git.vger.kernel.org archive mirror
 help / color / mirror / Atom feed
From: Jakub Narebski <jnareb@gmail.com>
To: Junio C Hamano <gitster@pobox.com>
Cc: git@vger.kernel.org, "Petr Baudis" <pasky@suse.cz>,
	"Alex Riesen" <raa.lkml@gmail.com>,
	"Ævar Arnfjörð Bjarmason" <avarab@gmail.com>
Subject: Re: [1.8.0] perl/Git.pm: moving away from using Error.pm module
Date: Mon, 21 Feb 2011 10:31:09 +0100	[thread overview]
Message-ID: <201102211031.11308.jnareb@gmail.com> (raw)
In-Reply-To: <7v4o7xluph.fsf@alter.siamese.dyndns.org>

On Mon, 21 Feb 2011, Junio C Hamano wrote:
> Jakub Narebski <jnareb@gmail.com> writes:
> 
> > Proposal:
> >
> > Replace use of Error.pm module in Git.pm with either Exception::Class
> > based error class, or using 'carp'/'croak' from Carp, or both by adding 
> > an option to set error handler in 'Git' class (like e.g. in 'CHI' 
> > module on CPAN).
> 
> Personally, I was never a big fan of the syntax magic with Error.pm, but I
> refrained from commenting on it as I am not heavily involved in that part
> of the system.  If we are going to change things so that everybody uses a
> more traditional "eval {}; if ($@) { ... }", it would be a welcome change
> from my point of view.

Structured exceptions are usually better than 'die <string>' if
you need to examine error in more detail and act on this detail:
  http://www.modernperlbooks.com/mt/2010/10/structured-data-and-knowing-versus-guessing.html
  http://www.modernperlbooks.com/mt/2010/08/the-stringceptional-difficulty-of-changing-error-messages.html
  http://www.modernperlbooks.com/mt/2010/07/dont-parse-that-string.html

I think that is why Git.pm uses Error module (which seemed like a good
choice in 2006), and it is why I propose using Exception::Class, perhaps
as an option. 

> > Migration plan:
> 
> Do we even need one?
> 
> As far as an external caller is concerned, it would have been expecting us
> to throw an exception by dying, and it wouldn't have mattered if it used
> Error.pm or "eval { $call_to_Git_pm }; if ($@) {...}", I think.

Well, it depends if external scripts use try_git_cmd sugar... and whether
they try to act on details of caught exception.

Also if Git->repository would return 'undef' on failure instead of throwing
an exception, this would require changes to external scripts.

-- 
Jakub Narebski
Poland

  reply	other threads:[~2011-02-21  9:31 UTC|newest]

Thread overview: 7+ messages / expand[flat|nested]  mbox.gz  Atom feed  top
2011-02-20 22:46 [1.8.0] perl/Git.pm: moving away from using Error.pm module Jakub Narebski
2011-02-21  7:20 ` Junio C Hamano
2011-02-21  9:31   ` Jakub Narebski [this message]
2011-02-21 11:02   ` Nick
2011-02-21 12:12     ` Jakub Narebski
2011-02-21 12:31     ` Ævar Arnfjörð Bjarmason
2011-04-15 23:35       ` Avner

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=201102211031.11308.jnareb@gmail.com \
    --to=jnareb@gmail.com \
    --cc=avarab@gmail.com \
    --cc=git@vger.kernel.org \
    --cc=gitster@pobox.com \
    --cc=pasky@suse.cz \
    --cc=raa.lkml@gmail.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 NNTP newsgroup(s).