[Buildroot] $(DATE) out of sync with system

[Buildroot] $(DATE) out of sync with system

[Mike Sander]

I have Custom rootfs suffix set to the default $(DATE).   I have noticed 
that a few hours before midnight my rootfs.arm-yyyymmdd.romfs, the dd 
portion advances prematurely. this breaks a script I have to copy 
binaries after the build.

I suspect it is somehow related to DATE not using the timezone correctly 
(maybe I dont have it set correctly?).  I have attempted to find where 
DATE is being set.  I would imaging that it is using the "date" function 
& format specifiers.   Unfortunately DATE is a VERY common symbol under 
buildroot.  I am unable to locate where this symbol is initially set.

Can someone tell me where to look?

Thanks,

Mike

[Buildroot] $(DATE) out of sync with system

[Ulf Samuelsson]

>I have Custom rootfs suffix set to the default $(DATE).   I have noticed 
> that a few hours before midnight my rootfs.arm-yyyymmdd.romfs, the dd 
> portion advances prematurely. this breaks a script I have to copy 
> binaries after the build.
> 
> I suspect it is somehow related to DATE not using the timezone correctly 
> (maybe I dont have it set correctly?).  I have attempted to find where 
> DATE is being set.  I would imaging that it is using the "date" function 
> & format specifiers.   Unfortunately DATE is a VERY common symbol under 
> buildroot.  I am unable to locate where this symbol is initially set.
> 
> Can someone tell me where to look?
> 


Top Makefile

Best Regards
Ulf Samuelsson

[Buildroot] $(DATE) out of sync with system

[Mike Sander]

Ulf Samuelsson wrote:
>> I have Custom rootfs suffix set to the default $(DATE).   I have noticed 
>> that a few hours before midnight my rootfs.arm-yyyymmdd.romfs, the dd 
>> portion advances prematurely. this breaks a script I have to copy 
>> binaries after the build.
>>
>> I suspect it is somehow related to DATE not using the timezone correctly 
>> (maybe I dont have it set correctly?).  I have attempted to find where 
>> DATE is being set.  I would imaging that it is using the "date" function 
>> & format specifiers.   Unfortunately DATE is a VERY common symbol under 
>> buildroot.  I am unable to locate where this symbol is initially set.
>>
>> Can someone tell me where to look?
>>
>>     
>
>
> Top Makefile
>   
Thanks.  I had done a multi level grep for "DATE" and "="..... but this 
generated literally hundreds of lines of output..... that proverbial 
needle in the hay stack.

The relevant line does the following DATE:= $(shell date -u +%Y%m%d).  
Anyone know why the "-u" which specifies Coordinated Universal Time?  
This explains the rollover I see a few hours before midnight.  Would it 
not make more sense for this just to be local time?    FWIW... I can 
update my script to be consistent with the DATE logic now that I know 
what is going on.

Regards,

Mike
> Best Regards
> Ulf Samuelsson
>
>
>
> =-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-
> Scanned with Copfilter Version 0.84beta3a (P3Scan 2.2.1)
> AntiSpam:  SpamAssassin 3.2.3
> by Markus Madlener @ http://www.copfilter.org
>
>   

[Buildroot] $(DATE) out of sync with system

[Ulf Samuelsson]

fre 2008-05-09 klockan 12:34 -0400 skrev Mike Sander:
> Ulf Samuelsson wrote:
> >> I have Custom rootfs suffix set to the default $(DATE).   I have noticed 
> >> that a few hours before midnight my rootfs.arm-yyyymmdd.romfs, the dd 
> >> portion advances prematurely. this breaks a script I have to copy 
> >> binaries after the build.
> >>
> >> I suspect it is somehow related to DATE not using the timezone correctly 
> >> (maybe I dont have it set correctly?).  I have attempted to find where 
> >> DATE is being set.  I would imaging that it is using the "date" function 
> >> & format specifiers.   Unfortunately DATE is a VERY common symbol under 
> >> buildroot.  I am unable to locate where this symbol is initially set.
> >>
> >> Can someone tell me where to look?
> >>
> >>     
> >
> >
> > Top Makefile
> >   
> Thanks.  I had done a multi level grep for "DATE" and "="..... but this 
> generated literally hundreds of lines of output..... that proverbial 
> needle in the hay stack.
> 
> The relevant line does the following DATE:= $(shell date -u +%Y%m%d).  
> Anyone know why the "-u" which specifies Coordinated Universal Time?  
> This explains the rollover I see a few hours before midnight.  Would it 
> not make more sense for this just to be local time?    FWIW... I can 
> update my script to be consistent with the DATE logic now that I know 
> what is going on.
> 
No you are right, I think it should use local time, so I will fix this.


-- 
Best Regards,
Ulf Samuelsson

[Buildroot] $(DATE) out of sync with system

[Mike Sander]

Ulf Samuelsson wrote:
> fre 2008-05-09 klockan 12:34 -0400 skrev Mike Sander:
>   
>> Ulf Samuelsson wrote:
>>     
>>>> I have Custom rootfs suffix set to the default $(DATE).   I have noticed 
>>>> that a few hours before midnight my rootfs.arm-yyyymmdd.romfs, the dd 
>>>> portion advances prematurely. this breaks a script I have to copy 
>>>> binaries after the build.
>>>>
>>>> I suspect it is somehow related to DATE not using the timezone correctly 
>>>> (maybe I dont have it set correctly?).  I have attempted to find where 
>>>> DATE is being set.  I would imaging that it is using the "date" function 
>>>> & format specifiers.   Unfortunately DATE is a VERY common symbol under 
>>>> buildroot.  I am unable to locate where this symbol is initially set.
>>>>
>>>> Can someone tell me where to look?
>>>>
>>>>     
>>>>         
>>> Top Makefile
>>>   
>>>       
>> Thanks.  I had done a multi level grep for "DATE" and "="..... but this 
>> generated literally hundreds of lines of output..... that proverbial 
>> needle in the hay stack.
>>
>> The relevant line does the following DATE:= $(shell date -u +%Y%m%d).  
>> Anyone know why the "-u" which specifies Coordinated Universal Time?  
>> This explains the rollover I see a few hours before midnight.  Would it 
>> not make more sense for this just to be local time?    FWIW... I can 
>> update my script to be consistent with the DATE logic now that I know 
>> what is going on.
>>
>>     
> No you are right, I think it should use local time, so I will fix this.
>
>
>   
Thank You..

Mike