From: Derrick Stolee <stolee@gmail.com>
To: Stefan Beller <sbeller@google.com>,
Derrick Stolee <dstolee@microsoft.com>
Cc: git <git@vger.kernel.org>
Subject: Re: [PATCH 3/9] midx: mark bad packed objects
Date: Tue, 21 Aug 2018 09:53:44 -0400 [thread overview]
Message-ID: <da0185cf-7d48-862b-eb25-5d5bd8f4e833@gmail.com> (raw)
In-Reply-To: <CAGZ79kY8i3y6_r=bQAx7V5=ckSmb_RgyByjJwtS+7Qr2y8x8LQ@mail.gmail.com>
On 8/20/2018 5:23 PM, Stefan Beller wrote:
> On Mon, Aug 20, 2018 at 9:52 AM Derrick Stolee <dstolee@microsoft.com> wrote:
>> When an object fails to decompress from a pack-file, we mark the object
>> as 'bad' so we can retry with a different copy of the object (if such a
>> copy exists).
>>
>> Before now, the multi-pack-index did not update the bad objects list for
>> the pack-files it contains, and we did not check the bad objects list
>> when reading an object. Now, do both.
> This sounds like a bug fix unlike patches 1 & 2 that sound like
> feature work(2) or making code more readable(1).
> (After studying the code, this doesn't sound like a bug fix any more,
> but a safety thing)
It is a safety thing. One codepath that needs this includes this comment:
/*
* We're probably in deep shit, but let's try to fetch
* the required base anyway from another pack or loose.
* This is costly but should happen only in the presence
* of a corrupted pack, and is better than failing outright.
*/
This goes back to 8eca0b47: implement some resilience against pack
corruptions. The logic is tested in t5303-pack-corruption-resilience.sh,
with the test title '... and a redundant pack allows for full recovery too'.
> Is it worth having this on a separate track coming in
> faster than the rest of this series?
ds/multi-pack-index is cooking in 'next' until after 2.19.0, so I'm not
sure it's worth splitting things up at this point.
>> Signed-off-by: Derrick Stolee <dstolee@microsoft.com>
>> ---
>> midx.c | 10 ++++++++++
>> 1 file changed, 10 insertions(+)
>>
>> diff --git a/midx.c b/midx.c
>> index 6824acf5f8..7fa75a37a3 100644
>> --- a/midx.c
>> +++ b/midx.c
>> @@ -280,6 +280,16 @@ static int nth_midxed_pack_entry(struct multi_pack_index *m, struct pack_entry *
>> if (!is_pack_valid(p))
>> return 0;
>>
>> + if (p->num_bad_objects) {
>> + uint32_t i;
> Is there a reason that i needs to be if 32 bits?
> Would size_t (for iterating) or 'int' (as a default
> like in many for loops) be ok, too?
i is bounded by p->num_bad_objects, which is also uint32_t.
Thanks,
-Stolee
next prev parent reply other threads:[~2018-08-21 13:53 UTC|newest]
Thread overview: 18+ messages / expand[flat|nested] mbox.gz Atom feed top
2018-08-20 16:51 [PATCH 0/9] multi-pack-index cleanups Derrick Stolee
2018-08-20 16:51 ` [PATCH 1/9] multi-pack-index: provide more helpful usage info Derrick Stolee
2018-08-20 16:51 ` [PATCH 2/9] multi-pack-index: store local property Derrick Stolee
2018-08-20 21:14 ` Stefan Beller
2018-08-20 16:51 ` [PATCH 3/9] midx: mark bad packed objects Derrick Stolee
2018-08-20 21:23 ` Stefan Beller
2018-08-21 13:53 ` Derrick Stolee [this message]
2018-08-20 16:51 ` [PATCH 4/9] midx: stop reporting garbage Derrick Stolee
2018-08-20 16:52 ` [PATCH 5/9] midx: fix bug that skips midx with alternates Derrick Stolee
2018-08-20 16:52 ` [PATCH 6/9] packfile: add all_packs list Derrick Stolee
2018-08-20 16:52 ` [PATCH 7/9] treewide: use get_all_packs Derrick Stolee
2018-08-20 22:01 ` Stefan Beller
2018-08-21 13:56 ` Derrick Stolee
2018-08-20 16:52 ` [PATCH 8/9] midx: test a few commands that " Derrick Stolee
2018-08-20 22:03 ` Stefan Beller
2018-08-20 16:52 ` [PATCH 9/9] pack-objects: consider packs in multi-pack-index Derrick Stolee
2018-08-21 14:34 ` [PATCH 0/9] multi-pack-index cleanups Duy Nguyen
2018-08-21 14:44 ` Derrick Stolee
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
List information: http://vger.kernel.org/majordomo-info.html
* Reply using the --to, --cc, and --in-reply-to
switches of git-send-email(1):
git send-email \
--in-reply-to=da0185cf-7d48-862b-eb25-5d5bd8f4e833@gmail.com \
--to=stolee@gmail.com \
--cc=dstolee@microsoft.com \
--cc=git@vger.kernel.org \
--cc=sbeller@google.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.
Code repositories for project(s) associated with this public inbox
https://80x24.org/mirrors/git.git
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).