[PATCH] waf: Fix the build on openbsd

[Michael Adam]

On 2015-02-09 at 15:36 +0100, Ralph Böhme wrote:
> On Mon, Feb 09, 2015 at 02:09:17PM +0100, Volker Lendecke wrote:
> > On Mon, Feb 09, 2015 at 12:41:56PM +0100, Ralph Böhme wrote:
> > > Subject: [PATCH 1/2] waf: add new file for wafadmin bugfixes
> > > 
> > > In order to facilitate differentation between
> > > 
> > > 1) pristine upstream waf in wafadmin
> > > 2) backported upstream commits/bugfixes in wafadmin
> > > 3) downstream bugfixes
> > > 
> > > add a new file samba_wafadmin_fixes.py which is restricted to only
> > > containing pure wafadmin bugfixes.
> > 
> > Why this is necessary?
> 
> I won't say it's necessary, but it may help seperating 2) and 3).
> 
> @metze, @obnox: can you please share your thoughts? :)

I agree with Ralph:  It is by no means necessary but I think this
is quite nice an idea.  I also sort of understand Volker's concern
of this being overkill, but I think the advantages outweigh the
disadvantages easily. Volker, remember that whenever we touched
wafadmin in the past, we started a discussion that we should not
do it. Whence the initial approach of changing wafsamba.
Of course it is, as you say, always possible to "just use git" to
compare our wafadmin and upstream, but at least currently this is
not super easy. At least Ralph and I have spent a significant time
comparing wafadmin from samba and upstream, with different versions.
Maybe you can do it just more easily or you don't mind doing silly
work over and over again instead of abstracting it away, but I
doubt that. ;-)

Also one thought regarding changing of wafsamba, which you
complained so bitterly about (I don't want to go in to half a
dozen of subthreads and reply to individual mails at this point):
I think that while you are right that code duplication is bad,
and is in general bad engineering, in cases like this it may
have advantages, in that it may even help keeping our setup
working across updates of waf itself without needing to deal
with patches not applying. Of course on the other hand
it may also conceal problems in an upgrade. But then, we are
not doing these updates especially frequently...

Coming back to the original problem that started this whole
thread, it is of course unfortunate that the patch that was
supposed to fix openbsd build problems lets your openbsd
build fail. We found the patch in upstream, it looked
reasonable, and we added it. It is our fault that we did
not test it on OpenBSD (I think we didn't). But we would
not have expected it to *break* OpenBSD... :-/

Cheers - Michael

-------------- next part --------------
A non-text attachment was scrubbed...
Name: not available
Type: application/pgp-signature
Size: 198 bytes
Desc: not available
URL: <http://lists.samba.org/pipermail/samba-technical/attachments/20150210/003ed8d8/attachment.pgp>