Re: [PATCH v16 00/14] Reftable support git-core

[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

 



Hi Han-Wen,

On Wed, 10 Jun 2020, Han-Wen Nienhuys wrote:

> On Wed, Jun 10, 2020 at 8:59 PM Johannes Schindelin
> <Johannes.Schindelin@xxxxxx> wrote:
>
> > > I am reasonably sure we use opendir() to iterate over existing files
> > > and subdirectories in a directory (e.g. in the codepaths to
> > > enumerate loose object files in .git/objects/??/ directories).
> > >
> > > I do not offhand know we also use fdopendir() elsewhere.  I strongly
> > > suspect we do not.  Perhaps some platforms do POSIX.1-2001 but not
> > > ready for POSIX.1-2008 or something silly like that?
> >
> > We don't. We also do not use `unlinkat()`. And we generally use
> > `remove_dir_recursively()` instead of implementing a separate version of
> > it that only handles one directory level ;-)
> >
> > This is what I needed in Git for Windows' `shears/pu` branch to make it
> > compile again:
>
> Thanks Dscho. Last time I folded a fix into
> github.com/google/reftable, you seemed quite upset.

I was not upset. I just pointed out how unnecessarily difficult and
cumbersome it is to have this ever-growing patch, with the _actual_ commit
history completely elsewhere.

My assessment has not changed, and the situation has not changed, and does
not look like it will change, so I'll just keep quiet from now on.

:-)

> How should I proceed with this fixup?

That's for you to decide, not for me. My suggestions in the past seem not
to have pleased you, so in the interest of keeping our relationship
friendly, I will refrain from adding more suggestions.

Ciao,
Johannes




[Index of Archives]     [Linux Kernel Development]     [Gcc Help]     [IETF Annouce]     [DCCP]     [Netdev]     [Networking]     [Security]     [V4L]     [Bugtraq]     [Yosemite]     [MIPS Linux]     [ARM Linux]     [Linux Security]     [Linux RAID]     [Linux SCSI]     [Fedora Users]

  Powered by Linux