All of lore.kernel.org
 help / color / mirror / Atom feed
From: Florian Westphal <fw@strlen.de>
To: Eric Leblond <eric@regit.org>
Cc: Arturo Borrero Gonzalez <arturo@netfilter.org>,
	Pablo Neira Ayuso <pablo@netfilter.org>,
	Netfilter Development Mailing list
	<netfilter-devel@vger.kernel.org>
Subject: Re: [nft PATCH 0/16] introduce libnftables
Date: Thu, 17 Aug 2017 12:35:52 +0200	[thread overview]
Message-ID: <20170817103552.GA17221@breakpoint.cc> (raw)
In-Reply-To: <1502960318.31564.1.camel@regit.org>

Eric Leblond <eric@regit.org> wrote:

Thanks a lot for working on this Eric!

> On Thu, 2017-08-17 at 10:32 +0200, Arturo Borrero Gonzalez wrote:
> > On 16 August 2017 at 22:42, Eric Leblond <eric@regit.org> wrote:
> > > 
> > > Hello,
> > > 
> > > This patchset adds a basi high level libnftables to nftables code.
> > > It is currently supporting running a command from a buffer or from
> > > a file as well as batch support allowing to chain commands and
> > > commit
> > > them at once.
> > > 
> > > The API is mostly using existing structures such as nft_ctx that
> > > are
> > > updated to contain enough information. It also adds a structure
> > > dedicated to batch.
> > > 
> > 
> > Great work Eric, thanks!
> > 
> > Some comments below.
> > 
> > > A simple program running a command is the following:
> > > 
> > >         nft_global_init();
> > >         nft = nft_context_new();
> > >         nft_context_set_print_func(nft, my_print, buf);
> > 
> > ^^^
> > A minor thing: Did you evaluate merging these two? Setting the print
> > function directly when allocating a new context.
> 
> Nope but could make sense.

I'd recommend to keep it like this, else we can run into
problems when we need a new func.

If it stays this way we can simply add
nft_context_set_foo_func() instead of breaking nft_context_new() abi
or adding nft_context_new2() (ugh...).

> > On a side note, I remember in NFWS 2017 we discussed the possibility
> > of libnftables being a separate source project, i.e a standalone
> > repository.
> > Now that I see your patches, what I see is that libnftables is mostly
> > all the code, while nft itself is very little code.
> > Still, with my Debian hat, I think that different repositories is
> > good to have.
> 
> I don't like the cascade idea with nftables -> libnftables -> libnftnl
> -> libmnl that this will induce. Also this means some potential
> breakage in versionning.

I would also like to keep it in same repo, else i fear we will
quickly have copy&paste programming...

We can always split later if we think that nft and libnft have matured
in a way that they are distinct after all.


  reply	other threads:[~2017-08-17 10:38 UTC|newest]

Thread overview: 37+ messages / expand[flat|nested]  mbox.gz  Atom feed  top
2017-08-16 20:42 [nft PATCH 0/16] introduce libnftables Eric Leblond
2017-08-16 20:42 ` [nft PATH 01/16] libnftables: introduce library Eric Leblond
2017-08-17  8:57   ` Phil Sutter
2017-08-17 17:09     ` Eric Leblond
2017-08-17 17:13       ` Phil Sutter
2017-08-19  8:43         ` Eric Leblond
2017-08-19 19:07           ` Jan Engelhardt
2017-08-21  8:19             ` Pablo Neira Ayuso
2017-08-16 20:42 ` [nft PATH 02/16] libnftables: add context new and free Eric Leblond
2017-08-17  9:04   ` Phil Sutter
2017-08-16 20:42 ` [nft PATH 03/16] libnftables: add nft_run_command_from_buffer Eric Leblond
2017-08-17  9:21   ` Phil Sutter
2017-08-16 20:42 ` [nft PATH 04/16] libnftables: add nft_run_command_from_filename Eric Leblond
2017-08-16 20:42 ` [nft PATH 05/16] libnftables: put nft_run in library Eric Leblond
2017-08-16 20:43 ` [nft PATH 06/16] libnftables: add missing variable to library Eric Leblond
2017-08-17  9:35   ` Phil Sutter
2017-08-19 11:02     ` Eric Leblond
2017-08-16 20:43 ` [nft PATH 07/16] libnftables: add NFT_EXIT_* " Eric Leblond
2017-08-16 20:43 ` [nft PATH 08/16] libnftables: add a nft_cache to nft_ctx Eric Leblond
2017-08-17  9:43   ` Phil Sutter
2017-08-16 20:43 ` [nft PATH 09/16] libnftables: move iface_cache_release to deinit Eric Leblond
2017-08-16 20:43 ` [nft PATH 10/16] libnftables: get rid of printf Eric Leblond
2017-08-17 10:01   ` Phil Sutter
2017-08-19  8:59     ` Eric Leblond
2017-08-16 20:43 ` [nft PATH 11/16] libnftables: add nft_context_set_print Eric Leblond
2017-08-16 20:43 ` [nft PATH 12/16] libnftables: transaction support Eric Leblond
2017-08-17 10:11   ` Phil Sutter
2017-08-16 20:43 ` [nft PATH 13/16] libnftables: set max_errors to 1 in library Eric Leblond
2017-08-16 20:43 ` [nft PATH 14/16] erec: add function to free list Eric Leblond
2017-08-16 20:43 ` [nft PATH 15/16] libnftables: add error handling Eric Leblond
2017-08-17 10:32   ` Phil Sutter
2017-08-19  9:04     ` Eric Leblond
2017-08-16 20:43 ` [nft PATH 16/16] libnftables: basic doxygen documentation Eric Leblond
2017-08-17  8:32 ` [nft PATCH 0/16] introduce libnftables Arturo Borrero Gonzalez
2017-08-17  8:58   ` Eric Leblond
2017-08-17 10:35     ` Florian Westphal [this message]
2017-08-17 10:47   ` Phil Sutter

Reply instructions:

You may reply publicly to this message via plain-text email
using any one of the following methods:

* Save the following mbox file, import it into your mail client,
  and reply-to-all from there: mbox

  Avoid top-posting and favor interleaved quoting:
  https://en.wikipedia.org/wiki/Posting_style#Interleaved_style

* Reply using the --to, --cc, and --in-reply-to
  switches of git-send-email(1):

  git send-email \
    --in-reply-to=20170817103552.GA17221@breakpoint.cc \
    --to=fw@strlen.de \
    --cc=arturo@netfilter.org \
    --cc=eric@regit.org \
    --cc=netfilter-devel@vger.kernel.org \
    --cc=pablo@netfilter.org \
    /path/to/YOUR_REPLY

  https://kernel.org/pub/software/scm/git/docs/git-send-email.html

* If your mail client supports setting the In-Reply-To header
  via mailto: links, try the mailto: link
Be sure your reply has a Subject: header at the top and a blank line before the message body.
This is an external index of several public inboxes,
see mirroring instructions on how to clone and mirror
all data and code used by this external index.