[U-Boot-Users] Request to mailing list U-Boot-Users rejected

[Jerry Van Baren]

Mike Frysinger wrote:
> On Tuesday 11 December 2007, Mike Frysinger wrote:
>> On Tuesday 11 December 2007, Wolfgang Denk wrote:
>>> In message <200712112009.06587.sr at denx.de> you wrote:
>>>>>> Message body is too big: 40979 bytes with a limit of 40 KB
>>> ...
>>>
>>>> 100k limit, please.
>>> And what happens whith the next posting with a size of 100k + N bytes?
>> with that logic, why allow 40k.  take it down to 10k.  or 1k.
>>
>> posting valid patches that exceed 40k is not uncommon.  patches exceeding
>> 100k is much less uncommon.
> 
> actually, if you're NAK-ing this by yourself, then i dont see why you'ed NAK a 
> valid patch in the first place.
> 
> someone posts a log file or configuration file or something that exceeds the 
> limit makes sense.  people developing source code when patches are supposed 
> to be sent to the list does not.
> -mike

I would agree with Mike... have a soft limit of 40K where bounced 
messages go to Wolfgang and he authorizes "reasonable" patches. 
Alternatively, set the limit higher (100K) and cut it back if it gets 
abused.  I suspect it won't.

I don't know how many "oversize" (>40K) messages Wolfgang sees.  Judging 
from the small number of complaints about oversize bounces on the list, 
it doesn't seem to be many.

Discarding oversize patches on a hard numerical criteria rather than on 
a softer validity criteria is easy for decision making but hard on the 
developers.  Arbitrarily breaking (in *both* senses of the word) patches 
just to fit them under the 40K limit causes more work and results in 
*less* benefit.

I like Koha's motto: "Every time a patch falls on the floor, a kitten 
dies." <http://wiki.koha.org/doku.php?id=en:development:git_usage>

We should be encouraging patches of a reasonable size and should be 
flexible on the definition of "reasonable."  We are willing to use 3rd 
party code (e.g. linux driver code) that may not quite meet our coding 
standards because it is useful and expedient to not have to rewrite 
perfectly good code.  If we have reasonable flexibility on coding 
standards, why not on patch sizes???

My 2 cents,
gvb

P.S. courtesy of Ogden Nash :-)

THE KITTEN
----------
The trouble with a kitten is
THAT
Eventually it becomes a
CAT.