Delete support for the LESSOPEN and LESSCLOSE environment variables
aka the "Input Preprocessor": it's been a source for multiple
security bugs in the past as everything has to deal with handling
arbitrary filenames and generally resulted in multiple TOCTOU issues.
The base system never included a default LESSOPEN setting like some
Linux distributions did, but it's a suds-filled sink full of knives
to try to use safely.

ok tb@ deraadt@ millert@

Delete what remains of the support for combining characters into
ligatures: it was incomplete (only for the Arabic script and only
for the single ligature LAM WITH ALEF) and it was implemented in a
way that is unsustainable (with a static table inside less).

If we ever want ligature support, we are better off making a fresh
start. However, for languages like Arabic and Persian, even that
wouldn't really be useful without having bidirectional support first.

OK millert@
(and also considering comments from Mohammadreza Abdollahzadeh,
Evan Silberman, and benno@)

New utility function mbtowc_left() because this functionality -
decoding a UTF-8 multibyte character to the left of a given byte -
is already needed at three places in line.c and will also be needed
for cleanup work in cmdbuf.c in the future.
OK millert@

trivial patch to at least make the buggy functions control_char()
and get_wchar() static for now - until they can be deleted

Clean up all major UTF-8 issues in cvt.c.
This also allows to delete the buggy, now unused function put_wchar().
OK millert@

Completely clean up UTF-8 handling in the file filename.c:
Employ the usual form of an mbtowc(3) loop, eliminating two calls
to the bad function step_char() and reducing the number of nested
loops by one. This also removes the last caller of the bad function
binary_char(), which is consequently deleted.
While here, count ASCII C0 non-whitespace control characters as
binary (except backspace and, with -R only, escape).
OK millert@

Start UTF-8 cleanup of store_char(), in particular getting rid of
a call to the flawed function step_char(-1), using the standard
function mbtowc(3) instead.
Merge in in_ansi_esc_seq(), simplifying the code, and make the
related functions is_ansi_end() and is_ascii_char() static because
they are used in line.c only.

OK nicm@, and no opposition when shown on tech@

Split only LESS on $.

As guenther@ said "STOP SPLITTING ANYTHING BUT $LESS ON '$' !".

anton@ came up with the same diff. ok nicm@

Remove the lintism PRINTFLIKE1, don't replace it with an attribute so we
can keep it portable.

also noticed by Michael Reed, discussed with deraadt@

remove needless extern keywords, a vestige of pre-ANSI C

explained by and ok millert@

Remove LINUNUM, a needless alias for off_t. Also, remove a redundant
condition.

ok nicm@

Missing prototype.

remove fancy null check function in favor of checking for null.
from Ricardo Mestre

Add a flag argument to flush() to stop it calling quit() on error, then
use this from quit() to stop less blowing up the stack looping through
quit()/flush() if stderr is closed (for example "less /missing
2</dev/null"). ok millert

close_getchr() is now a noop

replace save() with estrdup() and make estrdup() exit like all the other
ecalloc, etc. functions do.
ok mmcc nicm

time_t is not a long. remove incorrect abstraction around time().
ok nicm

Replace less with the cleaned-up fork of less 458 maintained by Garrett
D'Amore at https://github.com/gdamore/less-fork. This has significantly
less portability goop, has a tidied up code style, uses terminfo instead
of termcap, and is has stricter POSIX compliance.

Many of our local changes have been accepted upstream: substantial
remaining local changes are code to read help files from /usr/share
rather than compiling them in, man page and help improvements, and some
tweaks to the default options.

Review and testing by millert, ok deraadt

Give more(1) its own help file, shorter than the one for less(1).
The helpfile itself was prepared by jmc@, the glue by me.
OK millert@ jmc@

Merge Less 458 plus local patches

ok guenther@, ok jmc@ for man bits

Merge in less 444 plus local changes

ok nicm@

Use interruptible syscalls instead of setjmp/longjmp. This makes
less's signal handlers safe. No one has reported problems so far...

o go back to using a help file instead of embedding it in the binary
o deal with HELPFILE not being defined
o add a SMALL definition to make a smaller less for the boot floppies
This still needs to be pared down a bit for the SMALL case

Merge in less-381 w/ local changes and remove obsolete files.

$OpenBSD$

branches: 1.1.1;
Initial revision

Delete what remains of the support for combining characters into
ligatures: it was incomplete (only for the Arabic script and only
for the single ligature LAM WITH ALEF) and it was implemented in a
way that is unsustainable (with a static table inside less).

If we ever want ligature support, we are better off making a fresh
start. However, for languages like Arabic and Persian, even that
wouldn't really be useful without having bidirectional support first.

OK millert@
(and also considering comments from Mohammadreza Abdollahzadeh,
Evan Silberman, and benno@)

New utility function mbtowc_left() because this functionality -
decoding a UTF-8 multibyte character to the left of a given byte -
is already needed at three places in line.c and will also be needed
for cleanup work in cmdbuf.c in the future.
OK millert@

trivial patch to at least make the buggy functions control_char()
and get_wchar() static for now - until they can be deleted

Clean up all major UTF-8 issues in cvt.c.
This also allows to delete the buggy, now unused function put_wchar().
OK millert@

Completely clean up UTF-8 handling in the file filename.c:
Employ the usual form of an mbtowc(3) loop, eliminating two calls
to the bad function step_char() and reducing the number of nested
loops by one. This also removes the last caller of the bad function
binary_char(), which is consequently deleted.
While here, count ASCII C0 non-whitespace control characters as
binary (except backspace and, with -R only, escape).
OK millert@

Start UTF-8 cleanup of store_char(), in particular getting rid of
a call to the flawed function step_char(-1), using the standard
function mbtowc(3) instead.
Merge in in_ansi_esc_seq(), simplifying the code, and make the
related functions is_ansi_end() and is_ascii_char() static because
they are used in line.c only.

OK nicm@, and no opposition when shown on tech@

Split only LESS on $.

As guenther@ said "STOP SPLITTING ANYTHING BUT $LESS ON '$' !".

anton@ came up with the same diff. ok nicm@

Remove the lintism PRINTFLIKE1, don't replace it with an attribute so we
can keep it portable.

also noticed by Michael Reed, discussed with deraadt@

remove needless extern keywords, a vestige of pre-ANSI C

explained by and ok millert@

Remove LINUNUM, a needless alias for off_t. Also, remove a redundant
condition.

ok nicm@

Missing prototype.

remove fancy null check function in favor of checking for null.
from Ricardo Mestre

Add a flag argument to flush() to stop it calling quit() on error, then
use this from quit() to stop less blowing up the stack looping through
quit()/flush() if stderr is closed (for example "less /missing
2</dev/null"). ok millert

close_getchr() is now a noop

replace save() with estrdup() and make estrdup() exit like all the other
ecalloc, etc. functions do.
ok mmcc nicm

time_t is not a long. remove incorrect abstraction around time().
ok nicm

Replace less with the cleaned-up fork of less 458 maintained by Garrett
D'Amore at https://github.com/gdamore/less-fork. This has significantly
less portability goop, has a tidied up code style, uses terminfo instead
of termcap, and is has stricter POSIX compliance.

Many of our local changes have been accepted upstream: substantial
remaining local changes are code to read help files from /usr/share
rather than compiling them in, man page and help improvements, and some
tweaks to the default options.

Review and testing by millert, ok deraadt

Give more(1) its own help file, shorter than the one for less(1).
The helpfile itself was prepared by jmc@, the glue by me.
OK millert@ jmc@

Merge Less 458 plus local patches

ok guenther@, ok jmc@ for man bits

Merge in less 444 plus local changes

ok nicm@

Use interruptible syscalls instead of setjmp/longjmp. This makes
less's signal handlers safe. No one has reported problems so far...

o go back to using a help file instead of embedding it in the binary
o deal with HELPFILE not being defined
o add a SMALL definition to make a smaller less for the boot floppies
This still needs to be pared down a bit for the SMALL case

Merge in less-381 w/ local changes and remove obsolete files.

$OpenBSD$

branches: 1.1.1;
Initial revision

New utility function mbtowc_left() because this functionality -
decoding a UTF-8 multibyte character to the left of a given byte -
is already needed at three places in line.c and will also be needed
for cleanup work in cmdbuf.c in the future.
OK millert@

trivial patch to at least make the buggy functions control_char()
and get_wchar() static for now - until they can be deleted

Clean up all major UTF-8 issues in cvt.c.
This also allows to delete the buggy, now unused function put_wchar().
OK millert@

Completely clean up UTF-8 handling in the file filename.c:
Employ the usual form of an mbtowc(3) loop, eliminating two calls
to the bad function step_char() and reducing the number of nested
loops by one. This also removes the last caller of the bad function
binary_char(), which is consequently deleted.
While here, count ASCII C0 non-whitespace control characters as
binary (except backspace and, with -R only, escape).
OK millert@

Start UTF-8 cleanup of store_char(), in particular getting rid of
a call to the flawed function step_char(-1), using the standard
function mbtowc(3) instead.
Merge in in_ansi_esc_seq(), simplifying the code, and make the
related functions is_ansi_end() and is_ascii_char() static because
they are used in line.c only.

OK nicm@, and no opposition when shown on tech@

Split only LESS on $.

As guenther@ said "STOP SPLITTING ANYTHING BUT $LESS ON '$' !".

anton@ came up with the same diff. ok nicm@

Remove the lintism PRINTFLIKE1, don't replace it with an attribute so we
can keep it portable.

also noticed by Michael Reed, discussed with deraadt@

remove needless extern keywords, a vestige of pre-ANSI C

explained by and ok millert@

Remove LINUNUM, a needless alias for off_t. Also, remove a redundant
condition.

ok nicm@

Missing prototype.

remove fancy null check function in favor of checking for null.
from Ricardo Mestre

Add a flag argument to flush() to stop it calling quit() on error, then
use this from quit() to stop less blowing up the stack looping through
quit()/flush() if stderr is closed (for example "less /missing
2</dev/null"). ok millert

close_getchr() is now a noop

replace save() with estrdup() and make estrdup() exit like all the other
ecalloc, etc. functions do.
ok mmcc nicm

time_t is not a long. remove incorrect abstraction around time().
ok nicm

Replace less with the cleaned-up fork of less 458 maintained by Garrett
D'Amore at https://github.com/gdamore/less-fork. This has significantly
less portability goop, has a tidied up code style, uses terminfo instead
of termcap, and is has stricter POSIX compliance.

Many of our local changes have been accepted upstream: substantial
remaining local changes are code to read help files from /usr/share
rather than compiling them in, man page and help improvements, and some
tweaks to the default options.

Review and testing by millert, ok deraadt

Give more(1) its own help file, shorter than the one for less(1).
The helpfile itself was prepared by jmc@, the glue by me.
OK millert@ jmc@

Merge Less 458 plus local patches

ok guenther@, ok jmc@ for man bits

Merge in less 444 plus local changes

ok nicm@

Use interruptible syscalls instead of setjmp/longjmp. This makes
less's signal handlers safe. No one has reported problems so far...

o go back to using a help file instead of embedding it in the binary
o deal with HELPFILE not being defined
o add a SMALL definition to make a smaller less for the boot floppies
This still needs to be pared down a bit for the SMALL case

Merge in less-381 w/ local changes and remove obsolete files.

$OpenBSD$

branches: 1.1.1;
Initial revision

trivial patch to at least make the buggy functions control_char()
and get_wchar() static for now - until they can be deleted

Clean up all major UTF-8 issues in cvt.c.
This also allows to delete the buggy, now unused function put_wchar().
OK millert@

Completely clean up UTF-8 handling in the file filename.c:
Employ the usual form of an mbtowc(3) loop, eliminating two calls
to the bad function step_char() and reducing the number of nested
loops by one. This also removes the last caller of the bad function
binary_char(), which is consequently deleted.
While here, count ASCII C0 non-whitespace control characters as
binary (except backspace and, with -R only, escape).
OK millert@

Start UTF-8 cleanup of store_char(), in particular getting rid of
a call to the flawed function step_char(-1), using the standard
function mbtowc(3) instead.
Merge in in_ansi_esc_seq(), simplifying the code, and make the
related functions is_ansi_end() and is_ascii_char() static because
they are used in line.c only.

OK nicm@, and no opposition when shown on tech@

Split only LESS on $.

As guenther@ said "STOP SPLITTING ANYTHING BUT $LESS ON '$' !".

anton@ came up with the same diff. ok nicm@

Remove the lintism PRINTFLIKE1, don't replace it with an attribute so we
can keep it portable.

also noticed by Michael Reed, discussed with deraadt@

remove needless extern keywords, a vestige of pre-ANSI C

explained by and ok millert@

Remove LINUNUM, a needless alias for off_t. Also, remove a redundant
condition.

ok nicm@

Missing prototype.

remove fancy null check function in favor of checking for null.
from Ricardo Mestre

Add a flag argument to flush() to stop it calling quit() on error, then
use this from quit() to stop less blowing up the stack looping through
quit()/flush() if stderr is closed (for example "less /missing
2</dev/null"). ok millert

close_getchr() is now a noop

replace save() with estrdup() and make estrdup() exit like all the other
ecalloc, etc. functions do.
ok mmcc nicm

time_t is not a long. remove incorrect abstraction around time().
ok nicm

Replace less with the cleaned-up fork of less 458 maintained by Garrett
D'Amore at https://github.com/gdamore/less-fork. This has significantly
less portability goop, has a tidied up code style, uses terminfo instead
of termcap, and is has stricter POSIX compliance.

Many of our local changes have been accepted upstream: substantial
remaining local changes are code to read help files from /usr/share
rather than compiling them in, man page and help improvements, and some
tweaks to the default options.

Review and testing by millert, ok deraadt

Give more(1) its own help file, shorter than the one for less(1).
The helpfile itself was prepared by jmc@, the glue by me.
OK millert@ jmc@

Merge Less 458 plus local patches

ok guenther@, ok jmc@ for man bits

Merge in less 444 plus local changes

ok nicm@

Use interruptible syscalls instead of setjmp/longjmp. This makes
less's signal handlers safe. No one has reported problems so far...

o go back to using a help file instead of embedding it in the binary
o deal with HELPFILE not being defined
o add a SMALL definition to make a smaller less for the boot floppies
This still needs to be pared down a bit for the SMALL case

Merge in less-381 w/ local changes and remove obsolete files.

$OpenBSD$

branches: 1.1.1;
Initial revision

Completely clean up UTF-8 handling in the file filename.c:
Employ the usual form of an mbtowc(3) loop, eliminating two calls
to the bad function step_char() and reducing the number of nested
loops by one. This also removes the last caller of the bad function
binary_char(), which is consequently deleted.
While here, count ASCII C0 non-whitespace control characters as
binary (except backspace and, with -R only, escape).
OK millert@

Start UTF-8 cleanup of store_char(), in particular getting rid of
a call to the flawed function step_char(-1), using the standard
function mbtowc(3) instead.
Merge in in_ansi_esc_seq(), simplifying the code, and make the
related functions is_ansi_end() and is_ascii_char() static because
they are used in line.c only.

OK nicm@, and no opposition when shown on tech@

Split only LESS on $.

As guenther@ said "STOP SPLITTING ANYTHING BUT $LESS ON '$' !".

anton@ came up with the same diff. ok nicm@

Remove the lintism PRINTFLIKE1, don't replace it with an attribute so we
can keep it portable.

also noticed by Michael Reed, discussed with deraadt@

remove needless extern keywords, a vestige of pre-ANSI C

explained by and ok millert@

Remove LINUNUM, a needless alias for off_t. Also, remove a redundant
condition.

ok nicm@

Missing prototype.

remove fancy null check function in favor of checking for null.
from Ricardo Mestre

Add a flag argument to flush() to stop it calling quit() on error, then
use this from quit() to stop less blowing up the stack looping through
quit()/flush() if stderr is closed (for example "less /missing
2</dev/null"). ok millert

close_getchr() is now a noop

replace save() with estrdup() and make estrdup() exit like all the other
ecalloc, etc. functions do.
ok mmcc nicm

time_t is not a long. remove incorrect abstraction around time().
ok nicm

Replace less with the cleaned-up fork of less 458 maintained by Garrett
D'Amore at https://github.com/gdamore/less-fork. This has significantly
less portability goop, has a tidied up code style, uses terminfo instead
of termcap, and is has stricter POSIX compliance.

Many of our local changes have been accepted upstream: substantial
remaining local changes are code to read help files from /usr/share
rather than compiling them in, man page and help improvements, and some
tweaks to the default options.

Review and testing by millert, ok deraadt

Give more(1) its own help file, shorter than the one for less(1).
The helpfile itself was prepared by jmc@, the glue by me.
OK millert@ jmc@

Merge Less 458 plus local patches

ok guenther@, ok jmc@ for man bits

Merge in less 444 plus local changes

ok nicm@

Use interruptible syscalls instead of setjmp/longjmp. This makes
less's signal handlers safe. No one has reported problems so far...

o go back to using a help file instead of embedding it in the binary
o deal with HELPFILE not being defined
o add a SMALL definition to make a smaller less for the boot floppies
This still needs to be pared down a bit for the SMALL case

Merge in less-381 w/ local changes and remove obsolete files.

$OpenBSD$

branches: 1.1.1;
Initial revision

Start UTF-8 cleanup of store_char(), in particular getting rid of
a call to the flawed function step_char(-1), using the standard
function mbtowc(3) instead.
Merge in in_ansi_esc_seq(), simplifying the code, and make the
related functions is_ansi_end() and is_ascii_char() static because
they are used in line.c only.

OK nicm@, and no opposition when shown on tech@

Split only LESS on $.

As guenther@ said "STOP SPLITTING ANYTHING BUT $LESS ON '$' !".

anton@ came up with the same diff. ok nicm@

Remove the lintism PRINTFLIKE1, don't replace it with an attribute so we
can keep it portable.

also noticed by Michael Reed, discussed with deraadt@

remove needless extern keywords, a vestige of pre-ANSI C

explained by and ok millert@

Remove LINUNUM, a needless alias for off_t. Also, remove a redundant
condition.

ok nicm@

Missing prototype.

remove fancy null check function in favor of checking for null.
from Ricardo Mestre

Add a flag argument to flush() to stop it calling quit() on error, then
use this from quit() to stop less blowing up the stack looping through
quit()/flush() if stderr is closed (for example "less /missing
2</dev/null"). ok millert

close_getchr() is now a noop

replace save() with estrdup() and make estrdup() exit like all the other
ecalloc, etc. functions do.
ok mmcc nicm

time_t is not a long. remove incorrect abstraction around time().
ok nicm

Replace less with the cleaned-up fork of less 458 maintained by Garrett
D'Amore at https://github.com/gdamore/less-fork. This has significantly
less portability goop, has a tidied up code style, uses terminfo instead
of termcap, and is has stricter POSIX compliance.

Many of our local changes have been accepted upstream: substantial
remaining local changes are code to read help files from /usr/share
rather than compiling them in, man page and help improvements, and some
tweaks to the default options.

Review and testing by millert, ok deraadt

Give more(1) its own help file, shorter than the one for less(1).
The helpfile itself was prepared by jmc@, the glue by me.
OK millert@ jmc@

Merge Less 458 plus local patches

ok guenther@, ok jmc@ for man bits

Merge in less 444 plus local changes

ok nicm@

Use interruptible syscalls instead of setjmp/longjmp. This makes
less's signal handlers safe. No one has reported problems so far...

o go back to using a help file instead of embedding it in the binary
o deal with HELPFILE not being defined
o add a SMALL definition to make a smaller less for the boot floppies
This still needs to be pared down a bit for the SMALL case

Merge in less-381 w/ local changes and remove obsolete files.

$OpenBSD$

branches: 1.1.1;
Initial revision

Split only LESS on $.

As guenther@ said "STOP SPLITTING ANYTHING BUT $LESS ON '$' !".

anton@ came up with the same diff. ok nicm@

Remove the lintism PRINTFLIKE1, don't replace it with an attribute so we
can keep it portable.

also noticed by Michael Reed, discussed with deraadt@

remove needless extern keywords, a vestige of pre-ANSI C

explained by and ok millert@

Remove LINUNUM, a needless alias for off_t. Also, remove a redundant
condition.

ok nicm@

Missing prototype.

remove fancy null check function in favor of checking for null.
from Ricardo Mestre

Add a flag argument to flush() to stop it calling quit() on error, then
use this from quit() to stop less blowing up the stack looping through
quit()/flush() if stderr is closed (for example "less /missing
2</dev/null"). ok millert

close_getchr() is now a noop

replace save() with estrdup() and make estrdup() exit like all the other
ecalloc, etc. functions do.
ok mmcc nicm

time_t is not a long. remove incorrect abstraction around time().
ok nicm

Replace less with the cleaned-up fork of less 458 maintained by Garrett
D'Amore at https://github.com/gdamore/less-fork. This has significantly
less portability goop, has a tidied up code style, uses terminfo instead
of termcap, and is has stricter POSIX compliance.

Many of our local changes have been accepted upstream: substantial
remaining local changes are code to read help files from /usr/share
rather than compiling them in, man page and help improvements, and some
tweaks to the default options.

Review and testing by millert, ok deraadt

Give more(1) its own help file, shorter than the one for less(1).
The helpfile itself was prepared by jmc@, the glue by me.
OK millert@ jmc@

Merge Less 458 plus local patches

ok guenther@, ok jmc@ for man bits

Merge in less 444 plus local changes

ok nicm@

Use interruptible syscalls instead of setjmp/longjmp. This makes
less's signal handlers safe. No one has reported problems so far...

o go back to using a help file instead of embedding it in the binary
o deal with HELPFILE not being defined
o add a SMALL definition to make a smaller less for the boot floppies
This still needs to be pared down a bit for the SMALL case

Merge in less-381 w/ local changes and remove obsolete files.

$OpenBSD$

branches: 1.1.1;
Initial revision

Remove the lintism PRINTFLIKE1, don't replace it with an attribute so we
can keep it portable.

also noticed by Michael Reed, discussed with deraadt@

remove needless extern keywords, a vestige of pre-ANSI C

explained by and ok millert@

Remove LINUNUM, a needless alias for off_t. Also, remove a redundant
condition.

ok nicm@

Missing prototype.

remove fancy null check function in favor of checking for null.
from Ricardo Mestre

Add a flag argument to flush() to stop it calling quit() on error, then
use this from quit() to stop less blowing up the stack looping through
quit()/flush() if stderr is closed (for example "less /missing
2</dev/null"). ok millert

close_getchr() is now a noop

replace save() with estrdup() and make estrdup() exit like all the other
ecalloc, etc. functions do.
ok mmcc nicm

time_t is not a long. remove incorrect abstraction around time().
ok nicm

Replace less with the cleaned-up fork of less 458 maintained by Garrett
D'Amore at https://github.com/gdamore/less-fork. This has significantly
less portability goop, has a tidied up code style, uses terminfo instead
of termcap, and is has stricter POSIX compliance.

Many of our local changes have been accepted upstream: substantial
remaining local changes are code to read help files from /usr/share
rather than compiling them in, man page and help improvements, and some
tweaks to the default options.

Review and testing by millert, ok deraadt

Give more(1) its own help file, shorter than the one for less(1).
The helpfile itself was prepared by jmc@, the glue by me.
OK millert@ jmc@

Merge Less 458 plus local patches

ok guenther@, ok jmc@ for man bits

Merge in less 444 plus local changes

ok nicm@

Use interruptible syscalls instead of setjmp/longjmp. This makes
less's signal handlers safe. No one has reported problems so far...

o go back to using a help file instead of embedding it in the binary
o deal with HELPFILE not being defined
o add a SMALL definition to make a smaller less for the boot floppies
This still needs to be pared down a bit for the SMALL case

Merge in less-381 w/ local changes and remove obsolete files.

$OpenBSD$

branches: 1.1.1;
Initial revision