Follow @Openwall on Twitter for new release announcements and other news
[<prev] [next>] [<thread-prev] [thread-next>] [day] [month] [year] [list]
Date: Wed, 23 Nov 2011 19:10:13 +0100
From: magnum <>
Subject: Re: 1.7.9-jumbo

2011-11-23 17:42, Solar Designer wrote:
> On Wed, Nov 23, 2011 at 09:13:40AM -0600, jfoug wrote:
>> When we start building J1, do we simply upconvert 1.7.8-j8, rebasing as
>> 1.7.9-j1?   If that is the plan, would we then roll up patches 0001- to
>> 0030- into a 1.7.9-j2 (new features, etc)?
> This would work well for me, but I suspect that it's extra effort given
> that magnum already has an updated tree.
> I am really not sure.  I need to be able to review those patches somehow
> before I sign the new jumbo patch and tarball, and a fully updated tree
> is not the easiest thing to review.

I think I'll upload "Jumbo RC" patches, one that fairly close equals
jumbo-8 (close enough that the dynamic format is broken, because of
changed base), a couple new ones (eg. a separate patch that fixes
dynamic_fmt for the new MD5_std.c) and one that closely resembles the 30
incremental patches originally for jumbo-8.

In the end you can call the sum of them "Jumbo-1" or not, they are just
split this way for reviewing purposes.

> BTW, for the icc-generated assembly files, I need PGP signatures from
> magnum I guess.  I sort of reviewed the one that got into -jumbo-8, but
> it was time-consuming and unreliable.  I mean, the wiki server might get
> compromised, which would let an intruder modify a previously uploaded
> file without me detecting that.  So I have to be double-checking those
> patches for lack of backdoors.  ...Yes, this could be another reason to
> switch to git.

If you have icc you can just re-generate them using "make intrinsics". I
would prefer that.


Powered by blists - more mailing lists

Confused about mailing lists and their use? Read about mailing lists on Wikipedia and check out these guidelines on proper formatting of your messages.