public inbox for binutils@sourceware.org
 help / color / mirror / Atom feed
From: Luca Boccassi <bluca@debian.org>
To: "Jose E. Marchesi" <jose.marchesi@oracle.com>
Cc: binutils@sourceware.org, Nick Clifton <nickc@redhat.com>
Subject: Re: [PATCH] ld: add --package-metadata
Date: Tue, 24 May 2022 12:26:54 +0100	[thread overview]
Message-ID: <CAMw=ZnSNWUJeB2di0JMhy5mB_Tv2ELhZU-B4ED9Vy7s6e3AGjA@mail.gmail.com> (raw)
In-Reply-To: <87y1yr1dgy.fsf@oracle.com>

On Tue, 24 May 2022 at 12:13, Jose E. Marchesi <jose.marchesi@oracle.com> wrote:
>
>
> Hi Luca.
>
> >> I've added an optional usage of libjansson to provide additional validation,
> >> so that it doesn't have to be reimplemented everywhere a thousand
> >> times over.
> >> For bootstrapping purposes the dep is optional and can be skipped, for example
> >> in Debian/Ubuntu it would be annotated with <!stage1> so that everything
> >> works out of the box.
>
> As far as I can see at https://systemd.io/COREDUMP_PACKAGE_METADATA/ the
> payload encoded in JSON comprises a single object which must be a
> dictionary, whose entries all contain strings.  Basically:
>
>   {"..." : "..."[, "..." : "..."]*}
>
> Isn't it a bit overkill to add a dependency to an external library just
> to verify the format of the above?  A regular expression would suffice,
> even if a fastidious one due to the string format:
>
>   S -> JSON string
>   B -> [\n\t ]*
>
>   PAYLOAD -> B{BSB:BSB(,SB:BSB)*}B

Hi,

The specification actually is very explicit in supporting a full json
payload, with the only limitations (certain unicode characters, int
ranges) being explicitly called out. The example given is simple
because it's just an example. I do expect the usage to expand at some
point in the future.

As mentioned in the patch the validation is entirely optional, and
nothing will happen if the build dependency isn't provided, which will
be the default in all distros. And it can also be explicitly disabled
if desired. I could flip the default to force-disable if desired, but
in practice I don't think it would make much difference, given how
packages are built in minimal environments already.

As a distribution builder I would very very much like to find out
about build-time issues at build time, rather than at runtime. It's
much cheaper and immediate. When you have 50000+ packages it is very
much unrealistic to expect to runtime QA them all for every little
thing like this. An automated, implicit build time check is so much
cheaper, reliable and effective, and reduces the need to reimplement
it every single time.

Kind regards,
Luca Boccassi

  reply	other threads:[~2022-05-24 11:27 UTC|newest]

Thread overview: 22+ messages / expand[flat|nested]  mbox.gz  Atom feed  top
2022-05-15 19:18 luca.boccassi
2022-05-16 16:40 ` Fangrui Song
2022-05-17  6:03   ` Florian Weimer
2022-05-17 14:44     ` Luca Boccassi
2022-05-25  6:56       ` Fangrui Song
2022-05-25  7:53         ` Florian Weimer
2022-05-23 11:26 ` Luca Boccassi
2022-05-24  9:34   ` Jose E. Marchesi
2022-05-24 11:26     ` Luca Boccassi [this message]
2022-05-24 16:23 ` Nick Clifton
2022-05-24 18:38   ` Luca Boccassi
2022-05-25  8:45     ` Florian Weimer
2022-05-25 13:53       ` Luca Boccassi
2022-05-30 14:08         ` Florian Weimer
2022-05-24 16:28 ` Nick Clifton
2022-05-24 21:15 ` [PATCH v2] " luca.boccassi
2022-05-25  4:30   ` Alan Modra
2022-05-25  6:02     ` Alan Modra
2022-05-25 13:42       ` Luca Boccassi
2022-05-25 13:41 ` [PATCH v3] " luca.boccassi
2022-05-26  3:55   ` Alan Modra
2022-05-26 10:46     ` Luca Boccassi

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='CAMw=ZnSNWUJeB2di0JMhy5mB_Tv2ELhZU-B4ED9Vy7s6e3AGjA@mail.gmail.com' \
    --to=bluca@debian.org \
    --cc=binutils@sourceware.org \
    --cc=jose.marchesi@oracle.com \
    --cc=nickc@redhat.com \
    /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 read-only IMAP folder(s) and NNTP newsgroup(s).