netfilter-devel.vger.kernel.org archive mirror
 help / color / mirror / Atom feed
From: Pablo Neira Ayuso <pablo@netfilter.org>
To: Arturo Borrero Gonzalez <arturo.borrero.glez@gmail.com>
Cc: Netfilter Development Mailing list <netfilter-devel@vger.kernel.org>
Subject: Re: [libnftables PATCH 4/6] internal: add a selector for parsing ops
Date: Wed, 8 Jan 2014 14:36:55 +0100	[thread overview]
Message-ID: <20140108133655.GA25887@localhost> (raw)
In-Reply-To: <CAOkSjBiOw10oK4stjz48yqF63+Y_8-hOoH-uUh4FqO2e7P2DVA@mail.gmail.com>

On Wed, Jan 08, 2014 at 01:31:44PM +0100, Arturo Borrero Gonzalez wrote:
> On 8 January 2014 00:29, Pablo Neira Ayuso <pablo@netfilter.org> wrote:
> >
> > You should pass a callback function instead, eg.
> >
> > static int nft_chain_json_parse(struct nft_chain *c, const char *json,
> >                                 struct nft_parse_err *err,
> >                                 json_t *(*jsonbuilder)(const void *input_data,
> >                                                        const char *treename,
> >                                                        struct nft_parse_err *e))
> 
> Ok.
> 
> > But I don't understand yet what you save (in terms of lines of code)
> > by using this aproach.
> 
> I avoid doing something like:
> 
> nft_*_parse() {
> switch (type)
>  if XML return xml_parse()
>  if JSON return json_parse()
> }
> 
> nft_*_parse_file() {
> switch (type)
>  if XML return xml_parse_file()
>  if JSON return json_parse_file()
> }
> 
> We double the format switch, and also two functions per format are
> needed to do build and parsing.
> Total = 6 functions heavily duplicating code.
> 
> With my approach, we have 1 function that decides which format to
> parse, and a one function per format to build and do parsing.
> Total = 3 functions, no duplicate code.

I guess this is saving you code, but I think this abstraction needs to
be refined, your functions:

* nft_mxml_do_build_tree
* nft_mxml_do_build_tree_file

look almost the same, only difference is mxmlLoadString / mxmlLoadFile.

I really think you can save more code and make this look better if you
rework the internal nft_*_json_parse functions to receive the xml/json
trees, eg.

static int nft_set_json_parse(struct nft_set *s, json_t tree
                              struct nft_parse_err *err)
{
        ...
}

So you nft_set_do_parse() creates the tree and pass it to it.

You can also add an enum like:

enum {
        NFT_PARSE_BUFFER,
        NFT_PARSE_FILE,
};

json_t *nft_json_build_tree(uint32_t type, void *data)
{
        json_t *tree;

        switch (type) {
        case NFT_PARSE_BUFFER:
                tree = nft_json_build_tree(data, ...);
                break;
        case NFT_PARSE_FILE:
                tree = nft_json_build_tree_file(data, ...);
                break;
        }

        return tree;
}

that you can pass this enum to nft_set_do_parse() to indicate how the
tree need to be build, eg.

statiuc int nft_set_do_parse(..., uint32_t format)
{
        switch (format) {
        ...
        case NFT_PARSE_JSON:
                tree = nft_json_build_tree(type, data);
                break;
        ...
        }
        ...
}

Where data is the file descriptor / buffer area.

  reply	other threads:[~2014-01-08 13:37 UTC|newest]

Thread overview: 17+ messages / expand[flat|nested]  mbox.gz  Atom feed  top
2014-01-07 11:47 [libnftables PATCH 0/6] parsing update Arturo Borrero Gonzalez
2014-01-07 11:47 ` [libnftables PATCH 1/6] mxml: add error reference of the top node Arturo Borrero Gonzalez
2014-01-07 23:16   ` Pablo Neira Ayuso
2014-01-07 11:47 ` [libnftables PATCH 2/6] set_elem: add json parsing to API Arturo Borrero Gonzalez
2014-01-07 23:16   ` Pablo Neira Ayuso
2014-01-07 11:47 ` [libnftables PATCH 3/6] internal: rework parsing symbol logic Arturo Borrero Gonzalez
2014-01-07 23:18   ` Pablo Neira Ayuso
2014-01-08 12:20     ` Arturo Borrero Gonzalez
2014-01-08 12:25       ` Pablo Neira Ayuso
2014-01-07 11:47 ` [libnftables PATCH 4/6] internal: add a selector for parsing ops Arturo Borrero Gonzalez
2014-01-07 23:29   ` Pablo Neira Ayuso
2014-01-08 12:31     ` Arturo Borrero Gonzalez
2014-01-08 13:36       ` Pablo Neira Ayuso [this message]
2014-01-07 11:47 ` [libnftables PATCH 5/6] parsing: add interface to parse from file Arturo Borrero Gonzalez
2014-01-07 11:47 ` [libnftables PATCH 6/6] tests: update tests with nft_*_parse_file() Arturo Borrero Gonzalez
2014-01-07 23:32   ` Pablo Neira Ayuso
2014-01-08 12:21     ` Arturo Borrero Gonzalez

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=20140108133655.GA25887@localhost \
    --to=pablo@netfilter.org \
    --cc=arturo.borrero.glez@gmail.com \
    --cc=netfilter-devel@vger.kernel.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 a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox;
as well as URLs for NNTP newsgroup(s).