From kde-devel  Fri Jul 31 08:00:51 2009
From: Thiago Macieira <thiago () kde ! org>
Date: Fri, 31 Jul 2009 08:00:51 +0000
To: kde-devel
Subject: Re: Strange backtrace
Message-Id: <200907311000.51420.thiago () kde ! org>
X-MARC-Message: https://marc.info/?l=kde-devel&m=124902740023357
MIME-Version: 1
Content-Type: multipart/mixed; boundary="--===============0541099074=="

--===============0541099074==
Content-type: multipart/signed; boundary=nextPart1381624.2tQMgH2mdY;
 protocol="application/pgp-signature"; micalg=pgp-sha1
Content-transfer-encoding: 7bit

--nextPart1381624.2tQMgH2mdY
Content-Type: Text/Plain;
  charset="utf-8"
Content-Transfer-Encoding: quoted-printable

Lubos Lunak wrote:
>On Thursday 30 of July 2009, Thiago Macieira wrote:
>> Albert Astals Cid wrote:
>> >A Dijous, 30 de juliol de 2009, Thiago Macieira va escriure:
>> >> That's supposed to mean that Git is good and SVN is bad. I will
>> >> deal with branches in Git, I will not deal with branches in SVN.
>> >
>> >Yeah but you are doing it the hard way, if you want me to help,
>> > switch to git.
>> >
>> >Of course it's your decision helping or not, but to me it seems
>> > pushing too much.
>>
>> Well, consider that I don't backport my fixes because I don't deal
>> with SVN branches.
>
> Yeah, it must be hard, the failing memory and all those
> difficult-to-use svnbackport scripts ...

Let's skip the fact that I consider *backporting* fixes to be the wrong=20
model in the first place.

The reason I don't use those scripts is that I use git-svn and tracking=20
multiple branches in our SVN with that is close to impossible. Yes, I=20
chose to use git-svn. And I did that because I gain a lot from it. I won't=
=20
go into the benefits of DVCS because you all already know it.

If the drawback is that I can't access older branches (which I never build=
=20
anyway), so be it.

Now, backporting is the wrong model. You shouldn't backport fixes to=20
previous releases. The fix should be applied to the earliest release it=20
makes sense, then merged into all later releases. This guarantees that all=
=20
newer releases contain all fixes from previous releases.

That's the model we use in Qt with Git. It's also been the same model that=
=20
we used with Perforce, albeit slightly different because of Perforce=20
limitations.

svnbackport is a port of cvsbackport to SVN. The reason we don't do the=20
proper model is because both CVS and SVN have lousy merge abilities (to=20
say nothing of branching). This will be one of the things I'll be most=20
happy about when we switch to Git -- more so than the DVCS itself.

=2D-=20
  Thiago Macieira  -  thiago (AT) macieira.info - thiago (AT) kde.org
    PGP/GPG: 0x6EF45358; fingerprint:
    E067 918B B660 DBD1 105C  966C 33F5 F005 6EF4 5358

--nextPart1381624.2tQMgH2mdY
Content-Type: application/pgp-signature; name=signature.asc 
Content-Description: This is a digitally signed message part.

-----BEGIN PGP SIGNATURE-----
Version: GnuPG v1.4.9 (GNU/Linux)

iD8DBQBKcqSzM/XwBW70U1gRArq9AKCSRDJb3MPlRJlOeY4RpoUNmtnjkQCfWHbF
E66aE499eJmswKGPPPYOLZo=
=73gy
-----END PGP SIGNATURE-----

--nextPart1381624.2tQMgH2mdY--

--===============0541099074==
Content-Type: text/plain; charset="us-ascii"
MIME-Version: 1.0
Content-Transfer-Encoding: 7bit
Content-Disposition: inline

 
>> Visit http://mail.kde.org/mailman/listinfo/kde-devel#unsub to unsubscribe <<

--===============0541099074==--