Scheduled removals

Scheduled removals

[Jan Engelhardt]

Hi,


as per Documentation/feature-removal-schedule.txt, some legacy code is 
ripe for removal. Now while pursuing this I wondered whether the struct 
definitions for the old code in the header files (e.g. in 
linux/netfilter/xt_CONNMARK.h) can also be removed or whether people 
will argue that doing so would unnecessarily cause dismay on behalf the 
potential users (of which there are not any). I am not aware of any 
users, and neither is Search Engine Google, so I'd just take the bait 
and kill it. Vetoes?

Re: Scheduled removals

[Patrick McHardy]

Jan Engelhardt wrote:
> as per Documentation/feature-removal-schedule.txt, some legacy code is 
> ripe for removal. Now while pursuing this I wondered whether the struct 
> definitions for the old code in the header files (e.g. in 
> linux/netfilter/xt_CONNMARK.h) can also be removed or whether people 
> will argue that doing so would unnecessarily cause dismay on behalf the 
> potential users (of which there are not any). I am not aware of any 
> users, and neither is Search Engine Google, so I'd just take the bait 
> and kill it. Vetoes?

Every time I'm searching I do find some users of at least some
iptables headers. Maybe not this specific one, but that doesn't
mean much, fact is there might be users.

That being said, I'd be fine removing both the functionality
and the headers after a sufficiently long period. I'm not
really confident that the period has indeed be long enough
though. When did we schedule this for removal?

Re: Scheduled removals

[Jan Engelhardt]

On Monday 2009-01-12 05:32, Patrick McHardy wrote:
>> as per Documentation/feature-removal-schedule.txt, some legacy code is ripe
>> for removal. Now while pursuing this I wondered whether the struct definitions
>> for the old code in the header files (e.g. in linux/netfilter/xt_CONNMARK.h)
>> can also be removed or whether people will argue that doing so would
>> unnecessarily cause dismay on behalf the potential users (of which there are
>> not any). I am not aware of any users, and neither is Search Engine Google, so
>> I'd just take the bait and kill it. Vetoes?
>
> Every time I'm searching I do find some users of at least some
> iptables headers. Maybe not this specific one, but that doesn't
> mean much, fact is there might be users.
>
> That being said, I'd be fine removing both the functionality
> and the headers after a sufficiently long period. I'm not
> really confident that the period has indeed be long enough
> though. When did we schedule this for removal?

January 2008.

Re: Scheduled removals

[Patrick McHardy]

Jan Engelhardt wrote:
> On Monday 2009-01-12 05:32, Patrick McHardy wrote:
>>> as per Documentation/feature-removal-schedule.txt, some legacy code is ripe
>>> for removal. Now while pursuing this I wondered whether the struct definitions
>>> for the old code in the header files (e.g. in linux/netfilter/xt_CONNMARK.h)
>>> can also be removed or whether people will argue that doing so would
>>> unnecessarily cause dismay on behalf the potential users (of which there are
>>> not any). I am not aware of any users, and neither is Search Engine Google, so
>>> I'd just take the bait and kill it. Vetoes?
>> Every time I'm searching I do find some users of at least some
>> iptables headers. Maybe not this specific one, but that doesn't
>> mean much, fact is there might be users.
>>
>> That being said, I'd be fine removing both the functionality
>> and the headers after a sufficiently long period. I'm not
>> really confident that the period has indeed be long enough
>> though. When did we schedule this for removal?
> 
> January 2008.

OK, lets remove it in 2.6.30. At that point it should have been
almost 1.5 years, which seems to be enough.