style: One more typo, NFC

style: Fix typo, NFC

"it it" --> "it is" (across a newline)

style(5): Advise against new struct typedefs and explain why.

Proposed on tech-kern:
https://mail-index.netbsd.org/tech-kern/2023/07/11/msg028950.html

Positive feedback to general concept, negative feedback to specifics
and phrasing of the first iteration but no objections to latest
iteration after several weeks at:
https://mail-index.netbsd.org/tech-kern/2023/07/16/msg028994.html

style: remove mentions of ANSI C

The distinction between traditional C and ISO C90 or later is no longer
needed in the style guide. Any old-style code is caught by the
compilers.

https://mail-index.netbsd.org/tech-userlevel/2023/04/15/msg013768.html
https://mail-index.netbsd.org/tech-userlevel/2023/04/17/msg013776.html

share/misc/style: Don't prescribe using "extern.h" for all prototypes.

That's very 80s.

No objection (or even response, amazingly enough) on tech-userlevel in
two weeks.

style: fix spelling of plural parentheses

style: don't require sorting variables in functions

Sorting the variables by size would be platform-dependent and thus is
not possible.

Sorting the variables alphabetically may or may not make the code easier
to read, and the example given below that rule doesn't follow it,
otherwise the correct order would be 'eight, eleven, nine, ten,
thirteen, twelve'.

https://mail-index.netbsd.org/tech-userlevel/2023/04/11/msg013749.html

style(5): Omit redundant proscription against .c externs in previous.

style(5): Forbid extern in .c files.

Discussed on tech-userlevel (bcc tech-kern):
https://mail-index.netbsd.org/tech-userlevel/2023/03/15/msg013727.html

exit with EXIT_FAILURE, not 1, as elsewhere

Move a style recommendation closer to the syntax that it describes.

C99: Suggest using narrow scopes for loop indices.

Terminate a sentence with a full stop.

share/misc/style: lint no longer needs CONSTCOND in do-while-loops

Since usr.bin/xlint/lint1/tree.c 1.202 from 2021-01-31, lint recognizes
the common idiom 'do { ... } while (0)' and no longer complains about
its constant condition; see tests/usr.bin/xlint/lint1/msg_161.c.

share/misc/style: clarify enum example code

The only purpose of the enum example is to show that enum constants are
written in uppercase.

The previous code could be mistakenly interpreted as requiring each
definition of an enum type to also define a global variable of that
type, and to make the variable name the abbreviated type name. Even if
the enum type were prefixed with 'typedef', choosing different names for
the tag and the typedef would not be necessary.

To avoid these possible misinterpretations above, remove the variable
name from the declaration.

Clarify and explain the rationale for parentheses in sizeof and return as
discussed.

- EXIT_FAILURE instead of 1
- %j instead of PRI for *intmax_t
- != -1 instead of < 0 for syscalls

style: fix typo

Remove confusing advice about macros that might expand to nothing.

Such macros should not exist; as advised earlier in the file, they
should expand to `__nothing' (literally) from <sys/cdefs.h> which
itself expands to ((void)0) precisely so that the problem this advice
sought to avoid does not arise.

Update style around single-line braces according to discussion.

https://mail-index.netbsd.org/tech-userlevel/2020/07/12/msg012536.html
https://mail-index.netbsd.org/tech-kern/2020/07/12/msg026594.html

Retain some examples of technically unnecessary braces that likely
aid legibility from the previous commit.

style: revert previous

I misintepreted the consensus.

style: prefer braces for single statement control statements

Prefer to use { braces } around single statements after
control statements, instead of discouraging them.

Per discussion on tech-userlevel & tech-kern, where the significant
majority of developers who responded (including current and former
core members) prefer this new style.

style: fix typo about enum constants being capitalized

document how we want to use __nothing

branches: 1.53.16;
fix spelling mistakes

Amend variable initialization in the declarations comment.

PR/47632: Supratim Chakraborty: Spelling Mistake for Variadic functions

branches: 1.50.2;
Elaborate and clarify rules about #include order.

ok christos

branches: 1.49.2;
explain NULL casts

Prototyping main is silly and shouldn't be encouraged.

make usage dead.

Be explicit about sizeof(*var) being preferred over sizeof(type).

Avoid trailing whitespace. It is definitely against the style rules.

branches: 1.44.4;
move punctuation of a complete sentence into its parens

Remove the \n from the __COPYRIGHT example; what(1) stops extracting
at them, so we lose most of the message.

branches: 1.42.4; 1.42.8;
Don't mention <machine/int_fmtio.h>; code should just
include <inttypes.h> to get intmax_t, PRIdMAX, and friends.

getopt uses : when an argument is expected

branches: 1.40.4;
Fix reversed test and missing +; from Joerg Niendorf

Add entry about making include files self-sufficient.

update brace comments to reflect current reality.

Mention the preferred option order in
getopt(), usage(), and man page.
Suggest describing newly added options in usage() and man page.

Add some comments about using inttypes.h / int_fmtio.h macros.
Show use of PRIdMAX, PRIxMAX, PRId64, PRIx32, and refer to SCN*.
Inspired by comments from Roland Illig and others.

Partly reverted my last change. I had forgotten that off_t is a signed
data type.

Don't cast unsigned values to long long. They might overflow.

Use exit(EXIT_FAILURE) instead of exit(1).

More clarification on uintN_t.

Make comment more specific.

fix accidental join that made the comment look like functions that
return ``void *'' should not be cast'ed.

Add missing (void) cast to printf return.
Prefer returning from main rather than calling exit.

Add uint32_t as example (so that it is implicitly clear we prefer the C99 types).

Use uintN_t in code example instead of u_intN_t. Agrees with current
standards (and more importantly Klaus too :-).

Recommend EXIT_SUCCESS and EXIT_FAILURE for exit status, instead of directly
using 0 and 1 values, respectively.

check strtol for ERANGE overflow too...

There's no need anymore to protect __COPYRIGHT() and __RCSID() with:
#ifndef __lint
...
#endif

Return value style update (no parentheses)

Amend rev 1.21: there is insufficient concensus on #define<TAB>.
Permit any whitespace after #define, but mandate consistency of
TABs within a file.

Clarify that macros should be #define<TAB>; it's important to specify this
in the actual document, just not in the cvs logs (cf. rev 1.15).

branches: 1.20.2;
Inspired by rev. 1.18: consistency, pull in <assert.h> for _DIAGASSERT().

As discussed with Luke, add a short _DIAGASSERT(3) example.
Fixes PR lib/14324.

add include of stdlib.h, for use of setprogname() & getprogname()

nuke __progname, suggest setprogname() and getprogname()

Fix typos.

- fix spelo
- be consistent with #define<TAB> for macros

add example on how to printf 64 bit quantities

fix comment around CONSTCOND

major overhaul:
* require ANSI not K&R for prototypes, stdargs, etc
* document current practice for
- rcsids & copyright
- header file inclusion protection
- expressions, and multi-statement macros
- enum & struct decls
- empty and large statements
- wrapping long lines, prototypes, function decls

Fix error in example - from VaX#n8 <vax@linkdead.paranoia.com>

Add a pointer to the indent(1) profile, and a disclaimer regarding same.

provide a better parenthesis example. from Brian Grayson in [misc/6714]

Function type declarations are manatory now.

Make C sentense end with `;'.

RCS id police. Side effect: shows the RCS id style we prefer :-)

getopt() returns -1 when the arg list is exhausted, not EOF. Ironic
that it should show up here...

Minor grammatical nit: "have too" -> "have to."

s/externs.h/extern.h/

style guide

overall source code style guide. it is king. revere it.

style: Fix typo, NFC

"it it" --> "it is" (across a newline)

style(5): Advise against new struct typedefs and explain why.

Proposed on tech-kern:
https://mail-index.netbsd.org/tech-kern/2023/07/11/msg028950.html

Positive feedback to general concept, negative feedback to specifics
and phrasing of the first iteration but no objections to latest
iteration after several weeks at:
https://mail-index.netbsd.org/tech-kern/2023/07/16/msg028994.html

style: remove mentions of ANSI C

The distinction between traditional C and ISO C90 or later is no longer
needed in the style guide. Any old-style code is caught by the
compilers.

https://mail-index.netbsd.org/tech-userlevel/2023/04/15/msg013768.html
https://mail-index.netbsd.org/tech-userlevel/2023/04/17/msg013776.html

share/misc/style: Don't prescribe using "extern.h" for all prototypes.

That's very 80s.

No objection (or even response, amazingly enough) on tech-userlevel in
two weeks.

style: fix spelling of plural parentheses

style: don't require sorting variables in functions

Sorting the variables by size would be platform-dependent and thus is
not possible.

Sorting the variables alphabetically may or may not make the code easier
to read, and the example given below that rule doesn't follow it,
otherwise the correct order would be 'eight, eleven, nine, ten,
thirteen, twelve'.

https://mail-index.netbsd.org/tech-userlevel/2023/04/11/msg013749.html

style(5): Omit redundant proscription against .c externs in previous.

style(5): Forbid extern in .c files.

Discussed on tech-userlevel (bcc tech-kern):
https://mail-index.netbsd.org/tech-userlevel/2023/03/15/msg013727.html

exit with EXIT_FAILURE, not 1, as elsewhere

Move a style recommendation closer to the syntax that it describes.

C99: Suggest using narrow scopes for loop indices.

Terminate a sentence with a full stop.

share/misc/style: lint no longer needs CONSTCOND in do-while-loops

Since usr.bin/xlint/lint1/tree.c 1.202 from 2021-01-31, lint recognizes
the common idiom 'do { ... } while (0)' and no longer complains about
its constant condition; see tests/usr.bin/xlint/lint1/msg_161.c.

share/misc/style: clarify enum example code

The only purpose of the enum example is to show that enum constants are
written in uppercase.

The previous code could be mistakenly interpreted as requiring each
definition of an enum type to also define a global variable of that
type, and to make the variable name the abbreviated type name. Even if
the enum type were prefixed with 'typedef', choosing different names for
the tag and the typedef would not be necessary.

To avoid these possible misinterpretations above, remove the variable
name from the declaration.

Clarify and explain the rationale for parentheses in sizeof and return as
discussed.

- EXIT_FAILURE instead of 1
- %j instead of PRI for *intmax_t
- != -1 instead of < 0 for syscalls

style: fix typo

Remove confusing advice about macros that might expand to nothing.

Such macros should not exist; as advised earlier in the file, they
should expand to `__nothing' (literally) from <sys/cdefs.h> which
itself expands to ((void)0) precisely so that the problem this advice
sought to avoid does not arise.

Update style around single-line braces according to discussion.

https://mail-index.netbsd.org/tech-userlevel/2020/07/12/msg012536.html
https://mail-index.netbsd.org/tech-kern/2020/07/12/msg026594.html

Retain some examples of technically unnecessary braces that likely
aid legibility from the previous commit.

style: revert previous

I misintepreted the consensus.

style: prefer braces for single statement control statements

Prefer to use { braces } around single statements after
control statements, instead of discouraging them.

Per discussion on tech-userlevel & tech-kern, where the significant
majority of developers who responded (including current and former
core members) prefer this new style.

style: fix typo about enum constants being capitalized

document how we want to use __nothing

branches: 1.53.16;
fix spelling mistakes

Amend variable initialization in the declarations comment.

PR/47632: Supratim Chakraborty: Spelling Mistake for Variadic functions

branches: 1.50.2;
Elaborate and clarify rules about #include order.

ok christos

branches: 1.49.2;
explain NULL casts

Prototyping main is silly and shouldn't be encouraged.

make usage dead.

Be explicit about sizeof(*var) being preferred over sizeof(type).

Avoid trailing whitespace. It is definitely against the style rules.

branches: 1.44.4;
move punctuation of a complete sentence into its parens

Remove the \n from the __COPYRIGHT example; what(1) stops extracting
at them, so we lose most of the message.

branches: 1.42.4; 1.42.8;
Don't mention <machine/int_fmtio.h>; code should just
include <inttypes.h> to get intmax_t, PRIdMAX, and friends.

getopt uses : when an argument is expected

branches: 1.40.4;
Fix reversed test and missing +; from Joerg Niendorf

Add entry about making include files self-sufficient.

update brace comments to reflect current reality.

Mention the preferred option order in
getopt(), usage(), and man page.
Suggest describing newly added options in usage() and man page.

Add some comments about using inttypes.h / int_fmtio.h macros.
Show use of PRIdMAX, PRIxMAX, PRId64, PRIx32, and refer to SCN*.
Inspired by comments from Roland Illig and others.

Partly reverted my last change. I had forgotten that off_t is a signed
data type.

Don't cast unsigned values to long long. They might overflow.

Use exit(EXIT_FAILURE) instead of exit(1).

More clarification on uintN_t.

Make comment more specific.

fix accidental join that made the comment look like functions that
return ``void *'' should not be cast'ed.

Add missing (void) cast to printf return.
Prefer returning from main rather than calling exit.

Add uint32_t as example (so that it is implicitly clear we prefer the C99 types).

Use uintN_t in code example instead of u_intN_t. Agrees with current
standards (and more importantly Klaus too :-).

Recommend EXIT_SUCCESS and EXIT_FAILURE for exit status, instead of directly
using 0 and 1 values, respectively.

check strtol for ERANGE overflow too...

There's no need anymore to protect __COPYRIGHT() and __RCSID() with:
#ifndef __lint
...
#endif

Return value style update (no parentheses)

Amend rev 1.21: there is insufficient concensus on #define<TAB>.
Permit any whitespace after #define, but mandate consistency of
TABs within a file.

Clarify that macros should be #define<TAB>; it's important to specify this
in the actual document, just not in the cvs logs (cf. rev 1.15).

branches: 1.20.2;
Inspired by rev. 1.18: consistency, pull in <assert.h> for _DIAGASSERT().

As discussed with Luke, add a short _DIAGASSERT(3) example.
Fixes PR lib/14324.

add include of stdlib.h, for use of setprogname() & getprogname()

nuke __progname, suggest setprogname() and getprogname()

Fix typos.

- fix spelo
- be consistent with #define<TAB> for macros

add example on how to printf 64 bit quantities

fix comment around CONSTCOND

major overhaul:
* require ANSI not K&R for prototypes, stdargs, etc
* document current practice for
- rcsids & copyright
- header file inclusion protection
- expressions, and multi-statement macros
- enum & struct decls
- empty and large statements
- wrapping long lines, prototypes, function decls

Fix error in example - from VaX#n8 <vax@linkdead.paranoia.com>

Add a pointer to the indent(1) profile, and a disclaimer regarding same.

provide a better parenthesis example. from Brian Grayson in [misc/6714]

Function type declarations are manatory now.

Make C sentense end with `;'.

RCS id police. Side effect: shows the RCS id style we prefer :-)

getopt() returns -1 when the arg list is exhausted, not EOF. Ironic
that it should show up here...

Minor grammatical nit: "have too" -> "have to."

s/externs.h/extern.h/

style guide

overall source code style guide. it is king. revere it.

style(5): Advise against new struct typedefs and explain why.

Proposed on tech-kern:
https://mail-index.netbsd.org/tech-kern/2023/07/11/msg028950.html

Positive feedback to general concept, negative feedback to specifics
and phrasing of the first iteration but no objections to latest
iteration after several weeks at:
https://mail-index.netbsd.org/tech-kern/2023/07/16/msg028994.html

style: remove mentions of ANSI C

The distinction between traditional C and ISO C90 or later is no longer
needed in the style guide. Any old-style code is caught by the
compilers.

https://mail-index.netbsd.org/tech-userlevel/2023/04/15/msg013768.html
https://mail-index.netbsd.org/tech-userlevel/2023/04/17/msg013776.html

share/misc/style: Don't prescribe using "extern.h" for all prototypes.

That's very 80s.

No objection (or even response, amazingly enough) on tech-userlevel in
two weeks.

style: fix spelling of plural parentheses

style: don't require sorting variables in functions

Sorting the variables by size would be platform-dependent and thus is
not possible.

Sorting the variables alphabetically may or may not make the code easier
to read, and the example given below that rule doesn't follow it,
otherwise the correct order would be 'eight, eleven, nine, ten,
thirteen, twelve'.

https://mail-index.netbsd.org/tech-userlevel/2023/04/11/msg013749.html

style(5): Omit redundant proscription against .c externs in previous.

style(5): Forbid extern in .c files.

Discussed on tech-userlevel (bcc tech-kern):
https://mail-index.netbsd.org/tech-userlevel/2023/03/15/msg013727.html

exit with EXIT_FAILURE, not 1, as elsewhere

Move a style recommendation closer to the syntax that it describes.

C99: Suggest using narrow scopes for loop indices.

Terminate a sentence with a full stop.

share/misc/style: lint no longer needs CONSTCOND in do-while-loops

Since usr.bin/xlint/lint1/tree.c 1.202 from 2021-01-31, lint recognizes
the common idiom 'do { ... } while (0)' and no longer complains about
its constant condition; see tests/usr.bin/xlint/lint1/msg_161.c.

share/misc/style: clarify enum example code

The only purpose of the enum example is to show that enum constants are
written in uppercase.

The previous code could be mistakenly interpreted as requiring each
definition of an enum type to also define a global variable of that
type, and to make the variable name the abbreviated type name. Even if
the enum type were prefixed with 'typedef', choosing different names for
the tag and the typedef would not be necessary.

To avoid these possible misinterpretations above, remove the variable
name from the declaration.

Clarify and explain the rationale for parentheses in sizeof and return as
discussed.

- EXIT_FAILURE instead of 1
- %j instead of PRI for *intmax_t
- != -1 instead of < 0 for syscalls

style: fix typo

Remove confusing advice about macros that might expand to nothing.

Such macros should not exist; as advised earlier in the file, they
should expand to `__nothing' (literally) from <sys/cdefs.h> which
itself expands to ((void)0) precisely so that the problem this advice
sought to avoid does not arise.

Update style around single-line braces according to discussion.

https://mail-index.netbsd.org/tech-userlevel/2020/07/12/msg012536.html
https://mail-index.netbsd.org/tech-kern/2020/07/12/msg026594.html

Retain some examples of technically unnecessary braces that likely
aid legibility from the previous commit.

style: revert previous

I misintepreted the consensus.

style: prefer braces for single statement control statements

Prefer to use { braces } around single statements after
control statements, instead of discouraging them.

Per discussion on tech-userlevel & tech-kern, where the significant
majority of developers who responded (including current and former
core members) prefer this new style.

style: fix typo about enum constants being capitalized

document how we want to use __nothing

branches: 1.53.16;
fix spelling mistakes

Amend variable initialization in the declarations comment.

PR/47632: Supratim Chakraborty: Spelling Mistake for Variadic functions

branches: 1.50.2;
Elaborate and clarify rules about #include order.

ok christos

branches: 1.49.2;
explain NULL casts

Prototyping main is silly and shouldn't be encouraged.

make usage dead.

Be explicit about sizeof(*var) being preferred over sizeof(type).

Avoid trailing whitespace. It is definitely against the style rules.

branches: 1.44.4;
move punctuation of a complete sentence into its parens

Remove the \n from the __COPYRIGHT example; what(1) stops extracting
at them, so we lose most of the message.

branches: 1.42.4; 1.42.8;
Don't mention <machine/int_fmtio.h>; code should just
include <inttypes.h> to get intmax_t, PRIdMAX, and friends.

getopt uses : when an argument is expected

branches: 1.40.4;
Fix reversed test and missing +; from Joerg Niendorf

Add entry about making include files self-sufficient.

update brace comments to reflect current reality.

Mention the preferred option order in
getopt(), usage(), and man page.
Suggest describing newly added options in usage() and man page.

Add some comments about using inttypes.h / int_fmtio.h macros.
Show use of PRIdMAX, PRIxMAX, PRId64, PRIx32, and refer to SCN*.
Inspired by comments from Roland Illig and others.

Partly reverted my last change. I had forgotten that off_t is a signed
data type.

Don't cast unsigned values to long long. They might overflow.

Use exit(EXIT_FAILURE) instead of exit(1).

More clarification on uintN_t.

Make comment more specific.

fix accidental join that made the comment look like functions that
return ``void *'' should not be cast'ed.

Add missing (void) cast to printf return.
Prefer returning from main rather than calling exit.

Add uint32_t as example (so that it is implicitly clear we prefer the C99 types).

Use uintN_t in code example instead of u_intN_t. Agrees with current
standards (and more importantly Klaus too :-).

Recommend EXIT_SUCCESS and EXIT_FAILURE for exit status, instead of directly
using 0 and 1 values, respectively.

check strtol for ERANGE overflow too...

There's no need anymore to protect __COPYRIGHT() and __RCSID() with:
#ifndef __lint
...
#endif

Return value style update (no parentheses)

Amend rev 1.21: there is insufficient concensus on #define<TAB>.
Permit any whitespace after #define, but mandate consistency of
TABs within a file.

Clarify that macros should be #define<TAB>; it's important to specify this
in the actual document, just not in the cvs logs (cf. rev 1.15).

branches: 1.20.2;
Inspired by rev. 1.18: consistency, pull in <assert.h> for _DIAGASSERT().

As discussed with Luke, add a short _DIAGASSERT(3) example.
Fixes PR lib/14324.

add include of stdlib.h, for use of setprogname() & getprogname()

nuke __progname, suggest setprogname() and getprogname()

Fix typos.

- fix spelo
- be consistent with #define<TAB> for macros

add example on how to printf 64 bit quantities

fix comment around CONSTCOND

major overhaul:
* require ANSI not K&R for prototypes, stdargs, etc
* document current practice for
- rcsids & copyright
- header file inclusion protection
- expressions, and multi-statement macros
- enum & struct decls
- empty and large statements
- wrapping long lines, prototypes, function decls

Fix error in example - from VaX#n8 <vax@linkdead.paranoia.com>

Add a pointer to the indent(1) profile, and a disclaimer regarding same.

provide a better parenthesis example. from Brian Grayson in [misc/6714]

Function type declarations are manatory now.

Make C sentense end with `;'.

RCS id police. Side effect: shows the RCS id style we prefer :-)

getopt() returns -1 when the arg list is exhausted, not EOF. Ironic
that it should show up here...

Minor grammatical nit: "have too" -> "have to."

s/externs.h/extern.h/

style guide

overall source code style guide. it is king. revere it.

style: remove mentions of ANSI C

The distinction between traditional C and ISO C90 or later is no longer
needed in the style guide. Any old-style code is caught by the
compilers.

https://mail-index.netbsd.org/tech-userlevel/2023/04/15/msg013768.html
https://mail-index.netbsd.org/tech-userlevel/2023/04/17/msg013776.html

share/misc/style: Don't prescribe using "extern.h" for all prototypes.

That's very 80s.

No objection (or even response, amazingly enough) on tech-userlevel in
two weeks.

style: fix spelling of plural parentheses

style: don't require sorting variables in functions

Sorting the variables by size would be platform-dependent and thus is
not possible.

Sorting the variables alphabetically may or may not make the code easier
to read, and the example given below that rule doesn't follow it,
otherwise the correct order would be 'eight, eleven, nine, ten,
thirteen, twelve'.

https://mail-index.netbsd.org/tech-userlevel/2023/04/11/msg013749.html

style(5): Omit redundant proscription against .c externs in previous.

style(5): Forbid extern in .c files.

Discussed on tech-userlevel (bcc tech-kern):
https://mail-index.netbsd.org/tech-userlevel/2023/03/15/msg013727.html

exit with EXIT_FAILURE, not 1, as elsewhere

Move a style recommendation closer to the syntax that it describes.

C99: Suggest using narrow scopes for loop indices.

Terminate a sentence with a full stop.

share/misc/style: lint no longer needs CONSTCOND in do-while-loops

Since usr.bin/xlint/lint1/tree.c 1.202 from 2021-01-31, lint recognizes
the common idiom 'do { ... } while (0)' and no longer complains about
its constant condition; see tests/usr.bin/xlint/lint1/msg_161.c.

share/misc/style: clarify enum example code

The only purpose of the enum example is to show that enum constants are
written in uppercase.

The previous code could be mistakenly interpreted as requiring each
definition of an enum type to also define a global variable of that
type, and to make the variable name the abbreviated type name. Even if
the enum type were prefixed with 'typedef', choosing different names for
the tag and the typedef would not be necessary.

To avoid these possible misinterpretations above, remove the variable
name from the declaration.

Clarify and explain the rationale for parentheses in sizeof and return as
discussed.

- EXIT_FAILURE instead of 1
- %j instead of PRI for *intmax_t
- != -1 instead of < 0 for syscalls

style: fix typo

Remove confusing advice about macros that might expand to nothing.

Such macros should not exist; as advised earlier in the file, they
should expand to `__nothing' (literally) from <sys/cdefs.h> which
itself expands to ((void)0) precisely so that the problem this advice
sought to avoid does not arise.

Update style around single-line braces according to discussion.

https://mail-index.netbsd.org/tech-userlevel/2020/07/12/msg012536.html
https://mail-index.netbsd.org/tech-kern/2020/07/12/msg026594.html

Retain some examples of technically unnecessary braces that likely
aid legibility from the previous commit.

style: revert previous

I misintepreted the consensus.

style: prefer braces for single statement control statements

Prefer to use { braces } around single statements after
control statements, instead of discouraging them.

Per discussion on tech-userlevel & tech-kern, where the significant
majority of developers who responded (including current and former
core members) prefer this new style.

style: fix typo about enum constants being capitalized

document how we want to use __nothing

branches: 1.53.16;
fix spelling mistakes

Amend variable initialization in the declarations comment.

PR/47632: Supratim Chakraborty: Spelling Mistake for Variadic functions

branches: 1.50.2;
Elaborate and clarify rules about #include order.

ok christos

branches: 1.49.2;
explain NULL casts

Prototyping main is silly and shouldn't be encouraged.

make usage dead.

Be explicit about sizeof(*var) being preferred over sizeof(type).

Avoid trailing whitespace. It is definitely against the style rules.

branches: 1.44.4;
move punctuation of a complete sentence into its parens

Remove the \n from the __COPYRIGHT example; what(1) stops extracting
at them, so we lose most of the message.

branches: 1.42.4; 1.42.8;
Don't mention <machine/int_fmtio.h>; code should just
include <inttypes.h> to get intmax_t, PRIdMAX, and friends.

getopt uses : when an argument is expected

branches: 1.40.4;
Fix reversed test and missing +; from Joerg Niendorf

Add entry about making include files self-sufficient.

update brace comments to reflect current reality.

Mention the preferred option order in
getopt(), usage(), and man page.
Suggest describing newly added options in usage() and man page.

Add some comments about using inttypes.h / int_fmtio.h macros.
Show use of PRIdMAX, PRIxMAX, PRId64, PRIx32, and refer to SCN*.
Inspired by comments from Roland Illig and others.

Partly reverted my last change. I had forgotten that off_t is a signed
data type.

Don't cast unsigned values to long long. They might overflow.

Use exit(EXIT_FAILURE) instead of exit(1).

More clarification on uintN_t.

Make comment more specific.

fix accidental join that made the comment look like functions that
return ``void *'' should not be cast'ed.

Add missing (void) cast to printf return.
Prefer returning from main rather than calling exit.

Add uint32_t as example (so that it is implicitly clear we prefer the C99 types).

Use uintN_t in code example instead of u_intN_t. Agrees with current
standards (and more importantly Klaus too :-).

Recommend EXIT_SUCCESS and EXIT_FAILURE for exit status, instead of directly
using 0 and 1 values, respectively.

check strtol for ERANGE overflow too...

There's no need anymore to protect __COPYRIGHT() and __RCSID() with:
#ifndef __lint
...
#endif

Return value style update (no parentheses)

Amend rev 1.21: there is insufficient concensus on #define<TAB>.
Permit any whitespace after #define, but mandate consistency of
TABs within a file.

Clarify that macros should be #define<TAB>; it's important to specify this
in the actual document, just not in the cvs logs (cf. rev 1.15).

branches: 1.20.2;
Inspired by rev. 1.18: consistency, pull in <assert.h> for _DIAGASSERT().

As discussed with Luke, add a short _DIAGASSERT(3) example.
Fixes PR lib/14324.

add include of stdlib.h, for use of setprogname() & getprogname()

nuke __progname, suggest setprogname() and getprogname()

Fix typos.

- fix spelo
- be consistent with #define<TAB> for macros

add example on how to printf 64 bit quantities

fix comment around CONSTCOND

major overhaul:
* require ANSI not K&R for prototypes, stdargs, etc
* document current practice for
- rcsids & copyright
- header file inclusion protection
- expressions, and multi-statement macros
- enum & struct decls
- empty and large statements
- wrapping long lines, prototypes, function decls

Fix error in example - from VaX#n8 <vax@linkdead.paranoia.com>

Add a pointer to the indent(1) profile, and a disclaimer regarding same.

provide a better parenthesis example. from Brian Grayson in [misc/6714]

Function type declarations are manatory now.

Make C sentense end with `;'.

RCS id police. Side effect: shows the RCS id style we prefer :-)

getopt() returns -1 when the arg list is exhausted, not EOF. Ironic
that it should show up here...

Minor grammatical nit: "have too" -> "have to."

s/externs.h/extern.h/

style guide

overall source code style guide. it is king. revere it.

share/misc/style: Don't prescribe using "extern.h" for all prototypes.

That's very 80s.

No objection (or even response, amazingly enough) on tech-userlevel in
two weeks.

style: fix spelling of plural parentheses

style: don't require sorting variables in functions

Sorting the variables by size would be platform-dependent and thus is
not possible.

Sorting the variables alphabetically may or may not make the code easier
to read, and the example given below that rule doesn't follow it,
otherwise the correct order would be 'eight, eleven, nine, ten,
thirteen, twelve'.

https://mail-index.netbsd.org/tech-userlevel/2023/04/11/msg013749.html

style(5): Omit redundant proscription against .c externs in previous.

style(5): Forbid extern in .c files.

Discussed on tech-userlevel (bcc tech-kern):
https://mail-index.netbsd.org/tech-userlevel/2023/03/15/msg013727.html

exit with EXIT_FAILURE, not 1, as elsewhere

Move a style recommendation closer to the syntax that it describes.

C99: Suggest using narrow scopes for loop indices.

Terminate a sentence with a full stop.

share/misc/style: lint no longer needs CONSTCOND in do-while-loops

Since usr.bin/xlint/lint1/tree.c 1.202 from 2021-01-31, lint recognizes
the common idiom 'do { ... } while (0)' and no longer complains about
its constant condition; see tests/usr.bin/xlint/lint1/msg_161.c.

share/misc/style: clarify enum example code

The only purpose of the enum example is to show that enum constants are
written in uppercase.

The previous code could be mistakenly interpreted as requiring each
definition of an enum type to also define a global variable of that
type, and to make the variable name the abbreviated type name. Even if
the enum type were prefixed with 'typedef', choosing different names for
the tag and the typedef would not be necessary.

To avoid these possible misinterpretations above, remove the variable
name from the declaration.

Clarify and explain the rationale for parentheses in sizeof and return as
discussed.

- EXIT_FAILURE instead of 1
- %j instead of PRI for *intmax_t
- != -1 instead of < 0 for syscalls

style: fix typo

Remove confusing advice about macros that might expand to nothing.

Such macros should not exist; as advised earlier in the file, they
should expand to `__nothing' (literally) from <sys/cdefs.h> which
itself expands to ((void)0) precisely so that the problem this advice
sought to avoid does not arise.

Update style around single-line braces according to discussion.

https://mail-index.netbsd.org/tech-userlevel/2020/07/12/msg012536.html
https://mail-index.netbsd.org/tech-kern/2020/07/12/msg026594.html

Retain some examples of technically unnecessary braces that likely
aid legibility from the previous commit.

style: revert previous

I misintepreted the consensus.

style: prefer braces for single statement control statements

Prefer to use { braces } around single statements after
control statements, instead of discouraging them.

Per discussion on tech-userlevel & tech-kern, where the significant
majority of developers who responded (including current and former
core members) prefer this new style.

style: fix typo about enum constants being capitalized

document how we want to use __nothing

branches: 1.53.16;
fix spelling mistakes

Amend variable initialization in the declarations comment.

PR/47632: Supratim Chakraborty: Spelling Mistake for Variadic functions

branches: 1.50.2;
Elaborate and clarify rules about #include order.

ok christos

branches: 1.49.2;
explain NULL casts

Prototyping main is silly and shouldn't be encouraged.

make usage dead.

Be explicit about sizeof(*var) being preferred over sizeof(type).

Avoid trailing whitespace. It is definitely against the style rules.

branches: 1.44.4;
move punctuation of a complete sentence into its parens

Remove the \n from the __COPYRIGHT example; what(1) stops extracting
at them, so we lose most of the message.

branches: 1.42.4; 1.42.8;
Don't mention <machine/int_fmtio.h>; code should just
include <inttypes.h> to get intmax_t, PRIdMAX, and friends.

getopt uses : when an argument is expected

branches: 1.40.4;
Fix reversed test and missing +; from Joerg Niendorf

Add entry about making include files self-sufficient.

update brace comments to reflect current reality.

Mention the preferred option order in
getopt(), usage(), and man page.
Suggest describing newly added options in usage() and man page.

Add some comments about using inttypes.h / int_fmtio.h macros.
Show use of PRIdMAX, PRIxMAX, PRId64, PRIx32, and refer to SCN*.
Inspired by comments from Roland Illig and others.

Partly reverted my last change. I had forgotten that off_t is a signed
data type.

Don't cast unsigned values to long long. They might overflow.

Use exit(EXIT_FAILURE) instead of exit(1).

More clarification on uintN_t.

Make comment more specific.

fix accidental join that made the comment look like functions that
return ``void *'' should not be cast'ed.

Add missing (void) cast to printf return.
Prefer returning from main rather than calling exit.

Add uint32_t as example (so that it is implicitly clear we prefer the C99 types).

Use uintN_t in code example instead of u_intN_t. Agrees with current
standards (and more importantly Klaus too :-).

Recommend EXIT_SUCCESS and EXIT_FAILURE for exit status, instead of directly
using 0 and 1 values, respectively.

check strtol for ERANGE overflow too...

There's no need anymore to protect __COPYRIGHT() and __RCSID() with:
#ifndef __lint
...
#endif

Return value style update (no parentheses)

Amend rev 1.21: there is insufficient concensus on #define<TAB>.
Permit any whitespace after #define, but mandate consistency of
TABs within a file.

Clarify that macros should be #define<TAB>; it's important to specify this
in the actual document, just not in the cvs logs (cf. rev 1.15).

branches: 1.20.2;
Inspired by rev. 1.18: consistency, pull in <assert.h> for _DIAGASSERT().

As discussed with Luke, add a short _DIAGASSERT(3) example.
Fixes PR lib/14324.

add include of stdlib.h, for use of setprogname() & getprogname()

nuke __progname, suggest setprogname() and getprogname()

Fix typos.

- fix spelo
- be consistent with #define<TAB> for macros

add example on how to printf 64 bit quantities

fix comment around CONSTCOND

major overhaul:
* require ANSI not K&R for prototypes, stdargs, etc
* document current practice for
- rcsids & copyright
- header file inclusion protection
- expressions, and multi-statement macros
- enum & struct decls
- empty and large statements
- wrapping long lines, prototypes, function decls

Fix error in example - from VaX#n8 <vax@linkdead.paranoia.com>

Add a pointer to the indent(1) profile, and a disclaimer regarding same.

provide a better parenthesis example. from Brian Grayson in [misc/6714]

Function type declarations are manatory now.

Make C sentense end with `;'.

RCS id police. Side effect: shows the RCS id style we prefer :-)

getopt() returns -1 when the arg list is exhausted, not EOF. Ironic
that it should show up here...

Minor grammatical nit: "have too" -> "have to."

s/externs.h/extern.h/

style guide

overall source code style guide. it is king. revere it.

exit with EXIT_FAILURE, not 1, as elsewhere

Move a style recommendation closer to the syntax that it describes.

C99: Suggest using narrow scopes for loop indices.

Terminate a sentence with a full stop.

share/misc/style: lint no longer needs CONSTCOND in do-while-loops

Since usr.bin/xlint/lint1/tree.c 1.202 from 2021-01-31, lint recognizes
the common idiom 'do { ... } while (0)' and no longer complains about
its constant condition; see tests/usr.bin/xlint/lint1/msg_161.c.

share/misc/style: clarify enum example code

The only purpose of the enum example is to show that enum constants are
written in uppercase.

The previous code could be mistakenly interpreted as requiring each
definition of an enum type to also define a global variable of that
type, and to make the variable name the abbreviated type name. Even if
the enum type were prefixed with 'typedef', choosing different names for
the tag and the typedef would not be necessary.

To avoid these possible misinterpretations above, remove the variable
name from the declaration.

Clarify and explain the rationale for parentheses in sizeof and return as
discussed.

- EXIT_FAILURE instead of 1
- %j instead of PRI for *intmax_t
- != -1 instead of < 0 for syscalls

style: fix typo

Remove confusing advice about macros that might expand to nothing.

Such macros should not exist; as advised earlier in the file, they
should expand to `__nothing' (literally) from <sys/cdefs.h> which
itself expands to ((void)0) precisely so that the problem this advice
sought to avoid does not arise.

Update style around single-line braces according to discussion.

https://mail-index.netbsd.org/tech-userlevel/2020/07/12/msg012536.html
https://mail-index.netbsd.org/tech-kern/2020/07/12/msg026594.html

Retain some examples of technically unnecessary braces that likely
aid legibility from the previous commit.

style: revert previous

I misintepreted the consensus.

style: prefer braces for single statement control statements

Prefer to use { braces } around single statements after
control statements, instead of discouraging them.

Per discussion on tech-userlevel & tech-kern, where the significant
majority of developers who responded (including current and former
core members) prefer this new style.

style: fix typo about enum constants being capitalized

document how we want to use __nothing

branches: 1.53.16;
fix spelling mistakes

Amend variable initialization in the declarations comment.

PR/47632: Supratim Chakraborty: Spelling Mistake for Variadic functions

branches: 1.50.2;
Elaborate and clarify rules about #include order.

ok christos

branches: 1.49.2;
explain NULL casts

Prototyping main is silly and shouldn't be encouraged.

make usage dead.

Be explicit about sizeof(*var) being preferred over sizeof(type).

Avoid trailing whitespace. It is definitely against the style rules.

branches: 1.44.4;
move punctuation of a complete sentence into its parens

Remove the \n from the __COPYRIGHT example; what(1) stops extracting
at them, so we lose most of the message.

branches: 1.42.4; 1.42.8;
Don't mention <machine/int_fmtio.h>; code should just
include <inttypes.h> to get intmax_t, PRIdMAX, and friends.

getopt uses : when an argument is expected

branches: 1.40.4;
Fix reversed test and missing +; from Joerg Niendorf

Add entry about making include files self-sufficient.

update brace comments to reflect current reality.

Mention the preferred option order in
getopt(), usage(), and man page.
Suggest describing newly added options in usage() and man page.

Add some comments about using inttypes.h / int_fmtio.h macros.
Show use of PRIdMAX, PRIxMAX, PRId64, PRIx32, and refer to SCN*.
Inspired by comments from Roland Illig and others.

Partly reverted my last change. I had forgotten that off_t is a signed
data type.

Don't cast unsigned values to long long. They might overflow.

Use exit(EXIT_FAILURE) instead of exit(1).

More clarification on uintN_t.

Make comment more specific.

fix accidental join that made the comment look like functions that
return ``void *'' should not be cast'ed.

Add missing (void) cast to printf return.
Prefer returning from main rather than calling exit.

Add uint32_t as example (so that it is implicitly clear we prefer the C99 types).

Use uintN_t in code example instead of u_intN_t. Agrees with current
standards (and more importantly Klaus too :-).

Recommend EXIT_SUCCESS and EXIT_FAILURE for exit status, instead of directly
using 0 and 1 values, respectively.

check strtol for ERANGE overflow too...

There's no need anymore to protect __COPYRIGHT() and __RCSID() with:
#ifndef __lint
...
#endif

Return value style update (no parentheses)

Amend rev 1.21: there is insufficient concensus on #define<TAB>.
Permit any whitespace after #define, but mandate consistency of
TABs within a file.

Clarify that macros should be #define<TAB>; it's important to specify this
in the actual document, just not in the cvs logs (cf. rev 1.15).

branches: 1.20.2;
Inspired by rev. 1.18: consistency, pull in <assert.h> for _DIAGASSERT().

As discussed with Luke, add a short _DIAGASSERT(3) example.
Fixes PR lib/14324.

add include of stdlib.h, for use of setprogname() & getprogname()

nuke __progname, suggest setprogname() and getprogname()

Fix typos.

- fix spelo
- be consistent with #define<TAB> for macros

add example on how to printf 64 bit quantities

fix comment around CONSTCOND

major overhaul:
* require ANSI not K&R for prototypes, stdargs, etc
* document current practice for
- rcsids & copyright
- header file inclusion protection
- expressions, and multi-statement macros
- enum & struct decls
- empty and large statements
- wrapping long lines, prototypes, function decls

Fix error in example - from VaX#n8 <vax@linkdead.paranoia.com>

Add a pointer to the indent(1) profile, and a disclaimer regarding same.

provide a better parenthesis example. from Brian Grayson in [misc/6714]

Function type declarations are manatory now.

Make C sentense end with `;'.

RCS id police. Side effect: shows the RCS id style we prefer :-)

getopt() returns -1 when the arg list is exhausted, not EOF. Ironic
that it should show up here...

Minor grammatical nit: "have too" -> "have to."

s/externs.h/extern.h/

style guide

overall source code style guide. it is king. revere it.

Move a style recommendation closer to the syntax that it describes.

C99: Suggest using narrow scopes for loop indices.

Terminate a sentence with a full stop.

share/misc/style: lint no longer needs CONSTCOND in do-while-loops

Since usr.bin/xlint/lint1/tree.c 1.202 from 2021-01-31, lint recognizes
the common idiom 'do { ... } while (0)' and no longer complains about
its constant condition; see tests/usr.bin/xlint/lint1/msg_161.c.