gbranden pushed a commit to branch master
in repository groff.
commit 129605355239f448ba517a807e0f3bf46fdeb52e
Author: G. Branden Robinson <[email protected]>
AuthorDate: Thu Mar 19 11:34:34 2026 -0500
[troff]: Fix Savannah #66547.
* src/roff/troff/input.cpp (transparent_throughput_file_request):
Disallow all grout-invalid characters in "safer" transparent file
throughput (implemented by the `trf` request), including all C0
control code points except the line feed (Unix newline), and all C1
controls and Latin-1 extension code points. Revise warning diagnostic
message.
* doc/groff.texi.in (Host System Service Access, Warnings):
* man/groff.7.man (Request short reference):
* man/groff_diff.7.man (New requests):
* src/roff/troff/troff.1.man (Warnings):
* NEWS: Document it.
Fixes <https://savannah.gnu.org/bugs/?66547>.
Illustration:
$ printf '# foo\0\001\002\003\004\005\006\007\241bar\n# baz\n' >|
ATTIC/bad-trf
$ printf '.trf ATTIC/bad-trf\n' | ./build/test-groff -ww -Z
x T ps
x res 72000 1 1
x init
p1
troff:<standard input>:1: warning: character code 0 is invalid in
transparent file throughput; ignoring [-w input]
troff:<standard input>:1: warning: character code 1 is invalid in
transparent file throughput; ignoring [-w input]
troff:<standard input>:1: warning: character code 2 is invalid in
transparent file throughput; ignoring [-w input]
troff:<standard input>:1: warning: character code 3 is invalid in
transparent file throughput; ignoring [-w input]
troff:<standard input>:1: warning: character code 4 is invalid in
transparent file throughput; ignoring [-w input]
troff:<standard input>:1: warning: character code 5 is invalid in
transparent file throughput; ignoring [-w input]
troff:<standard input>:1: warning: character code 6 is invalid in
transparent file throughput; ignoring [-w input]
troff:<standard input>:1: warning: character code 7 is invalid in
transparent file throughput; ignoring [-w input]
troff:<standard input>:1: warning: character code 161 is invalid in
transparent file throughput; ignoring [-w input]
\# foobar
\# baz
x trailer
V792000
x stop
(Comment characters are backslash-prefixed when leading a line to keep
Git from discarding them from this commit log message.)
---
ChangeLog | 18 ++++++++++++++++++
NEWS | 5 +++++
doc/groff.texi.in | 28 ++++++++++++++++++++--------
man/groff.7.man | 18 ++++++++++++------
man/groff_diff.7.man | 30 ++++++++++++++++++++++++------
src/roff/troff/input.cpp | 6 +++---
src/roff/troff/troff.1.man | 4 +++-
7 files changed, 85 insertions(+), 24 deletions(-)
diff --git a/ChangeLog b/ChangeLog
index b2044b449..54480a7fe 100644
--- a/ChangeLog
+++ b/ChangeLog
@@ -1,3 +1,21 @@
+2026-03-19 G. Branden Robinson <[email protected]>
+
+ * src/roff/troff/input.cpp
+ (transparent_throughput_file_request): Disallow all
+ grout-invalid characters in "safer" transparent file throughput
+ {implemented by the `trf` request}, including all C0 control
+ code points except the line feed (Unix newline), and all C1
+ controls and Latin-1 extension code points. Revise warning
+ diagnostic message.
+
+ * doc/groff.texi.in (Host System Service Access, Warnings):
+ * man/groff.7.man (Request short reference):
+ * man/groff_diff.7.man (New requests):
+ * src/roff/troff/troff.1.man (Warnings):
+ * NEWS: Document it.
+
+ Fixes <https://savannah.gnu.org/bugs/?66547>.
+
2026-03-19 G. Branden Robinson <[email protected]>
* src/devices/grotty/tty.cpp (tty_printer::end_page): Further
diff --git a/NEWS b/NEWS
index e9c153903..3cb4aff27 100644
--- a/NEWS
+++ b/NEWS
@@ -18,6 +18,11 @@ troff
* GNU troff, the formatter, now issues warnings in the "delim",
"syntax", and "escape" categories by default.
+* The `trf` (transparent file throughput) request now discards input
+ characters that are invalid in GNU troff's output format, meaning all
+ 8-bit character codes less than decimal 32 (except 10, the line feed)
+ and greater than decimal 127.
+
* A new read-only register `.ul` reports the count of productive input
lines remaining to be be underlined in the environment (as configured
by either the `cu` or `ul` requests).
diff --git a/doc/groff.texi.in b/doc/groff.texi.in
index 7f6a87a22..507f10d78 100644
--- a/doc/groff.texi.in
+++ b/doc/groff.texi.in
@@ -20036,7 +20036,8 @@ in
@cindex mode, copy, and @code{trf} request
Break and copy the contents of
@var{file}
-as ``throughput'' to GNU
+as ``throughput'' to
+GNU
@command{troff}'s
output.
Each line of
@@ -20057,20 +20058,29 @@ GNU
@command{troff} @c GNU
performs the copy only when the diversion is emitted.
-@cindex @code{trf} request, and invalid characters
-@cindex characters, invalid for @code{trf} request
-@cindex invalid characters for @code{trf} request
@code{cf} copies the contents of @var{file} completely unprocessed;
it is therefore an error to use this request in safer mode,
which is the default.
-Invoke GNU
+Invoke
+GNU
@command{troff} @c GNU
or a front end with the
@option{-U}
option to enable unsafe mode.
-@code{trf} discards invalid input characters;
-recall @ref{Input Format}.
+@cindex @code{trf} request, and invalid characters
+@cindex characters, invalid for @code{trf} request
+@cindex invalid characters for @code{trf} request
+@code{trf}
+discards input characters that are invalid in
+GNU
+@command{troff}'s @c GNU
+output format,
+meaning all 8-bit character codes less than decimal 32
+(except 10,
+the line feed)
+and greater than decimal 127.@footnote{@xref{GNU
+@command{troff} Output}.}
For @code{cf}, within a diversion, ``completely unprocessed'' means that
each line of a file to be inserted is handled as if it were preceded by
@@ -21713,7 +21723,9 @@ A request was invoked with a mandatory argument absent.
@item input
@itemx 16384
-An invalid character occurred on the input stream.
+An invalid character occurred on the input stream,
+or was present in the contents of a file processed
+as transparent throughput.
@item escape
@itemx 32768
diff --git a/man/groff.7.man b/man/groff.7.man
index 5cc69fd48..fae1fcffc 100644
--- a/man/groff.7.man
+++ b/man/groff.7.man
@@ -5377,18 +5377,24 @@ output
(see
.MR groff_out @MAN5EXT@ ).
.
-Unlike
-.request cf ,
-characters invalid as input to GNU
-.I troff \" GNU
-are discarded.
-.
If
.IR file 's
contents do not end with a newline,
.B trf
appends one.
.
+Unlike
+.request cf ,
+.B trf
+discards input characters that are invalid in
+GNU
+.IR troff 's \" GNU
+output format,
+meaning all 8-bit character codes less than decimal 32
+(except 10,
+the line feed)
+and greater than decimal 127.
+.
.TPx
.REQ .trin abcd\fR\&.\|.\|.\&\fP
As
diff --git a/man/groff_diff.7.man b/man/groff_diff.7.man
index c26c82e4b..2b3405cb1 100644
--- a/man/groff_diff.7.man
+++ b/man/groff_diff.7.man
@@ -12,7 +12,7 @@ and AT&T
.\" ====================================================================
.\"
.\" Copyright 1989-2024 Free Software Foundation, Inc.
-.\" 2019-2025 G. Branden Robinson
+.\" 2019-2026 G. Branden Robinson
.\"
.\" This file is part of groff, the GNU roff typesetting system.
.\"
@@ -4526,12 +4526,10 @@ but does not append a newline.
.BI .trf\~ file
Break and copy
.I file
-as \[lq]throughput\[rq] to GNU
+as \[lq]throughput\[rq] to
+GNU
.I troff \" GNU
-output,
-discarding characters that are invalid as input;
-contrast with
-.BR cf .
+output.
.
Each line of
.I file
@@ -4549,6 +4547,26 @@ Invoking the request
with the no-break control character
suppresses the break.
.
+If a diversion is in use,
+GNU
+.I troff \" GNU
+performs the copy only when the diversion is emitted.
+.
+.
+.IP
+.B trf
+discards input characters that are invalid in
+GNU
+.IR troff 's \" GNU
+output format,
+meaning all 8-bit character codes less than decimal 32
+(except 10,
+the line feed)
+and greater than decimal 127.
+.
+See
+.MR groff_out @MAN5EXT@ .
+.
.
.TP
.BI .trin\~ abcd
diff --git a/src/roff/troff/input.cpp b/src/roff/troff/input.cpp
index 659350a2c..9b7ee46af 100644
--- a/src/roff/troff/input.cpp
+++ b/src/roff/troff/input.cpp
@@ -9563,9 +9563,9 @@ static void transparent_throughput_file_request()
int c = getc(fp);
if (c == EOF)
break;
- if (is_invalid_input_char(c))
- warning(WARN_INPUT, "invalid input character code %1",
- int(c));
+ if ((c != '\n') && ((c < 32) || (c > 127)))
+ warning(WARN_INPUT, "character code %1 is invalid in"
+ " transparent file throughput; ignoring", int(c));
else {
curdiv->transparent_output(c);
reading_beginning_of_input_line = c == '\n';
diff --git a/src/roff/troff/troff.1.man b/src/roff/troff/troff.1.man
index 1364e2a3b..cd8339586 100644
--- a/src/roff/troff/troff.1.man
+++ b/src/roff/troff/troff.1.man
@@ -708,7 +708,9 @@ T}
.
.sp \n[PD]u
input 16384 T{
-An invalid character occurred on the input stream.
+An invalid character occurred on the input stream,
+or was present in the contents of a file processed
+as transparent throughput.
T}
.
.
_______________________________________________
groff-commit mailing list
[email protected]
https://lists.gnu.org/mailman/listinfo/groff-commit
