff --git a/gcc/testsuite/c-c++-common/Wbidi-chars-9.c b/gcc/testsuite/c-c++-common/Wbidi-chars-9.c
new file mode 100644
index 00000000000..e2af1b1ca97
--- /dev/null
+++ b/gcc/testsuite/c-c++-common/Wbidi-chars-9.c
@@ -0,0 +1,29 @@
+/* PR preprocessor/103026 */
+/* { dg-do compile } */
+/* { dg-options "-Wbidi-chars=unpaired" } */
+/* Test that we properly separate bidi contexts (comment/identifier/character
+   constant/string literal).  */
+
+/* LRE ->‪<- */ int pdf_\u202c_1;
+/* { dg-warning "unpaired" "" { target *-*-* } .-1 } */
+/* RLE ->‫<- */ int pdf_\u202c_2;
+/* { dg-warning "unpaired" "" { target *-*-* } .-1 } */
+/* LRO ->‭<- */ int pdf_\u202c_3;
+/* { dg-warning "unpaired" "" { target *-*-* } .-1 } */
+/* RLO ->‮<- */ int pdf_\u202c_4;
+/* { dg-warning "unpaired" "" { target *-*-* } .-1 } */
+/* LRI ->⁦<-*/ int pdi_\u2069_1;
+/* { dg-warning "unpaired" "" { target *-*-* } .-1 } */
+/* RLI ->⁧<- */ int pdi_\u2069_12;
+/* { dg-warning "unpaired" "" { target *-*-* } .-1 } */
+/* FSI ->⁨<- */ int pdi_\u2069_3;
+/* { dg-warning "unpaired" "" { target *-*-* } .-1 } */
+
+const char *s1 = "LRE\u202a"; /* PDF ->‬<- */
+/* { dg-warning "unpaired" "" { target *-*-* } .-1 } */
+/* LRE ->‪<- */ const char *s2 = "PDF\u202c";
+/* { dg-warning "unpaired" "" { target *-*-* } .-1 } */
+const char *s3 = "LRE\u202a"; int pdf_\u202c_5;
+/* { dg-warning "unpaired" "" { target *-*-* } .-1 } */
+int lre_\u202a; const char *s4 = "PDF\u202c";
+/* { dg-warning "unpaired" "" { target *-*-* } .-1 } */
diff --git a/libcpp/include/cpplib.h b/libcpp/include/cpplib.h
index 176f8c5bbce..112b9c24751 100644
--- a/libcpp/include/cpplib.h
+++ b/libcpp/include/cpplib.h
@@ -319,6 +319,17 @@ enum cpp_main_search
   CMS_system,  /* Search the system INCLUDE path.  */
 };
 
+/* The possible bidirectional control characters checking levels, from least
+   restrictive to most.  */
+enum cpp_bidirectional_level {
+  /* No checking.  */
+  bidirectional_none,
+  /* Only detect unpaired uses of bidirectional control characters.  */
+  bidirectional_unpaired,
+  /* Detect any use of bidirectional control characters.  */
+  bidirectional_any
+};
+
 /* This structure is nested inside struct cpp_reader, and
    carries all the options visible to the command line.  */
 struct cpp_options
@@ -539,6 +550,10 @@ struct cpp_options
   /* True if warn about differences between C++98 and C++11.  */
   bool cpp_warn_cxx11_compat;
 
+  /* Nonzero if bidirectional control characters checking is on.  See enum
+     cpp_bidirectional_level.  */
+  unsigned char cpp_warn_bidirectional;
+
   /* Dependency generation.  */
   struct
   {
@@ -643,7 +658,8 @@ enum cpp_warning_reason {
   CPP_W_C90_C99_COMPAT,
   CPP_W_C11_C2X_COMPAT,
   CPP_W_CXX11_COMPAT,
-  CPP_W_EXPANSION_TO_DEFINED
+  CPP_W_EXPANSION_TO_DEFINED,
+  CPP_W_BIDIRECTIONAL
 };
 
 /* Callback for header lookup for HEADER, which is the name of a
diff --git a/libcpp/init.c b/libcpp/init.c
index 5a424e23553..f9a8f5f088f 100644
--- a/libcpp/init.c
+++ b/libcpp/init.c
@@ -223,6 +223,7 @@ cpp_create_reader (enum c_lang lang, cpp_hash_table *table,
       = ENABLE_CANONICAL_SYSTEM_HEADERS;
   CPP_OPTION (pfile, ext_numeric_literals) = 1;
   CPP_OPTION (pfile, warn_date_time) = 0;
+  CPP_OPTION (pfile, cpp_warn_bidirectional) = bidirectional_unpaired;
 
   /* Default CPP arithmetic to something sensible for the host for the
      benefit of dumb users like fix-header.  */
diff --git a/libcpp/internal.h b/libcpp/internal.h
index 8577cab6c83..0ce0246c5a2 100644
--- a/libcpp/internal.h
+++ b/libcpp/internal.h
@@ -597,6 +597,13 @@ struct cpp_reader
   /* Location identifying the main source file -- intended to be line
      zero of said file.  */
   location_t main_loc;
+
+  /* Returns true iff we should warn about UTF-8 bidirectional control
+     characters.  */
+  bool warn_bidi_p () const
+  {
+    return CPP_OPTION (this, cpp_warn_bidirectional) != bidirectional_none;
+  }
 };
 
 /* Character classes.  Based on the more primitive macros in safe-ctype.h.
diff --git a/libcpp/lex.c b/libcpp/lex.c
index fa2253d41c3..f5907ce8f08 100644
--- a/libcpp/lex.c
+++ b/libcpp/lex.c
@@ -1164,6 +1164,324 @@ _cpp_process_line_notes (cpp_reader *pfile, int in_comment)
     }
 }
 
+namespace bidi {
+  enum class kind {
+    NONE, LRE, RLE, LRO, RLO, LRI, RLI, FSI, PDF, PDI, LTR, RTL
+  };
+
+  /* All the UTF-8 encodings of bidi characters start with E2.  */
+  constexpr uchar utf8_start = 0xe2;
+
+  /* A vector holding currently open bidi contexts.  We use a char for
+     each context, its LSB is 1 if it represents a PDF context, 0 if it
+     represents a PDI context.  The next bit is 1 if this context was open
+     by a bidi character written as a UCN, and 0 when it was UTF-8.  */
+  semi_embedded_vec <unsigned char, 16> vec;
+
+  /* Close the whole comment/identifier/string literal/character constant
+     context.  */
+  void on_close ()
+  {
+    vec.truncate (0);
+  }
+
+  /* Pop the last element in the vector.  */
+  void pop ()
+  {
+    unsigned int len = vec.count ();
+    gcc_checking_assert (len > 0);
+    vec.truncate (len - 1);
+  }
+
+  /* Return the context of the Ith element.  */
+  kind ctx_at (unsigned int i)
+  {
+    return (vec[i] & 1) ? kind::PDF : kind::PDI;
+  }
+
+  /* Return which context is currently opened.  */
+  kind current_ctx ()
+  {
+    unsigned int len = vec.count ();
+    if (len == 0)
+      return kind::NONE;
+    return ctx_at (len - 1);
+  }
+
+  /* Return true if the current context comes from a UCN origin, that is,
+     the bidi char which started this bidi context was written as a UCN.  */
+  bool current_ctx_ucn_p ()
+  {
+    unsigned int len = vec.count ();
+    gcc_checking_assert (len > 0);
+    return (vec[len - 1] >> 1) & 1;
+  }
+
+  /* We've read a bidi char, update the current vector as necessary.  */
+  void on_char (kind k, bool ucn_p)
+  {
+    switch (k)
+      {
+      case kind::LRE:
+      case kind::RLE:
+      case kind::LRO:
+      case kind::RLO:
+	vec.push (ucn_p ? 3u : 1u);
+	break;
+      case kind::LRI:
+      case kind::RLI:
+      case kind::FSI:
+	vec.push (ucn_p ? 2u : 0u);
+	break;
+      /* PDF terminates the scope of the last LRE, RLE, LRO, or RLO
+	 whose scope has not yet been terminated.  */
+      case kind::PDF:
+	if (current_ctx () == kind::PDF)
+	  pop ();
+	break;
+      /* PDI terminates the scope of the last LRI, RLI, or FSI whose
+	 scope has not yet been terminated, as well as the scopes of
+	 any subsequent LREs, RLEs, LROs, or RLOs whose scopes have not
+	 yet been terminated.  */
+      case kind::PDI:
+	for (int i = vec.count () - 1; i >= 0; --i)
+	  if (ctx_at (i) == kind::PDI)
+	    {
+	      vec.truncate (i);
+	      break;
+	    }
+	break;
+      case kind::LTR:
+      case kind::RTL:
+	/* These aren't popped by a PDF/PDI.  */
+	break;
+      [[likely]] case kind::NONE:
+	break;
+      default:
+	abort ();
+      }
+  }
+
+  /* Return a descriptive string for K.  */
+  const char *to_str (kind k)
+  {
+    switch (k)
+      {
+      case kind::LRE:
+	return "U+202A (LEFT-TO-RIGHT EMBEDDING)";
+      case kind::RLE:
+	return "U+202B (RIGHT-TO-LEFT EMBEDDING)";
+      case kind::LRO:
+	return "U+202D (LEFT-TO-RIGHT OVERRIDE)";
+      case kind::RLO:
+	return "U+202E (RIGHT-TO-LEFT OVERRIDE)";
+      case kind::LRI:
+	return "U+2066 (LEFT-TO-RIGHT ISOLATE)";
+      case kind::RLI:
+	return "U+2067 (RIGHT-TO-LEFT ISOLATE)";
+      case kind::FSI:
+	return "U+2068 (FIRST STRONG ISOLATE)";
+      case kind::PDF:
+	return "U+202C (POP DIRECTIONAL FORMATTING)";
+      case kind::PDI:
+	return "U+2069 (POP DIRECTIONAL ISOLATE)";
+      case kind::LTR:
+	return "U+200E (LEFT-TO-RIGHT MARK)";
+      case kind::RTL:
+	return "U+200F (RIGHT-TO-LEFT MARK)";
+      default:
+	abort ();
+      }
+  }
+}
+
+/* Parse a sequence of 3 bytes starting with P and return its bidi code.  */
+
+static bidi::kind
+get_bidi_utf8 (const unsigned char *const p)
+{
+  gcc_checking_assert (p[0] == bidi::utf8_start);
+
+  if (p[1] == 0x80)
+    switch (p[2])
+      {
+      case 0xaa:
+	return bidi::kind::LRE;
+      case 0xab:
+	return bidi::kind::RLE;
+      case 0xac:
+	return bidi::kind::PDF;
+      case 0xad:
+	return bidi::kind::LRO;
+      case 0xae:
+	return bidi::kind::RLO;
+      case 0x8e:
+	return bidi::kind::LTR;
+      case 0x8f:
+	return bidi::kind::RTL;
+      default:
+	break;
+      }
+  else if (p[1] == 0x81)
+    switch (p[2])
+      {
+      case 0xa6:
+	return bidi::kind::LRI;
+      case 0xa7:
+	return bidi::kind::RLI;
+      case 0xa8:
+	return bidi::kind::FSI;
+      case 0xa9:
+	return bidi::kind::PDI;
+      default:
+	break;
+      }
+
+  return bidi::kind::NONE;
+}
+
+/* Parse a UCN where P points just past \u or \U and return its bidi code.  */
+
+static bidi::kind
+get_bidi_ucn (const unsigned char *p, bool is_U)
+{
+  /* 6.4.3 Universal Character Names
+      \u hex-quad
+      \U hex-quad hex-quad
+     where \unnnn means \U0000nnnn.  */
+
+  if (is_U)
+    {
+      if (p[0] != '0' || p[1] != '0' || p[2] != '0' || p[3] != '0')
+	return bidi::kind::NONE;
+      /* Skip 4B so we can treat \u and \U the same below.  */
+      p += 4;
+    }
+
+  /* All code points we are looking for start with 20xx.  */
+  if (p[0] != '2' || p[1] != '0')
+    return bidi::kind::NONE;
+  else if (p[2] == '2')
+    switch (p[3])
+      {
+      case 'a':
+      case 'A':
+	return bidi::kind::LRE;
+      case 'b':
+      case 'B':
+	return bidi::kind::RLE;
+      case 'c':
+      case 'C':
+	return bidi::kind::PDF;
+      case 'd':
+      case 'D':
+	return bidi::kind::LRO;
+      case 'e':
+      case 'E':
+	return bidi::kind::RLO;
+      default:
+	break;
+      }
+  else if (p[2] == '6')
+    switch (p[3])
+      {
+      case '6':
+	return bidi::kind::LRI;
+      case '7':
+	return bidi::kind::RLI;
+      case '8':
+	return bidi::kind::FSI;
+      case '9':
+	return bidi::kind::PDI;
+      default:
+	break;
+      }
+  else if (p[2] == '0')
+    switch (p[3])
+      {
+      case 'e':
+      case 'E':
+	return bidi::kind::LTR;
+      case 'f':
+      case 'F':
+	return bidi::kind::RTL;
+      default:
+	break;
+      }
+
+  return bidi::kind::NONE;
+}
+
+/* We're closing a bidi context, that is, we've encountered a newline,
+   are closing a C-style comment, or are at the end of a string literal,
+   character constant, or identifier.  Warn if this context was not
+   properly terminated by a PDI or PDF.  P points to the last character
+   in this context.  */
+
+static void
+maybe_warn_bidi_on_close (cpp_reader *pfile, const uchar *p)
+{
+  if (CPP_OPTION (pfile, cpp_warn_bidirectional) == bidirectional_unpaired
+      && bidi::vec.count () > 0)
+    {
+      const location_t loc
+	= linemap_position_for_column (pfile->line_table,
+				       CPP_BUF_COLUMN (pfile->buffer, p));
+      cpp_warning_with_line (pfile, CPP_W_BIDIRECTIONAL, loc, 0,
+			     "unpaired UTF-8 bidirectional character "
+			     "detected");
+    }
+  /* We're done with this context.  */
+  bidi::on_close ();
+}
+
+/* We're at the beginning or in the middle of an identifier/comment/string
+   literal/character constant.  Warn if we've encountered a bidi character.
+   KIND says which bidi character it was; P points to it in the character
+   stream.  UCN_P is true iff this bidi character was written as a UCN.  */
+
+static void
+maybe_warn_bidi_on_char (cpp_reader *pfile, const uchar *p, bidi::kind kind,
+			 bool ucn_p)
+{
+  if (__builtin_expect (kind == bidi::kind::NONE, 1))
+    return;
+
+  const auto warn_bidi = CPP_OPTION (pfile, cpp_warn_bidirectional);
+
+  if (warn_bidi != bidirectional_none)
+    {
+      const location_t loc
+	= linemap_position_for_column (pfile->line_table,
+				       CPP_BUF_COLUMN (pfile->buffer, p));
+      /* It seems excessive to warn about a PDI/PDF that is closing
+	 an opened context because we've already warned about the
+	 opening character.  Except warn when we have a UCN x UTF-8
+	 mismatch.  */
+      if (kind == bidi::current_ctx ())
+	{
+	  if (warn_bidi == bidirectional_unpaired
+	      && bidi::current_ctx_ucn_p () != ucn_p)
+	    cpp_warning_with_line (pfile, CPP_W_BIDIRECTIONAL, loc, 0,
+				   "UTF-8 vs UCN mismatch when closing "
+				   "a context by \"%s\"", bidi::to_str (kind));
+	}
+      else if (warn_bidi == bidirectional_any)
+	{
+	  if (kind == bidi::kind::PDF || kind == bidi::kind::PDI)
+	    cpp_warning_with_line (pfile, CPP_W_BIDIRECTIONAL, loc, 0,
+				   "\"%s\" is closing an unopened context",
+				   bidi::to_str (kind));
+	  else
+	    cpp_warning_with_line (pfile, CPP_W_BIDIRECTIONAL, loc, 0,
+				   "found problematic Unicode character \"%s\"",
+				   bidi::to_str (kind));
+	}
+    }
+  /* We're done with this context.  */
+  bidi::on_char (kind, ucn_p);
+}
+
 /* Skip a C-style block comment.  We find the end of the comment by
    seeing if an asterisk is before every '/' we encounter.  Returns
    nonzero if comment terminated by EOF, zero otherwise.
@@ -1175,6 +1493,7 @@ _cpp_skip_block_comment (cpp_reader *pfile)
   cpp_buffer *buffer = pfile->buffer;
   const uchar *cur = buffer->cur;
   uchar c;
+  const bool warn_bidi_p = pfile->warn_bidi_p ();
 
   cur++;
   if (*cur == '/')
@@ -1189,7 +1508,11 @@ _cpp_skip_block_comment (cpp_reader *pfile)
       if (c == '/')
 	{
 	  if (cur[-2] == '*')
-	    break;
+	    {
+	      if (warn_bidi_p)
+		maybe_warn_bidi_on_close (pfile, cur);
+	      break;
+	    }
 
 	  /* Warn about potential nested comments, but not if the '/'
 	     comes immediately before the true comment delimiter.
@@ -1208,6 +1531,8 @@ _cpp_skip_block_comment (cpp_reader *pfile)
 	{
 	  unsigned int cols;
 	  buffer->cur = cur - 1;
+	  if (warn_bidi_p)
+	    maybe_warn_bidi_on_close (pfile, cur);
 	  _cpp_process_line_notes (pfile, true);
 	  if (buffer->next_line >= buffer->rlimit)
 	    return true;
@@ -1218,6 +1543,13 @@ _cpp_skip_block_comment (cpp_reader *pfile)
 
 	  cur = buffer->cur;
 	}
+      /* If this is a beginning of a UTF-8 encoding, it might be
+	 a bidirectional character.  */
+      else if (__builtin_expect (c == bidi::utf8_start, 0) && warn_bidi_p)
+	{
+	  bidi::kind kind = get_bidi_utf8 (cur - 1);
+	  maybe_warn_bidi_on_char (pfile, cur, kind, /*ucn_p=*/false);
+	}
     }
 
   buffer->cur = cur;
@@ -1233,9 +1565,31 @@ skip_line_comment (cpp_reader *pfile)
 {
   cpp_buffer *buffer = pfile->buffer;
   location_t orig_line = pfile->line_table->highest_line;
+  const bool warn_bidi_p = pfile->warn_bidi_p ();
 
-  while (*buffer->cur != '\n')
-    buffer->cur++;
+  if (!warn_bidi_p)
+    while (*buffer->cur != '\n')
+      buffer->cur++;
+  else
+    {
+      while (*buffer->cur != '\n'
+	     && *buffer->cur != bidi::utf8_start)
+	buffer->cur++;
+      if (__builtin_expect (*buffer->cur == bidi::utf8_start, 0))
+	{
+	  while (*buffer->cur != '\n')
+	    {
+	      if (__builtin_expect (*buffer->cur == bidi::utf8_start, 0))
+		{
+		  bidi::kind kind = get_bidi_utf8 (buffer->cur);
+		  maybe_warn_bidi_on_char (pfile, buffer->cur, kind,
+					   /*ucn_p=*/false);
+		}
+	      buffer->cur++;
+	    }
+	  maybe_warn_bidi_on_close (pfile, buffer->cur);
+	}
+    }
 
   _cpp_process_line_notes (pfile, true);
   return orig_line != pfile->line_table->highest_line;
@@ -1346,11 +1700,13 @@ static const cppchar_t utf8_signifier = 0xC0;
 
 /* Returns TRUE if the sequence starting at buffer->cur is valid in
    an identifier.  FIRST is TRUE if this starts an identifier.  */
+
 static bool
 forms_identifier_p (cpp_reader *pfile, int first,
 		    struct normalize_state *state)
 {
   cpp_buffer *buffer = pfile->buffer;
+  const bool warn_bidi_p = pfile->warn_bidi_p ();
 
   if (*buffer->cur == '$')
     {
@@ -1373,6 +1729,13 @@ forms_identifier_p (cpp_reader *pfile, int first,
       cppchar_t s;
       if (*buffer->cur >= utf8_signifier)
 	{
+	  if (__builtin_expect (*buffer->cur == bidi::utf8_start, 0)
+	      && warn_bidi_p)
+	    {
+	      bidi::kind kind = get_bidi_utf8 (buffer->cur);
+	      maybe_warn_bidi_on_char (pfile, buffer->cur, kind,
+				       /*ucn_p=*/false);
+	    }
 	  if (_cpp_valid_utf8 (pfile, &buffer->cur, buffer->rlimit, 1 + !first,
 			       state, &s))
 	    return true;
@@ -1381,6 +1744,13 @@ forms_identifier_p (cpp_reader *pfile, int first,
 	       && (buffer->cur[1] == 'u' || buffer->cur[1] == 'U'))
 	{
 	  buffer->cur += 2;
+	  if (warn_bidi_p)
+	    {
+	      bidi::kind kind = get_bidi_ucn (buffer->cur,
+					      buffer->cur[-1] == 'U');
+	      maybe_warn_bidi_on_char (pfile, buffer->cur, kind,
+				       /*ucn_p=*/true);
+	    }
 	  if (_cpp_valid_ucn (pfile, &buffer->cur, buffer->rlimit, 1 + !first,
 			      state, &s, NULL, NULL))
 	    return true;
@@ -1489,6 +1859,7 @@ lex_identifier (cpp_reader *pfile, const uchar *base, bool starts_ucn,
   const uchar *cur;
   unsigned int len;
   unsigned int hash = HT_HASHSTEP (0, *base);
+  const bool warn_bidi_p = pfile->warn_bidi_p ();
 
   cur = pfile->buffer->cur;
   if (! starts_ucn)
@@ -1505,13 +1876,17 @@ lex_identifier (cpp_reader *pfile, const uchar *base, bool starts_ucn,
     {
       /* Slower version for identifiers containing UCNs
 	 or extended chars (including $).  */
-      do {
-	while (ISIDNUM (*pfile->buffer->cur))
-	  {
-	    NORMALIZE_STATE_UPDATE_IDNUM (nst, *pfile->buffer->cur);
-	    pfile->buffer->cur++;
-	  }
-      } while (forms_identifier_p (pfile, false, nst));
+      do
+	{
+	  while (ISIDNUM (*pfile->buffer->cur))
+	    {
+	      NORMALIZE_STATE_UPDATE_IDNUM (nst, *pfile->buffer->cur);
+	      pfile->buffer->cur++;
+	    }
+	}
+      while (forms_identifier_p (pfile, false, nst));
+      if (warn_bidi_p)
+	maybe_warn_bidi_on_close (pfile, pfile->buffer->cur);
       result = _cpp_interpret_identifier (pfile, base,
 					  pfile->buffer->cur - base);
       *spelling = cpp_lookup (pfile, base, pfile->buffer->cur - base);
@@ -1758,6 +2133,7 @@ static void
 lex_raw_string (cpp_reader *pfile, cpp_token *token, const uchar *base)
 {
   const uchar *pos = base;
+  const bool warn_bidi_p = pfile->warn_bidi_p ();
 
   /* 'tis a pity this information isn't passed down from the lexer's
      initial categorization of the token.  */
@@ -1994,8 +2370,15 @@ lex_raw_string (cpp_reader *pfile, cpp_token *token, const uchar *base)
 	  pos = base = pfile->buffer->cur;
 	  note = &pfile->buffer->notes[pfile->buffer->cur_note];
 	}
+      else if (__builtin_expect ((unsigned char) c == bidi::utf8_start, 0)
+	       && warn_bidi_p)
+	maybe_warn_bidi_on_char (pfile, pos - 1, get_bidi_utf8 (pos - 1),
+				 /*ucn_p=*/false);
     }
 
+  if (warn_bidi_p)
+    maybe_warn_bidi_on_close (pfile, pos);
+
   if (CPP_OPTION (pfile, user_literals))
     {
       /* If a string format macro, say from inttypes.h, is placed touching
@@ -2090,15 +2473,27 @@ lex_string (cpp_reader *pfile, cpp_token *token, const uchar *base)
   else
     terminator = '>', type = CPP_HEADER_NAME;
 
+  const bool warn_bidi_p = pfile->warn_bidi_p ();
   for (;;)
     {
       cppchar_t c = *cur++;
 
       /* In #include-style directives, terminators are not escapable.  */
       if (c == '\\' && !pfile->state.angled_headers && *cur != '\n')
-	cur++;
+	{
+	  if ((cur[0] == 'u' || cur[0] == 'U') && warn_bidi_p)
+	    {
+	      bidi::kind kind = get_bidi_ucn (cur + 1, cur[0] == 'U');
+	      maybe_warn_bidi_on_char (pfile, cur, kind, /*ucn_p=*/true);
+	    }
+	  cur++;
+	}
       else if (c == terminator)
-	break;
+	{
+	  if (warn_bidi_p)
+	    maybe_warn_bidi_on_close (pfile, cur - 1);
+	  break;
+	}
       else if (c == '\n')
 	{
 	  cur--;
@@ -2115,6 +2510,11 @@ lex_string (cpp_reader *pfile, cpp_token *token, const uchar *base)
 	}
       else if (c == '\0')
 	saw_NUL = true;
+      else if (__builtin_expect (c == bidi::utf8_start, 0) && warn_bidi_p)
+	{
+	  bidi::kind kind = get_bidi_utf8 (cur - 1);
+	  maybe_warn_bidi_on_char (pfile, cur - 1, kind, /*ucn_p=*/false);
+	}
     }
 
   if (saw_NUL && !pfile->state.skipping)

base-commit: 4cdf7db9a39d18bd536d816a5751d4d3cf23808b
-- 
2.33.1



More information about the Gcc-patches mailing list