Re: [RFC/PATCH] read-cache: write all indexes with the same permissions

["Ævar Arnfjörð Bjarmason" <avarab@gmail.com>]

On Tue, Nov 13 2018, Duy Nguyen wrote:

> On Tue, Nov 13, 2018 at 4:32 PM Ævar Arnfjörð Bjarmason
> <avarab@gmail.com> wrote:
>> I won't have time to finish this today, as noted in
>> https://public-inbox.org/git/874lcl2e9t.fsf@evledraar.gmail.com/
>> there's a pretty major bug here in that we're now writing out literal
>> sharedindex_XXXXXX files.
>
> It's not the end of the world because create_tempfile opens with
> O_EXCL so if two processes try to create it at the same time, one will
> fail, but no corruption or such.

Right, no race there.

>> Obviously that needs to be fixed, and the fix is trivial, I can use
>> another one of the mks_*() functions with the same mode we use to
>> create the index.
>>
>> But we really ought to have tests for the bug this patch introduces,
>> and as noted in the E-Mail linked above we don't.
>>
>> So hopefully Duy or someone with more knowledge of the split index
>> will chime in to say what's missing there...
>
> I don't have any bright idea how to catch the literal _XXXXX file.
> It's a temporary file and will not last long enough for us to verify
> unless we intercept open() calls with LD_PRELOAD.

Sorry for being unclear. I don't mean how can we catch this specific
bug, that would be uninteresting and hard to test for.

I'm asking whether the bug in this patch isn't revealing an existing
issue with us not having any tests for N number of sharedindex.*
files. I.e. we have >1 of them, merge them and prune them, don't we?