All of lore.kernel.org
 help / color / mirror / Atom feed
* Feature Request git clone shallow-include
@ 2019-02-20  0:05 Joe Enzminger
  2019-02-20  1:22 ` Duy Nguyen
  0 siblings, 1 reply; 6+ messages in thread
From: Joe Enzminger @ 2019-02-20  0:05 UTC (permalink / raw)
  To: git

Currently, git clone supports shallow-exclude=<tag-name>.  The client
will clone up to, but not including, the commit with the tag.

It would be useful to have the ability to include the commit with the
tag.  The suggestion would be to add a "shallow-include" options to
clone to support this behavior.

I have tried to use shallow-exclude with a follow on git fetch
--deepen=1, but it always returns "fatal: error in object; unshallow
<sha1>"

^ permalink raw reply	[flat|nested] 6+ messages in thread

* Re: Feature Request git clone shallow-include
  2019-02-20  0:05 Feature Request git clone shallow-include Joe Enzminger
@ 2019-02-20  1:22 ` Duy Nguyen
  2019-02-20 18:06   ` Joe Enzminger
  0 siblings, 1 reply; 6+ messages in thread
From: Duy Nguyen @ 2019-02-20  1:22 UTC (permalink / raw)
  To: Joe Enzminger; +Cc: Git Mailing List

On Wed, Feb 20, 2019 at 7:07 AM Joe Enzminger
<joe.enzminger@exactasystems.com> wrote:
>
> Currently, git clone supports shallow-exclude=<tag-name>.  The client
> will clone up to, but not including, the commit with the tag.
>
> It would be useful to have the ability to include the commit with the
> tag.  The suggestion would be to add a "shallow-include" options to
> clone to support this behavior.

So exclude the tag's parents and everything before, but keep the tag, correct?

I think if we support --shallow-exclude=<tag>^ then it should work the
way you want (if the tag is a normal merge you may need to add
--shallow-exclude=<tag>^2 as well). And you can do even fancier thing
like --shallow-exclude=<tag>~3 (i.e. exclude the  grand grand parent
of the tag, but keep the tag and grand parents). We will need to
restrict extended SHA-1 syntax to a safe subset of course.

> I have tried to use shallow-exclude with a follow on git fetch
> --deepen=1, but it always returns "fatal: error in object; unshallow
> <sha1>"
-- 
Duy

^ permalink raw reply	[flat|nested] 6+ messages in thread

* Re: Feature Request git clone shallow-include
  2019-02-20  1:22 ` Duy Nguyen
@ 2019-02-20 18:06   ` Joe Enzminger
  2019-02-21 13:06     ` Duy Nguyen
  0 siblings, 1 reply; 6+ messages in thread
From: Joe Enzminger @ 2019-02-20 18:06 UTC (permalink / raw)
  To: Duy Nguyen; +Cc: Git Mailing List

That is correct.  What you suggest is actually what I tried (using
sha-1 syntax).  For my purposes, excluding the tag's parent's but
including the tag is sufficient, but if is fairly straightforward to
extend support to the other use cases I'm sure someone would find is
useful.

Joe


On Tue, Feb 19, 2019 at 7:22 PM Duy Nguyen <pclouds@gmail.com> wrote:
>
> On Wed, Feb 20, 2019 at 7:07 AM Joe Enzminger
> <joe.enzminger@exactasystems.com> wrote:
> >
> > Currently, git clone supports shallow-exclude=<tag-name>.  The client
> > will clone up to, but not including, the commit with the tag.
> >
> > It would be useful to have the ability to include the commit with the
> > tag.  The suggestion would be to add a "shallow-include" options to
> > clone to support this behavior.
>
> So exclude the tag's parents and everything before, but keep the tag, correct?
>
> I think if we support --shallow-exclude=<tag>^ then it should work the
> way you want (if the tag is a normal merge you may need to add
> --shallow-exclude=<tag>^2 as well). And you can do even fancier thing
> like --shallow-exclude=<tag>~3 (i.e. exclude the  grand grand parent
> of the tag, but keep the tag and grand parents). We will need to
> restrict extended SHA-1 syntax to a safe subset of course.
>
> > I have tried to use shallow-exclude with a follow on git fetch
> > --deepen=1, but it always returns "fatal: error in object; unshallow
> > <sha1>"
> --
> Duy

^ permalink raw reply	[flat|nested] 6+ messages in thread

* Re: Feature Request git clone shallow-include
  2019-02-20 18:06   ` Joe Enzminger
@ 2019-02-21 13:06     ` Duy Nguyen
  2019-03-29 22:02       ` Joe Enzminger
  0 siblings, 1 reply; 6+ messages in thread
From: Duy Nguyen @ 2019-02-21 13:06 UTC (permalink / raw)
  To: Joe Enzminger; +Cc: Git Mailing List

On Thu, Feb 21, 2019 at 1:07 AM Joe Enzminger
<joe.enzminger@exactasystems.com> wrote:
>
> That is correct.  What you suggest is actually what I tried (using
> sha-1 syntax).  For my purposes, excluding the tag's parent's but
> including the tag is sufficient, but if is fairly straightforward to
> extend support to the other use cases I'm sure someone would find is
> useful.

It's not hard to do. I hope I will find some time to do it soon. My
only concern is whether reuse the current code or write new. The
former makes it easy to accidentally accept some extended sha-1 syntax
that should not run on the server side. On the other hand, the latter
will not be as thoroughly tested because it only runs by shallow code.
That's my problem though. I think I might be able to find a third
option somewhere in between.

>
> Joe
>
>
> On Tue, Feb 19, 2019 at 7:22 PM Duy Nguyen <pclouds@gmail.com> wrote:
> >
> > On Wed, Feb 20, 2019 at 7:07 AM Joe Enzminger
> > <joe.enzminger@exactasystems.com> wrote:
> > >
> > > Currently, git clone supports shallow-exclude=<tag-name>.  The client
> > > will clone up to, but not including, the commit with the tag.
> > >
> > > It would be useful to have the ability to include the commit with the
> > > tag.  The suggestion would be to add a "shallow-include" options to
> > > clone to support this behavior.
> >
> > So exclude the tag's parents and everything before, but keep the tag, correct?
> >
> > I think if we support --shallow-exclude=<tag>^ then it should work the
> > way you want (if the tag is a normal merge you may need to add
> > --shallow-exclude=<tag>^2 as well). And you can do even fancier thing
> > like --shallow-exclude=<tag>~3 (i.e. exclude the  grand grand parent
> > of the tag, but keep the tag and grand parents). We will need to
> > restrict extended SHA-1 syntax to a safe subset of course.
> >
> > > I have tried to use shallow-exclude with a follow on git fetch
> > > --deepen=1, but it always returns "fatal: error in object; unshallow
> > > <sha1>"
> > --
> > Duy



-- 
Duy

^ permalink raw reply	[flat|nested] 6+ messages in thread

* Re: Feature Request git clone shallow-include
  2019-02-21 13:06     ` Duy Nguyen
@ 2019-03-29 22:02       ` Joe Enzminger
  2019-03-30 11:38         ` Duy Nguyen
  0 siblings, 1 reply; 6+ messages in thread
From: Joe Enzminger @ 2019-03-29 22:02 UTC (permalink / raw)
  To: Duy Nguyen; +Cc: Git Mailing List

Duy -

Any updates on this feature request?

Joe

On Thu, Feb 21, 2019 at 7:06 AM Duy Nguyen <pclouds@gmail.com> wrote:
>
> On Thu, Feb 21, 2019 at 1:07 AM Joe Enzminger
> <joe.enzminger@exactasystems.com> wrote:
> >
> > That is correct.  What you suggest is actually what I tried (using
> > sha-1 syntax).  For my purposes, excluding the tag's parent's but
> > including the tag is sufficient, but if is fairly straightforward to
> > extend support to the other use cases I'm sure someone would find is
> > useful.
>
> It's not hard to do. I hope I will find some time to do it soon. My
> only concern is whether reuse the current code or write new. The
> former makes it easy to accidentally accept some extended sha-1 syntax
> that should not run on the server side. On the other hand, the latter
> will not be as thoroughly tested because it only runs by shallow code.
> That's my problem though. I think I might be able to find a third
> option somewhere in between.
>
> >
> > Joe
> >
> >
> > On Tue, Feb 19, 2019 at 7:22 PM Duy Nguyen <pclouds@gmail.com> wrote:
> > >
> > > On Wed, Feb 20, 2019 at 7:07 AM Joe Enzminger
> > > <joe.enzminger@exactasystems.com> wrote:
> > > >
> > > > Currently, git clone supports shallow-exclude=<tag-name>.  The client
> > > > will clone up to, but not including, the commit with the tag.
> > > >
> > > > It would be useful to have the ability to include the commit with the
> > > > tag.  The suggestion would be to add a "shallow-include" options to
> > > > clone to support this behavior.
> > >
> > > So exclude the tag's parents and everything before, but keep the tag, correct?
> > >
> > > I think if we support --shallow-exclude=<tag>^ then it should work the
> > > way you want (if the tag is a normal merge you may need to add
> > > --shallow-exclude=<tag>^2 as well). And you can do even fancier thing
> > > like --shallow-exclude=<tag>~3 (i.e. exclude the  grand grand parent
> > > of the tag, but keep the tag and grand parents). We will need to
> > > restrict extended SHA-1 syntax to a safe subset of course.
> > >
> > > > I have tried to use shallow-exclude with a follow on git fetch
> > > > --deepen=1, but it always returns "fatal: error in object; unshallow
> > > > <sha1>"
> > > --
> > > Duy
>
>
>
> --
> Duy

^ permalink raw reply	[flat|nested] 6+ messages in thread

* Re: Feature Request git clone shallow-include
  2019-03-29 22:02       ` Joe Enzminger
@ 2019-03-30 11:38         ` Duy Nguyen
  0 siblings, 0 replies; 6+ messages in thread
From: Duy Nguyen @ 2019-03-30 11:38 UTC (permalink / raw)
  To: Joe Enzminger; +Cc: Git Mailing List

On Sat, Mar 30, 2019 at 5:02 AM Joe Enzminger
<joe.enzminger@exactasystems.com> wrote:
>
> Duy -
>
> Any updates on this feature request?

Nope. I've been busy with other stuff. I did have a look at the
possibility of reusing code in sha1-name.c and concluded that it's not
quite safe.

> Joe
>
> On Thu, Feb 21, 2019 at 7:06 AM Duy Nguyen <pclouds@gmail.com> wrote:
> >
> > On Thu, Feb 21, 2019 at 1:07 AM Joe Enzminger
> > <joe.enzminger@exactasystems.com> wrote:
> > >
> > > That is correct.  What you suggest is actually what I tried (using
> > > sha-1 syntax).  For my purposes, excluding the tag's parent's but
> > > including the tag is sufficient, but if is fairly straightforward to
> > > extend support to the other use cases I'm sure someone would find is
> > > useful.
> >
> > It's not hard to do. I hope I will find some time to do it soon. My
> > only concern is whether reuse the current code or write new. The
> > former makes it easy to accidentally accept some extended sha-1 syntax
> > that should not run on the server side. On the other hand, the latter
> > will not be as thoroughly tested because it only runs by shallow code.
> > That's my problem though. I think I might be able to find a third
> > option somewhere in between.
> >
> > >
> > > Joe
> > >
> > >
> > > On Tue, Feb 19, 2019 at 7:22 PM Duy Nguyen <pclouds@gmail.com> wrote:
> > > >
> > > > On Wed, Feb 20, 2019 at 7:07 AM Joe Enzminger
> > > > <joe.enzminger@exactasystems.com> wrote:
> > > > >
> > > > > Currently, git clone supports shallow-exclude=<tag-name>.  The client
> > > > > will clone up to, but not including, the commit with the tag.
> > > > >
> > > > > It would be useful to have the ability to include the commit with the
> > > > > tag.  The suggestion would be to add a "shallow-include" options to
> > > > > clone to support this behavior.
> > > >
> > > > So exclude the tag's parents and everything before, but keep the tag, correct?
> > > >
> > > > I think if we support --shallow-exclude=<tag>^ then it should work the
> > > > way you want (if the tag is a normal merge you may need to add
> > > > --shallow-exclude=<tag>^2 as well). And you can do even fancier thing
> > > > like --shallow-exclude=<tag>~3 (i.e. exclude the  grand grand parent
> > > > of the tag, but keep the tag and grand parents). We will need to
> > > > restrict extended SHA-1 syntax to a safe subset of course.
> > > >
> > > > > I have tried to use shallow-exclude with a follow on git fetch
> > > > > --deepen=1, but it always returns "fatal: error in object; unshallow
> > > > > <sha1>"
> > > > --
> > > > Duy
> >
> >
> >
> > --
> > Duy



-- 
Duy

^ permalink raw reply	[flat|nested] 6+ messages in thread

end of thread, other threads:[~2019-03-30 11:39 UTC | newest]

Thread overview: 6+ messages (download: mbox.gz / follow: Atom feed)
-- links below jump to the message on this page --
2019-02-20  0:05 Feature Request git clone shallow-include Joe Enzminger
2019-02-20  1:22 ` Duy Nguyen
2019-02-20 18:06   ` Joe Enzminger
2019-02-21 13:06     ` Duy Nguyen
2019-03-29 22:02       ` Joe Enzminger
2019-03-30 11:38         ` Duy Nguyen

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.