Hi Michal,
The following patch removes the timestamp in .config file. This is a
prerequiresite for the revised version of the printer merge I sent back
in December[0]. The issue is that heading comment generations will be
merged in a single function, and that would revert the deterministic
change you made in mid-April.
The other options are either to have timestamp always enabled and
disable them selectively, but that might be a pain to implement, or to
have timestamp disable by default and an environment variable, say
KCONFIG_TIMESTAMP, to turn everything on manually.
Regards,
- Arnaud
[0]: https://patchwork.kernel.org/patch/380631/
--
From: Arnaud Lacombe <[email protected]>
Subject: kconfig: do not record timestamp in .config
Signed-off-by: Arnaud Lacombe <[email protected]>
---
Documentation/kbuild/kconfig.txt | 5 -----
scripts/kconfig/confdata.c | 12 +-----------
2 files changed, 1 insertions(+), 16 deletions(-)
diff --git a/Documentation/kbuild/kconfig.txt b/Documentation/kbuild/kconfig.txt
index cca46b1..c313d71 100644
--- a/Documentation/kbuild/kconfig.txt
+++ b/Documentation/kbuild/kconfig.txt
@@ -48,11 +48,6 @@ KCONFIG_OVERWRITECONFIG
If you set KCONFIG_OVERWRITECONFIG in the environment, Kconfig will not
break symlinks when .config is a symlink to somewhere else.
-KCONFIG_NOTIMESTAMP
---------------------------------------------------
-If this environment variable exists and is non-null, the timestamp line
-in generated .config files is omitted.
-
______________________________________________________________________
Environment variables for '{allyes/allmod/allno/rand}config'
diff --git a/scripts/kconfig/confdata.c b/scripts/kconfig/confdata.c
index 834eecb..2bafd9a 100644
--- a/scripts/kconfig/confdata.c
+++ b/scripts/kconfig/confdata.c
@@ -560,8 +560,6 @@ int conf_write(const char *name)
const char *basename;
const char *str;
char dirname[PATH_MAX+1], tmpname[PATH_MAX+1], newname[PATH_MAX+1];
- time_t now;
- int use_timestamp = 1;
char *env;
dirname[0] = 0;
@@ -598,19 +596,11 @@ int conf_write(const char *name)
if (!out)
return 1;
- time(&now);
- env = getenv("KCONFIG_NOTIMESTAMP");
- if (env && *env)
- use_timestamp = 0;
-
fprintf(out, _("#\n"
"# Automatically generated make config: don't edit\n"
"# %s\n"
- "%s%s"
"#\n"),
- rootmenu.prompt->text,
- use_timestamp ? "# " : "",
- use_timestamp ? ctime(&now) : "");
+ rootmenu.prompt->text);
if (!conf_get_changed())
sym_clear_all_valid();
--
1.7.3.4.574.g608b.dirty
On 16.5.2011 05:22, Arnaud Lacombe wrote:
> Hi Michal,
>
> The following patch removes the timestamp in .config file. This is a
> prerequiresite for the revised version of the printer merge I sent back
> in December[0]. The issue is that heading comment generations will be
> merged in a single function, and that would revert the deterministic
> change you made in mid-April.
>
> The other options are either to have timestamp always enabled and
> disable them selectively, but that might be a pain to implement, or to
> have timestamp disable by default and an environment variable, say
> KCONFIG_TIMESTAMP, to turn everything on manually.
I personally am not a big fan of embedded timestamps, so
Acked-by: Michal Marek <[email protected]>
But I'll wait some time before merging this, in case there are
objections from others...
Michal
>
> Regards,
> - Arnaud
>
> [0]: https://patchwork.kernel.org/patch/380631/
>
> --
>
> From: Arnaud Lacombe<[email protected]>
> Subject: kconfig: do not record timestamp in .config
>
> Signed-off-by: Arnaud Lacombe<[email protected]>
> ---
> Documentation/kbuild/kconfig.txt | 5 -----
> scripts/kconfig/confdata.c | 12 +-----------
> 2 files changed, 1 insertions(+), 16 deletions(-)
>
> diff --git a/Documentation/kbuild/kconfig.txt b/Documentation/kbuild/kconfig.txt
> index cca46b1..c313d71 100644
> --- a/Documentation/kbuild/kconfig.txt
> +++ b/Documentation/kbuild/kconfig.txt
> @@ -48,11 +48,6 @@ KCONFIG_OVERWRITECONFIG
> If you set KCONFIG_OVERWRITECONFIG in the environment, Kconfig will not
> break symlinks when .config is a symlink to somewhere else.
>
> -KCONFIG_NOTIMESTAMP
> ---------------------------------------------------
> -If this environment variable exists and is non-null, the timestamp line
> -in generated .config files is omitted.
> -
> ______________________________________________________________________
> Environment variables for '{allyes/allmod/allno/rand}config'
>
> diff --git a/scripts/kconfig/confdata.c b/scripts/kconfig/confdata.c
> index 834eecb..2bafd9a 100644
> --- a/scripts/kconfig/confdata.c
> +++ b/scripts/kconfig/confdata.c
> @@ -560,8 +560,6 @@ int conf_write(const char *name)
> const char *basename;
> const char *str;
> char dirname[PATH_MAX+1], tmpname[PATH_MAX+1], newname[PATH_MAX+1];
> - time_t now;
> - int use_timestamp = 1;
> char *env;
>
> dirname[0] = 0;
> @@ -598,19 +596,11 @@ int conf_write(const char *name)
> if (!out)
> return 1;
>
> - time(&now);
> - env = getenv("KCONFIG_NOTIMESTAMP");
> - if (env&& *env)
> - use_timestamp = 0;
> -
> fprintf(out, _("#\n"
> "# Automatically generated make config: don't edit\n"
> "# %s\n"
> - "%s%s"
> "#\n"),
> - rootmenu.prompt->text,
> - use_timestamp ? "# " : "",
> - use_timestamp ? ctime(&now) : "");
> + rootmenu.prompt->text);
>
> if (!conf_get_changed())
> sym_clear_all_valid();
Hi Michal,
2011/5/16 Michal Marek <[email protected]>:
> On 16.5.2011 05:22, Arnaud Lacombe wrote:
>>
>> Hi Michal,
>>
>> The following patch removes the timestamp in .config file. This is a
>> prerequiresite for the revised version of the printer merge I sent back
>> in December[0]. The issue is that heading comment generations will be
>> merged in a single function, and that would revert the deterministic
>> change you made in mid-April.
>>
>> The other options are either to have timestamp always enabled and
>> disable them selectively, but that might be a pain to implement, or to
>> have timestamp disable by default and an environment variable, say
>> KCONFIG_TIMESTAMP, to turn everything on manually.
>
> I personally am not a big fan of embedded timestamps, so
>
> Acked-by: Michal Marek <[email protected]>
>
Do you intend to merge that patch for .1 (or whatever Linus decides to
call the next kernel) ?
- Arnaud
On Tue, May 24, 2011 at 10:16:11AM -0400, Arnaud Lacombe wrote:
> Hi Michal,
>
> 2011/5/16 Michal Marek <[email protected]>:
> > On 16.5.2011 05:22, Arnaud Lacombe wrote:
> >>
> >> Hi Michal,
> >>
> >> The following patch removes the timestamp in .config file. This is a
> >> prerequiresite for the revised version of the printer merge I sent back
> >> in December[0]. The issue is that heading comment generations will be
> >> merged in a single function, and that would revert the deterministic
> >> change you made in mid-April.
> >>
> >> The other options are either to have timestamp always enabled and
> >> disable them selectively, but that might be a pain to implement, or to
> >> have timestamp disable by default and an environment variable, say
> >> KCONFIG_TIMESTAMP, to turn everything on manually.
> >
> > I personally am not a big fan of embedded timestamps, so
> >
> > Acked-by: Michal Marek <[email protected]>
> >
> Do you intend to merge that patch for .1 (or whatever Linus decides to
> call the next kernel) ?
I merged it now.
Michal