Re: Suggest to upgrade some software in base

On 06/28/04 07:52, Oliver Eikemeier wrote:
> Jon Noack wrote:
>> On 06/27/04 12:02, David O'Brien wrote:
>>> On Sun, Jun 27, 2004 at 04:54:08PM +0200, Divacky Roman wrote:
>>>> I digged through our base system and looked for versions of
>>>> contributed soft. I found these program which could (and I think
>>>> should) be easily and painlessly upgraded (before 5.3 as 5-STABLE)
>>>> because they are outdated etc... these are:
>>>>
>>>> file - 3.41 ->  4.09
>>>> Painless upgrade and the benefit is much newer magic file
>>>> ftp://ftp.astron.com/pub/file/
>>>
>>> Only semi-painless.  The code and how it is built has changed around a
>>> lot, else I would have upgraded it by now.  That said, in progress; but
>>> lower priority than my toolchain work.
>>> And why does this have to happen before 5-STABLE?  I can certainly MFC
>>> something like this.
>>
>> Don't import until FILE 4.10 is released.  I've submitted a patch to 
>> Christos Zoulas for inclusion in 4.10 that *greatly* increases the 
>> accuracy of FILE for FreeBSD.  As soon as I see FILE 4.10 released 
>> (with my patch), I'll be pleading for an import...
> 
> Seems like Christos has been swamped by readelf.c patches :) I must 
> admit that I didn't care about 4.6.2...
> 
> Anyway, 4.09 is an improvement, and the import of 4.10 should be trivial 
> afterwards, so why wait?

I agree that 4.09 is a huge improvement (although it's wrong for FreeBSD 
4.10+, at least it correctly detects 5.x for the time being.  however, 
as soon as we we get 5-STABLE it'll still say it's -CURRENT).  The 
biggest issue will be the upgrade from 3.41 -> 4.x, so the 4.09 -> 4.10 
upgrade should be trivial as you say.

My only concern was conservation of limited developer resources.  If 
someone wants to import 4.09 with the intent of following up to 4.10, go 
for it.  However, in my opinion FILE 4.10 should be primary goal.

Jon