All of lore.kernel.org
 help / color / mirror / Atom feed
From: David Turner <dturner@twopensource.com>
To: Thomas Gummerer <t.gummerer@gmail.com>
Cc: git@vger.kernel.org, mhagger@alum.mit.edu
Subject: Re: [PATCH v3 19/20] refs: add LMDB refs backend
Date: Tue, 19 Jan 2016 13:55:33 -0500	[thread overview]
Message-ID: <1453229733.16226.33.camel@twopensource.com> (raw)
In-Reply-To: <20160115133317.GJ10612@hank>

On Fri, 2016-01-15 at 14:33 +0100, Thomas Gummerer wrote:
> [I don't know too much about the refs code, but was still interested
> in this patch series, so below are the few things that I noticed
> while
> reading through it.]

Thanks for taking a look.

> On 01/14, David Turner wrote:

> > +exists.  The header has the same format as an ordinary reflog, but
> > with
> > +a timeztamp of all zeros and an empty value.
> 
> s/timeztamp/timestamp/

Fixed, thanks.

> > +test "$(git config extensions.refstorage)" = "lmdb" && die "git
> > -new-workdir is incompatible with the refs lmdb storage"
> > +
> 
> Is it expected that other potential ref backends are compatible with
> git-new-workdir?  Otherwise I think it would make more sense to
> whitelist the files backend here instead of blacklisting the lmdb
> backend, so we don't risk forgetting about this when adding another
> backend.

I no longer remember what the exact issue was, but we can be safe and
restrict git-new-workdir to the files backend (since it looks likely
that worktrees will replace it anyway).

> >  case "$git_dir" in
> >  .git)
> >  	git_dir="$orig_git/.git"
> 
> [...]
> 
> > +static int lmdb_init_db(struct strbuf *err, int shared)
> > +{
> > +	/*
> > +	 * To create a db, all we need to do is make a directory
> > for
> > +	 * it to live in; lmdb will do the rest.
> > +	 */
> > +
> > +	if (!db_path)
> > +		db_path =
> > xstrdup(real_path(get_refdb_path(get_git_common_dir())));
> 
> I think we're leaking some memory from get_refdb_path() here.
> get_refdb_path() uses strbuf_detach(), which according to its
> docstring
> makes its caller take care of the memory of the returned string.
> real_path() then uses strbuf_addstr() to add the string to its
> internal strbuf, but leaves the string we get from get_refdb_path()
> alone, so it leaks.

That only happens once per run, since db_path is set afterwards, but
I'll change get_refdb_path() to return a pointer to a statically
allocated string (as many similar funcs do).  That will be easier to
read.

  reply	other threads:[~2016-01-19 18:55 UTC|newest]

Thread overview: 49+ messages / expand[flat|nested]  mbox.gz  Atom feed  top
2016-01-14 16:25 [PATCH v3 00/20] refs backend rebase on pu David Turner
2016-01-14 16:25 ` [PATCH v3 01/20] refs: add a backend method structure with transaction functions David Turner
2016-01-14 16:25 ` [PATCH v3 02/20] refs: add methods for misc ref operations David Turner
2016-01-14 16:26 ` [PATCH v3 03/20] refs: add methods for the ref iterators David Turner
2016-01-14 16:26 ` [PATCH v3 04/20] refs: add do_for_each_per_worktree_ref David Turner
2016-01-14 16:26 ` [PATCH v3 05/20] refs: add methods for reflog David Turner
2016-01-14 16:26 ` [PATCH v3 06/20] refs: add method for initial ref transaction commit David Turner
2016-01-14 16:26 ` [PATCH v3 07/20] refs: add method for delete_refs David Turner
2016-01-14 16:26 ` [PATCH v3 08/20] refs: add methods to init refs db David Turner
2016-01-14 16:26 ` [PATCH v3 09/20] refs: add method to rename refs David Turner
2016-01-14 16:26 ` [PATCH v3 10/20] refs: make lock generic David Turner
2016-01-14 16:26 ` [PATCH v3 11/20] refs: move duplicate check to common code David Turner
2016-01-14 16:26 ` [PATCH v3 12/20] refs: allow log-only updates David Turner
2016-01-14 16:26 ` [PATCH v3 13/20] refs: resolve symbolic refs first David Turner
2016-02-04  7:37   ` Jeff King
2016-02-04 19:24     ` David Turner
2016-01-14 16:26 ` [PATCH v3 14/20] refs: always handle non-normal refs in files backend David Turner
2016-01-14 16:26 ` [PATCH v3 15/20] init: allow alternate backends to be set for new repos David Turner
2016-01-15 11:33   ` SZEDER Gábor
2016-01-15 12:51   ` Thomas Gummerer
2016-01-19 19:12     ` David Turner
2016-02-04  9:48   ` Duy Nguyen
2016-02-04 20:05     ` David Turner
2016-01-14 16:26 ` [PATCH v3 16/20] refs: check submodules ref storage config David Turner
2016-01-14 16:26 ` [PATCH v3 17/20] refs: allow ref backend to be set for clone David Turner
2016-01-15 11:32   ` SZEDER Gábor
2016-01-19 17:06     ` David Turner
2016-01-21  9:08       ` SZEDER Gábor
2016-01-14 16:26 ` [PATCH v3 18/20] svn: learn ref-storage argument David Turner
2016-01-15 11:34   ` SZEDER Gábor
2016-01-14 16:26 ` [PATCH v3 19/20] refs: add LMDB refs backend David Turner
2016-01-15 13:33   ` Thomas Gummerer
2016-01-19 18:55     ` David Turner [this message]
2016-02-04  9:58   ` Duy Nguyen
2016-02-04 19:33     ` David Turner
2016-01-14 16:26 ` [PATCH v3 20/20] refs: tests for lmdb backend David Turner
2016-02-02 20:08 ` [PATCH v3 00/20] refs backend rebase on pu David Turner
2016-02-02 22:13   ` Junio C Hamano
2016-02-04  0:09     ` Junio C Hamano
2016-02-04  1:12       ` David Turner
2016-02-04  1:54         ` Ramsay Jones
2016-02-04  2:58           ` Jeff King
2016-02-04 22:44             ` David Turner
2016-02-04 10:09   ` Duy Nguyen
2016-02-04 21:39     ` David Turner
2016-02-04 11:42   ` Duy Nguyen
2016-02-04 20:25     ` David Turner
2016-02-04 20:39       ` Ramsay Jones
2016-02-04 21:23         ` David Turner

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=1453229733.16226.33.camel@twopensource.com \
    --to=dturner@twopensource.com \
    --cc=git@vger.kernel.org \
    --cc=mhagger@alum.mit.edu \
    --cc=t.gummerer@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 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.