[Buildroot] [PATCH 0/3] expect: bump to version 5.45.3

[Buildroot] [PATCH 0/3] expect: bump to version 5.45.3

[Fabio Porcedda]

This patch set update the package expect and because the version
5.45.3 version is aviable only in cvs without a tag add the
possibility to use a date to download from cvs.

Best regards

Fabio Porcedda (3):
  support/download/cvs: add support to use a date instead of a tag
  manual: cvs: document that a date can be used instead of a tag
  expect: bump to version 5.45.3

 docs/manual/adding-packages-generic.txt |  2 +-
 package/expect/expect.hash              |  2 --
 package/expect/expect.mk                |  5 ++---
 support/download/cvs                    | 10 +++++++++-
 4 files changed, 12 insertions(+), 7 deletions(-)
 delete mode 100644 package/expect/expect.hash

-- 
2.3.5

[Buildroot] [PATCH 1/3] support/download/cvs: add support to use a date instead of a tag

[Fabio Porcedda]

This is useful when a tag is not avaiable.

Signed-off-by: Fabio Porcedda <fabio.porcedda@gmail.com>
---
 support/download/cvs | 10 +++++++++-
 1 file changed, 9 insertions(+), 1 deletion(-)

diff --git a/support/download/cvs b/support/download/cvs
index 2c3a666..6491cc8 100755
--- a/support/download/cvs
+++ b/support/download/cvs
@@ -26,7 +26,15 @@ rev="${3}"
 rawname="${4}"
 basename="${5}"
 
+if [[ ${rev} =~ ^[0-9] ]]; then
+    # Date, because a tag cannot begin with a number
+    select="-D ${rev}"
+else
+    # Tag
+    select=-"r :${rev}"
+fi
+
 ${CVS} ${verbose} -z3 -d":pserver:anonymous@${repo}" \
-       co -d "${basename}" -r ":${rev}" -P "${rawname}"
+       co -d "${basename}" ${select} -P "${rawname}"
 
 tar czf "${output}" "${basename}"
-- 
2.3.5

[Buildroot] [PATCH 1/3] support/download/cvs: add support to use a date instead of a tag

[Yann E. MORIN]

Fabio, All,

On 2015-04-18 18:54 +0200, Fabio Porcedda spake thusly:
> This is useful when a tag is not avaiable.
> 
> Signed-off-by: Fabio Porcedda <fabio.porcedda@gmail.com>

Acked-by: "Yann E. MORIN" <yann.morin.1998@free.fr>

Although I'm not a big fan of the [[...]] bashism, that's OK since it
makes things so much simpler.

Regards,
Yann E. MORIN.

> ---
>  support/download/cvs | 10 +++++++++-
>  1 file changed, 9 insertions(+), 1 deletion(-)
> 
> diff --git a/support/download/cvs b/support/download/cvs
> index 2c3a666..6491cc8 100755
> --- a/support/download/cvs
> +++ b/support/download/cvs
> @@ -26,7 +26,15 @@ rev="${3}"
>  rawname="${4}"
>  basename="${5}"
>  
> +if [[ ${rev} =~ ^[0-9] ]]; then
> +    # Date, because a tag cannot begin with a number
> +    select="-D ${rev}"
> +else
> +    # Tag
> +    select=-"r :${rev}"
> +fi
> +
>  ${CVS} ${verbose} -z3 -d":pserver:anonymous@${repo}" \
> -       co -d "${basename}" -r ":${rev}" -P "${rawname}"
> +       co -d "${basename}" ${select} -P "${rawname}"
>  
>  tar czf "${output}" "${basename}"
> -- 
> 2.3.5
> 

-- 
.-----------------.--------------------.------------------.--------------------.
|  Yann E. MORIN  | Real-Time Embedded | /"\ ASCII RIBBON | Erics' conspiracy: |
| +33 662 376 056 | Software  Designer | \ / CAMPAIGN     |  ___               |
| +33 223 225 172 `------------.-------:  X  AGAINST      |  \e/  There is no  |
| http://ymorin.is-a-geek.org/ | _/*\_ | / \ HTML MAIL    |   v   conspiracy.  |
'------------------------------^-------^------------------^--------------------'

[Buildroot] [PATCH 2/3] manual: cvs: document that a date can be used instead of a tag

[Fabio Porcedda]

Also instead of using the generic word "timestamp" use the word "tag".

Signed-off-by: Fabio Porcedda <fabio.porcedda@gmail.com>
---
 docs/manual/adding-packages-generic.txt | 2 +-
 1 file changed, 1 insertion(+), 1 deletion(-)

diff --git a/docs/manual/adding-packages-generic.txt b/docs/manual/adding-packages-generic.txt
index 8cf6bb6..be754a2 100644
--- a/docs/manual/adding-packages-generic.txt
+++ b/docs/manual/adding-packages-generic.txt
@@ -280,7 +280,7 @@ information is (assuming the package name is +libfoo+) :
      Only anonymous pserver mode is supported.
      +LIBFOO_SITE+ 'must' contain the source URL as well as the remote
      repository directory. The module is the package name.
-     +LIBFOO_VERSION+ is 'mandatory' and 'must' be a timestamp.
+     +LIBFOO_VERSION+ is 'mandatory' and 'must' be a tag or a date.
   ** +git+ for retrieving source code from a Git repository. Used by
      default when +LIBFOO_SITE+ begins with +git://+. The downloaded
      source code is cached as with the +svn+
-- 
2.3.5

[Buildroot] [PATCH 2/3] manual: cvs: document that a date can be used instead of a tag

[Yann E. MORIN]

Fabio, All,

On 2015-04-18 18:54 +0200, Fabio Porcedda spake thusly:
> Also instead of using the generic word "timestamp" use the word "tag".
> 
> Signed-off-by: Fabio Porcedda <fabio.porcedda@gmail.com>
> ---
>  docs/manual/adding-packages-generic.txt | 2 +-
>  1 file changed, 1 insertion(+), 1 deletion(-)
> 
> diff --git a/docs/manual/adding-packages-generic.txt b/docs/manual/adding-packages-generic.txt
> index 8cf6bb6..be754a2 100644
> --- a/docs/manual/adding-packages-generic.txt
> +++ b/docs/manual/adding-packages-generic.txt
> @@ -280,7 +280,7 @@ information is (assuming the package name is +libfoo+) :
>       Only anonymous pserver mode is supported.
>       +LIBFOO_SITE+ 'must' contain the source URL as well as the remote
>       repository directory. The module is the package name.
> -     +LIBFOO_VERSION+ is 'mandatory' and 'must' be a timestamp.
> +     +LIBFOO_VERSION+ is 'mandatory' and 'must' be a tag or a date.

I'd like we document the format of the date we recognise. Because, IIRC,
cvs accepts 'yesterday' as a date format, and that would be interpreted
as a tag with the current code.

So maybe, just state something like:

    ... or a date (YYYYMMDD:hhmmss)

Or whatever is a date format recognised.

Regards,
Yann E. MORIN.

>    ** +git+ for retrieving source code from a Git repository. Used by
>       default when +LIBFOO_SITE+ begins with +git://+. The downloaded
>       source code is cached as with the +svn+
> -- 
> 2.3.5
> 

-- 
.-----------------.--------------------.------------------.--------------------.
|  Yann E. MORIN  | Real-Time Embedded | /"\ ASCII RIBBON | Erics' conspiracy: |
| +33 662 376 056 | Software  Designer | \ / CAMPAIGN     |  ___               |
| +33 223 225 172 `------------.-------:  X  AGAINST      |  \e/  There is no  |
| http://ymorin.is-a-geek.org/ | _/*\_ | / \ HTML MAIL    |   v   conspiracy.  |
'------------------------------^-------^------------------^--------------------'

[Buildroot] [PATCH 2/3] manual: cvs: document that a date can be used instead of a tag

[Fabio Porcedda]

On Sun, Apr 19, 2015 at 10:10 AM, Yann E. MORIN <yann.morin.1998@free.fr> wrote:
> Fabio, All,
>
> On 2015-04-18 18:54 +0200, Fabio Porcedda spake thusly:
>> Also instead of using the generic word "timestamp" use the word "tag".
>>
>> Signed-off-by: Fabio Porcedda <fabio.porcedda@gmail.com>
>> ---
>>  docs/manual/adding-packages-generic.txt | 2 +-
>>  1 file changed, 1 insertion(+), 1 deletion(-)
>>
>> diff --git a/docs/manual/adding-packages-generic.txt b/docs/manual/adding-packages-generic.txt
>> index 8cf6bb6..be754a2 100644
>> --- a/docs/manual/adding-packages-generic.txt
>> +++ b/docs/manual/adding-packages-generic.txt
>> @@ -280,7 +280,7 @@ information is (assuming the package name is +libfoo+) :
>>       Only anonymous pserver mode is supported.
>>       +LIBFOO_SITE+ 'must' contain the source URL as well as the remote
>>       repository directory. The module is the package name.
>> -     +LIBFOO_VERSION+ is 'mandatory' and 'must' be a timestamp.
>> +     +LIBFOO_VERSION+ is 'mandatory' and 'must' be a tag or a date.
>
> I'd like we document the format of the date we recognise. Because, IIRC,
> cvs accepts 'yesterday' as a date format, and that would be interpreted
> as a tag with the current code.
>
> So maybe, just state something like:
>
>     ... or a date (YYYYMMDD:hhmmss)

What about this:
(<YYYY>-<MM>-<DD>[T<HH><MM>[<SS>]] e.g 2015-12-20 or 2015-12-20T1010)

> Or whatever is a date format recognised.
>

Thanks for reviewing.
-- 
Fabio Porcedda

[Buildroot] [PATCH 2/3] manual: cvs: document that a date can be used instead of a tag

[Arnout Vandecappelle]

On 19/04/15 12:08, Fabio Porcedda wrote:
> On Sun, Apr 19, 2015 at 10:10 AM, Yann E. MORIN <yann.morin.1998@free.fr> wrote:
>> Fabio, All,
>>
>> On 2015-04-18 18:54 +0200, Fabio Porcedda spake thusly:
>>> Also instead of using the generic word "timestamp" use the word "tag".
>>>
>>> Signed-off-by: Fabio Porcedda <fabio.porcedda@gmail.com>
>>> ---
>>>  docs/manual/adding-packages-generic.txt | 2 +-
>>>  1 file changed, 1 insertion(+), 1 deletion(-)
>>>
>>> diff --git a/docs/manual/adding-packages-generic.txt b/docs/manual/adding-packages-generic.txt
>>> index 8cf6bb6..be754a2 100644
>>> --- a/docs/manual/adding-packages-generic.txt
>>> +++ b/docs/manual/adding-packages-generic.txt
>>> @@ -280,7 +280,7 @@ information is (assuming the package name is +libfoo+) :
>>>       Only anonymous pserver mode is supported.
>>>       +LIBFOO_SITE+ 'must' contain the source URL as well as the remote
>>>       repository directory. The module is the package name.
>>> -     +LIBFOO_VERSION+ is 'mandatory' and 'must' be a timestamp.
>>> +     +LIBFOO_VERSION+ is 'mandatory' and 'must' be a tag or a date.
>>
>> I'd like we document the format of the date we recognise. Because, IIRC,
>> cvs accepts 'yesterday' as a date format, and that would be interpreted
>> as a tag with the current code.
>>
>> So maybe, just state something like:
>>
>>     ... or a date (YYYYMMDD:hhmmss)
> 
> What about this:
> (<YYYY>-<MM>-<DD>[T<HH><MM>[<SS>]] e.g 2015-12-20 or 2015-12-20T1010)

 Can a timezone be added to that? Playing with dates and times without timezone
is dangerous.

 Regards,
 Arnout

-- 
Arnout Vandecappelle                          arnout at mind be
Senior Embedded Software Architect            +32-16-286500
Essensium/Mind                                http://www.mind.be
G.Geenslaan 9, 3001 Leuven, Belgium           BE 872 984 063 RPR Leuven
LinkedIn profile: http://www.linkedin.com/in/arnoutvandecappelle
GPG fingerprint:  7CB5 E4CC 6C2E EFD4 6E3D A754 F963 ECAB 2450 2F1F

[Buildroot] [PATCH 2/3] manual: cvs: document that a date can be used instead of a tag

[Fabio Porcedda]

On Sun, Apr 19, 2015 at 5:45 PM, Arnout Vandecappelle <arnout@mind.be> wrote:
> On 19/04/15 12:08, Fabio Porcedda wrote:
>> On Sun, Apr 19, 2015 at 10:10 AM, Yann E. MORIN <yann.morin.1998@free.fr> wrote:
>>> Fabio, All,
>>>
>>> On 2015-04-18 18:54 +0200, Fabio Porcedda spake thusly:
>>>> Also instead of using the generic word "timestamp" use the word "tag".
>>>>
>>>> Signed-off-by: Fabio Porcedda <fabio.porcedda@gmail.com>
>>>> ---
>>>>  docs/manual/adding-packages-generic.txt | 2 +-
>>>>  1 file changed, 1 insertion(+), 1 deletion(-)
>>>>
>>>> diff --git a/docs/manual/adding-packages-generic.txt b/docs/manual/adding-packages-generic.txt
>>>> index 8cf6bb6..be754a2 100644
>>>> --- a/docs/manual/adding-packages-generic.txt
>>>> +++ b/docs/manual/adding-packages-generic.txt
>>>> @@ -280,7 +280,7 @@ information is (assuming the package name is +libfoo+) :
>>>>       Only anonymous pserver mode is supported.
>>>>       +LIBFOO_SITE+ 'must' contain the source URL as well as the remote
>>>>       repository directory. The module is the package name.
>>>> -     +LIBFOO_VERSION+ is 'mandatory' and 'must' be a timestamp.
>>>> +     +LIBFOO_VERSION+ is 'mandatory' and 'must' be a tag or a date.
>>>
>>> I'd like we document the format of the date we recognise. Because, IIRC,
>>> cvs accepts 'yesterday' as a date format, and that would be interpreted
>>> as a tag with the current code.
>>>
>>> So maybe, just state something like:
>>>
>>>     ... or a date (YYYYMMDD:hhmmss)
>>
>> What about this:
>> (<YYYY>-<MM>-<DD>[T<HH><MM>[<SS>]] e.g 2015-12-20 or 2015-12-20T1010)
>
>  Can a timezone be added to that? Playing with dates and times without timezone
> is dangerous.

This one works: (<YYYY>-<MM>-<DD>[T<HH>:<MM>[:<SS>][-<ZZ>] e.g
2015-12-20 or 2015-12-20T10:10-00)

But to made it works I must replace the ":" with "_":

index bdb26b8..2beb529 100644
--- a/package/pkg-generic.mk
+++ b/package/pkg-generic.mk
@@ -325,14 +325,14 @@ $(2)_RAWNAME                      =  $$(patsubst
host-%,%,$(1))
 ifndef $(2)_VERSION
  ifdef $(3)_VERSION
   $(2)_DL_VERSION := $$(strip $$($(3)_VERSION))
-  $(2)_VERSION := $$(subst /,_,$$(strip $$($(3)_VERSION)))
+  $(2)_VERSION := $$(subst :,_,$$(subst /,_,$$(strip $$($(3)_VERSION))))
  else
   $(2)_VERSION = undefined
   $(2)_DL_VERSION = undefined
  endif
 else
   $(2)_DL_VERSION := $$(strip $$($(2)_VERSION))
-  $(2)_VERSION := $$(strip $$(subst /,_,$$($(2)_VERSION)))
+  $(2)_VERSION := $$(strip $$(subst :,_,$$(subst /,_,$$($(2)_VERSION))))
 endif

BR
-- 
Fabio Porcedda

[Buildroot] [PATCH 2/3] manual: cvs: document that a date can be used instead of a tag

[Arnout Vandecappelle]

On 20/04/15 04:36, Fabio Porcedda wrote:
> On Sun, Apr 19, 2015 at 5:45 PM, Arnout Vandecappelle <arnout@mind.be> wrote:
>> On 19/04/15 12:08, Fabio Porcedda wrote:
>>> On Sun, Apr 19, 2015 at 10:10 AM, Yann E. MORIN <yann.morin.1998@free.fr> wrote:
>>>> Fabio, All,
>>>>
>>>> On 2015-04-18 18:54 +0200, Fabio Porcedda spake thusly:
>>>>> Also instead of using the generic word "timestamp" use the word "tag".
>>>>>
>>>>> Signed-off-by: Fabio Porcedda <fabio.porcedda@gmail.com>
>>>>> ---
>>>>>  docs/manual/adding-packages-generic.txt | 2 +-
>>>>>  1 file changed, 1 insertion(+), 1 deletion(-)
>>>>>
>>>>> diff --git a/docs/manual/adding-packages-generic.txt b/docs/manual/adding-packages-generic.txt
>>>>> index 8cf6bb6..be754a2 100644
>>>>> --- a/docs/manual/adding-packages-generic.txt
>>>>> +++ b/docs/manual/adding-packages-generic.txt
>>>>> @@ -280,7 +280,7 @@ information is (assuming the package name is +libfoo+) :
>>>>>       Only anonymous pserver mode is supported.
>>>>>       +LIBFOO_SITE+ 'must' contain the source URL as well as the remote
>>>>>       repository directory. The module is the package name.
>>>>> -     +LIBFOO_VERSION+ is 'mandatory' and 'must' be a timestamp.
>>>>> +     +LIBFOO_VERSION+ is 'mandatory' and 'must' be a tag or a date.
>>>>
>>>> I'd like we document the format of the date we recognise. Because, IIRC,
>>>> cvs accepts 'yesterday' as a date format, and that would be interpreted
>>>> as a tag with the current code.
>>>>
>>>> So maybe, just state something like:
>>>>
>>>>     ... or a date (YYYYMMDD:hhmmss)
>>>
>>> What about this:
>>> (<YYYY>-<MM>-<DD>[T<HH><MM>[<SS>]] e.g 2015-12-20 or 2015-12-20T1010)
>>
>>  Can a timezone be added to that? Playing with dates and times without timezone
>> is dangerous.
> 
> This one works: (<YYYY>-<MM>-<DD>[T<HH>:<MM>[:<SS>][-<ZZ>] e.g
> 2015-12-20 or 2015-12-20T10:10-00)

 Well, the example should include a timezone and mention that adding a timezone
is advisable.

 Alternatively, we could specifies that times are in UTC and set TZ=UTC in the
cvs helper.

> 
> But to made it works I must replace the ":" with "_":

 Why? The / substitution is only needed because VERSION is used in filenames,
and : is fine in filenames, no?

 Regards,
 Arnout

> 
> index bdb26b8..2beb529 100644
> --- a/package/pkg-generic.mk
> +++ b/package/pkg-generic.mk
> @@ -325,14 +325,14 @@ $(2)_RAWNAME                      =  $$(patsubst
> host-%,%,$(1))
>  ifndef $(2)_VERSION
>   ifdef $(3)_VERSION
>    $(2)_DL_VERSION := $$(strip $$($(3)_VERSION))
> -  $(2)_VERSION := $$(subst /,_,$$(strip $$($(3)_VERSION)))
> +  $(2)_VERSION := $$(subst :,_,$$(subst /,_,$$(strip $$($(3)_VERSION))))
>   else
>    $(2)_VERSION = undefined
>    $(2)_DL_VERSION = undefined
>   endif
>  else
>    $(2)_DL_VERSION := $$(strip $$($(2)_VERSION))
> -  $(2)_VERSION := $$(strip $$(subst /,_,$$($(2)_VERSION)))
> +  $(2)_VERSION := $$(strip $$(subst :,_,$$(subst /,_,$$($(2)_VERSION))))
>  endif
> 
> BR
> 


-- 
Arnout Vandecappelle                          arnout at mind be
Senior Embedded Software Architect            +32-16-286500
Essensium/Mind                                http://www.mind.be
G.Geenslaan 9, 3001 Leuven, Belgium           BE 872 984 063 RPR Leuven
LinkedIn profile: http://www.linkedin.com/in/arnoutvandecappelle
GPG fingerprint:  7CB5 E4CC 6C2E EFD4 6E3D A754 F963 ECAB 2450 2F1F

[Buildroot] [PATCH 2/3] manual: cvs: document that a date can be used instead of a tag

[Fabio Porcedda]

On Mon, Apr 20, 2015 at 11:06 PM, Arnout Vandecappelle <arnout@mind.be> wrote:
> On 20/04/15 04:36, Fabio Porcedda wrote:
>> On Sun, Apr 19, 2015 at 5:45 PM, Arnout Vandecappelle <arnout@mind.be> wrote:
>>> On 19/04/15 12:08, Fabio Porcedda wrote:
>>>> On Sun, Apr 19, 2015 at 10:10 AM, Yann E. MORIN <yann.morin.1998@free.fr> wrote:
>>>>> Fabio, All,
>>>>>
>>>>> On 2015-04-18 18:54 +0200, Fabio Porcedda spake thusly:
>>>>>> Also instead of using the generic word "timestamp" use the word "tag".
>>>>>>
>>>>>> Signed-off-by: Fabio Porcedda <fabio.porcedda@gmail.com>
>>>>>> ---
>>>>>>  docs/manual/adding-packages-generic.txt | 2 +-
>>>>>>  1 file changed, 1 insertion(+), 1 deletion(-)
>>>>>>
>>>>>> diff --git a/docs/manual/adding-packages-generic.txt b/docs/manual/adding-packages-generic.txt
>>>>>> index 8cf6bb6..be754a2 100644
>>>>>> --- a/docs/manual/adding-packages-generic.txt
>>>>>> +++ b/docs/manual/adding-packages-generic.txt
>>>>>> @@ -280,7 +280,7 @@ information is (assuming the package name is +libfoo+) :
>>>>>>       Only anonymous pserver mode is supported.
>>>>>>       +LIBFOO_SITE+ 'must' contain the source URL as well as the remote
>>>>>>       repository directory. The module is the package name.
>>>>>> -     +LIBFOO_VERSION+ is 'mandatory' and 'must' be a timestamp.
>>>>>> +     +LIBFOO_VERSION+ is 'mandatory' and 'must' be a tag or a date.
>>>>>
>>>>> I'd like we document the format of the date we recognise. Because, IIRC,
>>>>> cvs accepts 'yesterday' as a date format, and that would be interpreted
>>>>> as a tag with the current code.
>>>>>
>>>>> So maybe, just state something like:
>>>>>
>>>>>     ... or a date (YYYYMMDD:hhmmss)
>>>>
>>>> What about this:
>>>> (<YYYY>-<MM>-<DD>[T<HH><MM>[<SS>]] e.g 2015-12-20 or 2015-12-20T1010)
>>>
>>>  Can a timezone be added to that? Playing with dates and times without timezone
>>> is dangerous.
>>
>> This one works: (<YYYY>-<MM>-<DD>[T<HH>:<MM>[:<SS>][-<ZZ>] e.g
>> 2015-12-20 or 2015-12-20T10:10-00)
>
>  Well, the example should include a timezone and mention that adding a timezone
> is advisable.

Ok.

>  Alternatively, we could specifies that times are in UTC and set TZ=UTC in the
> cvs helper.

Fine for me.
Yann what do you think about it?

>
>>
>> But to made it works I must replace the ":" with "_":
>
>  Why? The / substitution is only needed because VERSION is used in filenames,
> and : is fine in filenames, no?

The colon is fine for filenames, neverthless if i don't replace it, I
get this error:

$ make expect-source
package/expect/expect.mk:19: *** target pattern contains no '%'.  Stop.

>> index bdb26b8..2beb529 100644
>> --- a/package/pkg-generic.mk
>> +++ b/package/pkg-generic.mk
>> @@ -325,14 +325,14 @@ $(2)_RAWNAME                      =  $$(patsubst
>> host-%,%,$(1))
>>  ifndef $(2)_VERSION
>>   ifdef $(3)_VERSION
>>    $(2)_DL_VERSION := $$(strip $$($(3)_VERSION))
>> -  $(2)_VERSION := $$(subst /,_,$$(strip $$($(3)_VERSION)))
>> +  $(2)_VERSION := $$(subst :,_,$$(subst /,_,$$(strip $$($(3)_VERSION))))
>>   else
>>    $(2)_VERSION = undefined
>>    $(2)_DL_VERSION = undefined
>>   endif
>>  else
>>    $(2)_DL_VERSION := $$(strip $$($(2)_VERSION))
>> -  $(2)_VERSION := $$(strip $$(subst /,_,$$($(2)_VERSION)))
>> +  $(2)_VERSION := $$(strip $$(subst :,_,$$(subst /,_,$$($(2)_VERSION))))
>>  endif

BR
-- 
Fabio Porcedda

[Buildroot] [PATCH 2/3] manual: cvs: document that a date can be used instead of a tag

[Yann E. MORIN]

Fabio, Gustavo, All,

On 2015-04-22 10:03 +0200, Fabio Porcedda spake thusly:
> On Mon, Apr 20, 2015 at 11:06 PM, Arnout Vandecappelle <arnout@mind.be> wrote:
> > On 20/04/15 04:36, Fabio Porcedda wrote:
[--SNIP--]
> >> This one works: (<YYYY>-<MM>-<DD>[T<HH>:<MM>[:<SS>][-<ZZ>] e.g
> >> 2015-12-20 or 2015-12-20T10:10-00)
> >
> >  Well, the example should include a timezone and mention that adding a timezone
> > is advisable.
> Ok.
> 
> >  Alternatively, we could specifies that times are in UTC and set TZ=UTC in the
> > cvs helper.
> 
> Fine for me.
> Yann what do you think about it?

I don't care. But what if a user specifies a timezone?

The tricky part with dealing with date (aka human date and timei) is
exactly that: having to deal with it. It is very complex to parse a date
in a reliable manner.

So we can't really validate a date. I'd rather we allow the full range
of dates supported by cvs, and let it deal with whatever the user
provides, and if that's incorrect, let cvs fail, not us (well, we'd
eventually fail, but not on our will).

> >> But to made it works I must replace the ":" with "_":
> >
> >  Why? The / substitution is only needed because VERSION is used in filenames,
> > and : is fine in filenames, no?
> 
> The colon is fine for filenames, neverthless if i don't replace it, I
> get this error:
> 
> $ make expect-source
> package/expect/expect.mk:19: *** target pattern contains no '%'.  Stop.

Yes, because the $(@D) contains the version string, and it is a make
goal. So, we'd end up wth something like;

    foo-2015-04-22T18:26/.stamp_extracted:
        blabla

Notice that there are two columns, and that is not valid make syntax
(AFAIR).

Which make me think: "2015-04-22 18:26" is an equally valid date, and we
do not really support spaces in paths, so we'd have to also replaces
spaces as well.

Regards,
Yann E. MORIN.

-- 
.-----------------.--------------------.------------------.--------------------.
|  Yann E. MORIN  | Real-Time Embedded | /"\ ASCII RIBBON | Erics' conspiracy: |
| +33 662 376 056 | Software  Designer | \ / CAMPAIGN     |  ___               |
| +33 223 225 172 `------------.-------:  X  AGAINST      |  \e/  There is no  |
| http://ymorin.is-a-geek.org/ | _/*\_ | / \ HTML MAIL    |   v   conspiracy.  |
'------------------------------^-------^------------------^--------------------'

[Buildroot] [PATCH 2/3] manual: cvs: document that a date can be used instead of a tag

[Arnout Vandecappelle]

On 04/22/15 18:29, Yann E. MORIN wrote:
> Fabio, Gustavo, All,
> 
> On 2015-04-22 10:03 +0200, Fabio Porcedda spake thusly:
>> On Mon, Apr 20, 2015 at 11:06 PM, Arnout Vandecappelle <arnout@mind.be> wrote:
>>> On 20/04/15 04:36, Fabio Porcedda wrote:
> [--SNIP--]
>>>> This one works: (<YYYY>-<MM>-<DD>[T<HH>:<MM>[:<SS>][-<ZZ>] e.g
>>>> 2015-12-20 or 2015-12-20T10:10-00)
>>>
>>>  Well, the example should include a timezone and mention that adding a timezone
>>> is advisable.
>> Ok.
>>
>>>  Alternatively, we could specifies that times are in UTC and set TZ=UTC in the
>>> cvs helper.
>>
>> Fine for me.
>> Yann what do you think about it?
> 
> I don't care. But what if a user specifies a timezone?

 Then the user-specified timezone overrides the environment, so the user gets
exactly what he expects. So I think setting TZ=UTC is the best option.

> 
> The tricky part with dealing with date (aka human date and timei) is
> exactly that: having to deal with it. It is very complex to parse a date
> in a reliable manner.
> 
> So we can't really validate a date. I'd rather we allow the full range
> of dates supported by cvs, and let it deal with whatever the user
> provides, and if that's incorrect, let cvs fail, not us (well, we'd
> eventually fail, but not on our will).

 Well, it was never the intention that we would parse it, just that we provide
an example of a valid date.

> 
>>>> But to made it works I must replace the ":" with "_":
>>>
>>>  Why? The / substitution is only needed because VERSION is used in filenames,
>>> and : is fine in filenames, no?
>>
>> The colon is fine for filenames, neverthless if i don't replace it, I
>> get this error:
>>
>> $ make expect-source
>> package/expect/expect.mk:19: *** target pattern contains no '%'.  Stop.
> 
> Yes, because the $(@D) contains the version string, and it is a make
> goal. So, we'd end up wth something like;
> 
>     foo-2015-04-22T18:26/.stamp_extracted:
>         blabla

 Actually it's in this rule:

$(1)-install-target:            $$($(2)_TARGET_INSTALL_TARGET)
which expands to
foo-install-target: foo-2015-04-22T18:26/.stamp_target_installed

> 
> Notice that there are two columns, and that is not valid make syntax
> (AFAIR).
> 
> Which make me think: "2015-04-22 18:26" is an equally valid date, and we
> do not really support spaces in paths, so we'd have to also replaces
> spaces as well.

 Right!


 Regards,
 Arnout

> 
> Regards,
> Yann E. MORIN.
> 


-- 
Arnout Vandecappelle                          arnout at mind be
Senior Embedded Software Architect            +32-16-286500
Essensium/Mind                                http://www.mind.be
G.Geenslaan 9, 3001 Leuven, Belgium           BE 872 984 063 RPR Leuven
LinkedIn profile: http://www.linkedin.com/in/arnoutvandecappelle
GPG fingerprint:  7CB5 E4CC 6C2E EFD4 6E3D A754 F963 ECAB 2450 2F1F

[Buildroot] [PATCH 2/3] manual: cvs: document that a date can be used instead of a tag

[Yann E. MORIN]

Fabio, All,

On 2015-04-19 12:08 +0200, Fabio Porcedda spake thusly:
> On Sun, Apr 19, 2015 at 10:10 AM, Yann E. MORIN <yann.morin.1998@free.fr> wrote:
> > On 2015-04-18 18:54 +0200, Fabio Porcedda spake thusly:
[--SNIP--]
> >> -     +LIBFOO_VERSION+ is 'mandatory' and 'must' be a timestamp.
> >> +     +LIBFOO_VERSION+ is 'mandatory' and 'must' be a tag or a date.
> >
> > I'd like we document the format of the date we recognise. Because, IIRC,
> > cvs accepts 'yesterday' as a date format, and that would be interpreted
> > as a tag with the current code.
> >
> > So maybe, just state something like:
> >
> >     ... or a date (YYYYMMDD:hhmmss)
> 
> What about this:
> (<YYYY>-<MM>-<DD>[T<HH><MM>[<SS>]] e.g 2015-12-20 or 2015-12-20T1010)

Well, this is exactly RFC3339;
    https://tools.ietf.org/html/rfc3339#section-5.8

So we may just state smehing along the lines of:

    ... or a date (formatted as defined in
    https://tools.ietf.org/html/rfc3339[RFC 3339]).

But I'd have to agree that it is not user-friendly, so we still want a
complete example, like so:

    ... or a date (formatted as defined in https://tools.ietf.org/html/rfc3339[RFC 3339],
    like <YYYY>-<MM>-<DD>[T<HH>:<MM>[:<SS>[.sss]]] e.g 2015-12-20 or 2015-12-20T10:10 or
    even 2015-12-20T16:42:12.467923; for complet ereference, see the
    http://sunsite.ualberta.ca/Documentation/Gnu/cvs-1.12.11/cvs.html#Date-input-formats[CVS
    manual]).

Regards,
Yann E. MORIN.

-- 
.-----------------.--------------------.------------------.--------------------.
|  Yann E. MORIN  | Real-Time Embedded | /"\ ASCII RIBBON | Erics' conspiracy: |
| +33 662 376 056 | Software  Designer | \ / CAMPAIGN     |  ___               |
| +33 223 225 172 `------------.-------:  X  AGAINST      |  \e/  There is no  |
| http://ymorin.is-a-geek.org/ | _/*\_ | / \ HTML MAIL    |   v   conspiracy.  |
'------------------------------^-------^------------------^--------------------'

[Buildroot] [PATCH 3/3] expect: bump to version 5.45.3

[Fabio Porcedda]

Download it from the cvs repository because it is not avaialbe as a
downloadable file.
Also because a tag isn't avaiable use the date.

Signed-off-by: Fabio Porcedda <fabio.porcedda@gmail.com>
---
 package/expect/expect.hash | 2 --
 package/expect/expect.mk   | 5 ++---
 2 files changed, 2 insertions(+), 5 deletions(-)
 delete mode 100644 package/expect/expect.hash

diff --git a/package/expect/expect.hash b/package/expect/expect.hash
deleted file mode 100644
index 09e68e7..0000000
--- a/package/expect/expect.hash
+++ /dev/null
@@ -1,2 +0,0 @@
-# Locally computed:
-sha256  b28dca90428a3b30e650525cdc16255d76bb6ccd65d448be53e620d95d5cc040  expect5.45.tar.gz
diff --git a/package/expect/expect.mk b/package/expect/expect.mk
index 8ef7e90..513be8f 100644
--- a/package/expect/expect.mk
+++ b/package/expect/expect.mk
@@ -4,9 +4,8 @@
 #
 ################################################################################
 
-EXPECT_VERSION = 5.45
-EXPECT_SOURCE = expect$(EXPECT_VERSION).tar.gz
-EXPECT_SITE = http://downloads.sourceforge.net/project/expect/Expect/$(EXPECT_VERSION)
+EXPECT_VERSION = 2014/05/02 # v5.45.3
+EXPECT_SITE = cvs://expect.cvs.sourceforge.net:/cvsroot/expect
 EXPECT_LICENSE = Public domain
 EXPECT_LICENSE_FILES = README
 
-- 
2.3.5

[Buildroot] [PATCH 3/3] expect: bump to version 5.45.3

[Yann E. MORIN]

Fabio, All,

On 2015-04-18 18:54 +0200, Fabio Porcedda spake thusly:
> Download it from the cvs repository because it is not avaialbe as a
> downloadable file.
> Also because a tag isn't avaiable use the date.
> 
> Signed-off-by: Fabio Porcedda <fabio.porcedda@gmail.com>
> ---
>  package/expect/expect.hash | 2 --
>  package/expect/expect.mk   | 5 ++---
>  2 files changed, 2 insertions(+), 5 deletions(-)
>  delete mode 100644 package/expect/expect.hash
> 
> diff --git a/package/expect/expect.hash b/package/expect/expect.hash
> deleted file mode 100644
> index 09e68e7..0000000
> --- a/package/expect/expect.hash
> +++ /dev/null
> @@ -1,2 +0,0 @@
> -# Locally computed:
> -sha256  b28dca90428a3b30e650525cdc16255d76bb6ccd65d448be53e620d95d5cc040  expect5.45.tar.gz
> diff --git a/package/expect/expect.mk b/package/expect/expect.mk
> index 8ef7e90..513be8f 100644
> --- a/package/expect/expect.mk
> +++ b/package/expect/expect.mk
> @@ -4,9 +4,8 @@
>  #
>  ################################################################################
>  
> -EXPECT_VERSION = 5.45
> -EXPECT_SOURCE = expect$(EXPECT_VERSION).tar.gz
> -EXPECT_SITE = http://downloads.sourceforge.net/project/expect/Expect/$(EXPECT_VERSION)
> +EXPECT_VERSION = 2014/05/02 # v5.45.3

Do not put comments after the variable definition, it breaks things.

Note that the example we have in the manual is wrong. Care to address
that as well, please?

Regards,
Yann E. MORIN.

> +EXPECT_SITE = cvs://expect.cvs.sourceforge.net:/cvsroot/expect
>  EXPECT_LICENSE = Public domain
>  EXPECT_LICENSE_FILES = README
>  
> -- 
> 2.3.5
> 

-- 
.-----------------.--------------------.------------------.--------------------.
|  Yann E. MORIN  | Real-Time Embedded | /"\ ASCII RIBBON | Erics' conspiracy: |
| +33 662 376 056 | Software  Designer | \ / CAMPAIGN     |  ___               |
| +33 223 225 172 `------------.-------:  X  AGAINST      |  \e/  There is no  |
| http://ymorin.is-a-geek.org/ | _/*\_ | / \ HTML MAIL    |   v   conspiracy.  |
'------------------------------^-------^------------------^--------------------'

[Buildroot] [PATCH 3/3] expect: bump to version 5.45.3

[Fabio Porcedda]

On Sun, Apr 19, 2015 at 10:11 AM, Yann E. MORIN <yann.morin.1998@free.fr> wrote:
> Fabio, All,
>
> On 2015-04-18 18:54 +0200, Fabio Porcedda spake thusly:
>> Download it from the cvs repository because it is not avaialbe as a
>> downloadable file.
>> Also because a tag isn't avaiable use the date.
>>
>> Signed-off-by: Fabio Porcedda <fabio.porcedda@gmail.com>
>> ---
>>  package/expect/expect.hash | 2 --
>>  package/expect/expect.mk   | 5 ++---
>>  2 files changed, 2 insertions(+), 5 deletions(-)
>>  delete mode 100644 package/expect/expect.hash
>>
>> diff --git a/package/expect/expect.hash b/package/expect/expect.hash
>> deleted file mode 100644
>> index 09e68e7..0000000
>> --- a/package/expect/expect.hash
>> +++ /dev/null
>> @@ -1,2 +0,0 @@
>> -# Locally computed:
>> -sha256  b28dca90428a3b30e650525cdc16255d76bb6ccd65d448be53e620d95d5cc040  expect5.45.tar.gz
>> diff --git a/package/expect/expect.mk b/package/expect/expect.mk
>> index 8ef7e90..513be8f 100644
>> --- a/package/expect/expect.mk
>> +++ b/package/expect/expect.mk
>> @@ -4,9 +4,8 @@
>>  #
>>  ################################################################################
>>
>> -EXPECT_VERSION = 5.45
>> -EXPECT_SOURCE = expect$(EXPECT_VERSION).tar.gz
>> -EXPECT_SITE = http://downloads.sourceforge.net/project/expect/Expect/$(EXPECT_VERSION)
>> +EXPECT_VERSION = 2014/05/02 # v5.45.3
>
> Do not put comments after the variable definition, it breaks things.

Could you explain it?

> Note that the example we have in the manual is wrong. Care to address
> that as well, please?

Sure

Thanks for reviewing.
-- 
Fabio Porcedda

[Buildroot] [PATCH 3/3] expect: bump to version 5.45.3

[Yann E. MORIN]

Fabio, All,

On 2015-04-19 12:12 +0200, Fabio Porcedda spake thusly:
> On Sun, Apr 19, 2015 at 10:11 AM, Yann E. MORIN <yann.morin.1998@free.fr> wrote:
> > On 2015-04-18 18:54 +0200, Fabio Porcedda spake thusly:
[--SNIP--]
> >> +EXPECT_VERSION = 2014/05/02 # v5.45.3
> >
> > Do not put comments after the variable definition, it breaks things.
> 
> Could you explain it?

I'm not sure what's going on here, but we've observed failures with
that, most notably when downloading via git, and removing the comment
made the error vanish.

One wild guess is that the space(s) between the version and the '#' are
part of the value, and that breaks later on. Really, just a wild
guess...

> > Note that the example we have in the manual is wrong. Care to address
> > that as well, please?
> 
> Sure

Thanks! :-)

Regards,
Yann E. MORIN.

-- 
.-----------------.--------------------.------------------.--------------------.
|  Yann E. MORIN  | Real-Time Embedded | /"\ ASCII RIBBON | Erics' conspiracy: |
| +33 662 376 056 | Software  Designer | \ / CAMPAIGN     |  ___               |
| +33 223 225 172 `------------.-------:  X  AGAINST      |  \e/  There is no  |
| http://ymorin.is-a-geek.org/ | _/*\_ | / \ HTML MAIL    |   v   conspiracy.  |
'------------------------------^-------^------------------^--------------------'