[nftables] TODO: Replace yy_switch_to_buffer by yypop_buffer_state and yypush_buffer_state

[nftables] TODO: Replace yy_switch_to_buffer by yypop_buffer_state and yypush_buffer_state

[Gopal Yadav]

Hi Netfilter Team,

I am looking to resolve a todo task in function yy_switch_to_buffer in
nftables/src/scanner.c file.
I am not familiar with lex but searching around I found that scanner.c
is produced by lex according to the scanner.l file.
Therefore my guess is, changing the scanner.c file directly is not the
solution, right?
Changes would have to be done in scanner.l, right?
How should I proceed to complete this todo?

I browsed bugzilla to find some other issues to solve, but I feel
lost. Are there any beginner friendly issues to solve or any other
starting point?

Thanks
Gopal

Re: [nftables] TODO: Replace yy_switch_to_buffer by yypop_buffer_state and yypush_buffer_state

[Florian Westphal]

Gopal Yadav <gopunop@gmail.com> wrote:
> Hi Netfilter Team,
> 
> I am looking to resolve a todo task in function yy_switch_to_buffer in
> nftables/src/scanner.c file.
> I am not familiar with lex but searching around I found that scanner.c
> is produced by lex according to the scanner.l file.
> Therefore my guess is, changing the scanner.c file directly is not the
> solution, right?

Right.

> Changes would have to be done in scanner.l, right?

No.

> How should I proceed to complete this todo?

This TODO is coming from flex itself, so, this is not an nftables
task.

> I browsed bugzilla to find some other issues to solve, but I feel
> lost. Are there any beginner friendly issues to solve or any other
> starting point?

This one for example:
https://bugzilla.netfilter.org/show_bug.cgi?id=1305

Its "just" a documentation issue.  You could work from comment 4
and improve the nft documentation to clarify 'accept' behaviour.

For many other bugs it would help if we had testcases that demonstrated
this problem in the nftables.git repo.

So, if you can translate a BZ ticker to e.g. a new test case in
tests/shell that show problem still exists in current nftables.git then
you could submit that test case as a patch, even if the problem is not
yet resolved.

Readily available test reproducers help a lot.

Re: [nftables] TODO: Replace yy_switch_to_buffer by yypop_buffer_state and yypush_buffer_state

[Gopal Yadav]

On Wed, Sep 9, 2020 at 5:53 PM Florian Westphal <fw@strlen.de> wrote:
>
> Gopal Yadav <gopunop@gmail.com> wrote:
> > Are there any beginner friendly issues to solve or any other
> > starting point?
>
> This one for example:
> https://bugzilla.netfilter.org/show_bug.cgi?id=1305
>
> Its "just" a documentation issue.  You could work from comment 4
> and improve the nft documentation to clarify 'accept' behaviour.
>
> For many other bugs it would help if we had testcases that demonstrated
> this problem in the nftables.git repo.
>
> So, if you can translate a BZ ticker to e.g. a new test case in
> tests/shell that show problem still exists in current nftables.git then
> you could submit that test case as a patch, even if the problem is not
> yet resolved.
>
> Readily available test reproducers help a lot.

Thanks for the reply. Pablo suggested to work on
https://bugzilla.netfilter.org/show_bug.cgi?id=1388
but I think his mail was not sent to the mailing list.
I am currently trying to solve that one. Will get back to look at 1305
after that.

Re: [nftables] TODO: Replace yy_switch_to_buffer by yypop_buffer_state and yypush_buffer_state

[Gopal Yadav]

On Wed, Sep 9, 2020 at 9:40 PM Gopal Yadav <gopunop@gmail.com> wrote:
>
> On Wed, Sep 9, 2020 at 5:53 PM Florian Westphal <fw@strlen.de> wrote:
> >
> > Gopal Yadav <gopunop@gmail.com> wrote:
> > > Are there any beginner friendly issues to solve or any other
> > > starting point?
> >
> > This one for example:
> > https://bugzilla.netfilter.org/show_bug.cgi?id=1305
> >
> > Its "just" a documentation issue.  You could work from comment 4
> > and improve the nft documentation to clarify 'accept' behaviour.
> >
> > For many other bugs it would help if we had testcases that demonstrated
> > this problem in the nftables.git repo.
> >
> > So, if you can translate a BZ ticker to e.g. a new test case in
> > tests/shell that show problem still exists in current nftables.git then
> > you could submit that test case as a patch, even if the problem is not
> > yet resolved.
> >
> > Readily available test reproducers help a lot.
>
> Thanks for the reply. Pablo suggested to work on
> https://bugzilla.netfilter.org/show_bug.cgi?id=1388
> but I think his mail was not sent to the mailing list.
> I am currently trying to solve that one. Will get back to look at 1305
> after that.

I have submitted a patch for 1388.
For the doc issue I would have to read the entire manpage.
Will try doing that meanwhile any other issue to work on?