[BUG] `git reset --hard` fails with `update = none` submodules

# What did you do before the bug happened? (Steps to reproduce your issue)
1. Clone a git repository that sets `update = none` in .gitmodules:
$ git clone --recurse-submodules https://github.com/ubolonton/tree-sitter-langs

2. Perform a hard reset:
$ cd tree-sitter-langs
$ git reset --hard

# What did you expect to happen? (Expected behavior)
The reset should succeed and do nothing.

# What happened instead? (Actual behavior)
The reset command fails with

fatal: not a git repository: ../../.git/modules/repos/agda
fatal: could not reset submodule index

[System Info]
git version:
git version 2.32.0
cpu: x86_64
no commit associated with this build
sizeof-long: 8
sizeof-size_t: 8
shell-path: /bin/sh
uname: Linux 5.12.10-hardened1-1-hardened #1 SMP PREEMPT Thu, 10 Jun
2021 21:12:42 +0000 x86_64
compiler info: gnuc: 11.1
libc info: glibc: 2.33
$SHELL (typically, interactive shell): /usr/bin/zsh

On 2021-06-16 at 00:16:06, Rose Kunkel wrote:

# What did you do before the bug happened? (Steps to reproduce your issue)
1. Clone a git repository that sets `update = none` in .gitmodules:
$ git clone --recurse-submodules https://github.com/ubolonton/tree-sitter-langs

2. Perform a hard reset:
$ cd tree-sitter-langs
$ git reset --hard

# What did you expect to happen? (Expected behavior)
The reset should succeed and do nothing.

I think we're in agreement on this.  This should be a fresh clone and so
a hard reset should change nothing.

# What happened instead? (Actual behavior)
The reset command fails with

fatal: not a git repository: ../../.git/modules/repos/agda
fatal: could not reset submodule index

Hmmm, I can't reproduce this behavior.  What I see is this:

  $ git reset --hard
  HEAD is now at 5d362ce Release 0.10.0

I'm running git version 2.32.0.272.g935e593368 on Debian sid (with the
experimental packages).

Can you try the clone and run a "git status" command in the repository
to see if anything is modified after your clone?  Are the submodules
checked out when you perform the clone?  In my case, I see lines like
this:

  Skipping submodule 'repos/agda'

If you're seeing something different, then that might contribute to the
different behavior we're seeing.
-- 
brian m. carlson (he/him or they/them)
Toronto, Ontario, CA

Running `git status` in the resulting repository gives

On branch master
Your branch is up to date with 'origin/master'.

nothing to commit, working tree clean

This is the output from the clone command:

Cloning into 'tree-sitter-langs'...
remote: Enumerating objects: 609, done.
remote: Counting objects: 100% (83/83), done.
remote: Compressing objects: 100% (52/52), done.
remote: Total 609 (delta 40), reused 58 (delta 24), pack-reused 526
Receiving objects: 100% (609/609), 117.17 KiB | 1.05 MiB/s, done.
Resolving deltas: 100% (322/322), done.
Submodule 'repos/agda'
(https://github.com/tree-sitter/tree-sitter-agda) registered for path
'repos/agda'
Submodule 'repos/bash'
(https://github.com/tree-sitter/tree-sitter-bash) registered for path
'repos/bash'
Submodule 'repos/c' (https://github.com/tree-sitter/tree-sitter-c)
registered for path 'repos/c'
Submodule 'repos/c-sharp'
(https://github.com/tree-sitter/tree-sitter-c-sharp) registered for
path 'repos/c-sharp'
Submodule 'repos/cpp' (https://github.com/tree-sitter/tree-sitter-cpp)
registered for path 'repos/cpp'
Submodule 'repos/css' (https://github.com/tree-sitter/tree-sitter-css)
registered for path 'repos/css'
Submodule 'repos/elm' (https://github.com/razzeee/tree-sitter-elm)
registered for path 'repos/elm'
Submodule 'repos/fluent'
(https://github.com/tree-sitter/tree-sitter-fluent) registered for
path 'repos/fluent'
Submodule 'repos/go' (https://github.com/tree-sitter/tree-sitter-go)
registered for path 'repos/go'
Submodule 'repos/html'
(https://github.com/tree-sitter/tree-sitter-html) registered for path
'repos/html'
Submodule 'repos/janet-simple'
(https://codeberg.org/sogaiu/tree-sitter-janet-simple) registered for
path 'repos/janet-simple'
Submodule 'repos/java'
(https://github.com/tree-sitter/tree-sitter-java) registered for path
'repos/java'
Submodule 'repos/javascript'
(https://github.com/tree-sitter/tree-sitter-javascript) registered for
path 'repos/javascript'
Submodule 'repos/jsdoc'
(https://github.com/tree-sitter/tree-sitter-jsdoc) registered for path
'repos/jsdoc'
Submodule 'repos/json'
(https://github.com/tree-sitter/tree-sitter-json) registered for path
'repos/json'
Submodule 'repos/julia'
(https://github.com/tree-sitter/tree-sitter-julia) registered for path
'repos/julia'
Submodule 'repos/ocaml'
(https://github.com/tree-sitter/tree-sitter-ocaml) registered for path
'repos/ocaml'
Submodule 'repos/php' (https://github.com/tree-sitter/tree-sitter-php)
registered for path 'repos/php'
Submodule 'repos/python'
(https://github.com/tree-sitter/tree-sitter-python) registered for
path 'repos/python'
Submodule 'repos/ruby'
(https://github.com/tree-sitter/tree-sitter-ruby) registered for path
'repos/ruby'
Submodule 'repos/rust'
(https://github.com/tree-sitter/tree-sitter-rust) registered for path
'repos/rust'
Submodule 'repos/scala'
(https://github.com/tree-sitter/tree-sitter-scala) registered for path
'repos/scala'
Submodule 'repos/swift'
(https://github.com/tree-sitter/tree-sitter-swift) registered for path
'repos/swift'
Submodule 'repos/typescript'
(https://github.com/tree-sitter/tree-sitter-typescript) registered for
path 'repos/typescript'
Skipping submodule 'repos/agda'
Skipping submodule 'repos/bash'
Skipping submodule 'repos/c'
Skipping submodule 'repos/c-sharp'
Skipping submodule 'repos/cpp'
Skipping submodule 'repos/css'
Skipping submodule 'repos/elm'
Skipping submodule 'repos/fluent'
Skipping submodule 'repos/go'
Skipping submodule 'repos/html'
Skipping submodule 'repos/janet-simple'
Skipping submodule 'repos/java'
Skipping submodule 'repos/javascript'
Skipping submodule 'repos/jsdoc'
Skipping submodule 'repos/json'
Skipping submodule 'repos/julia'
Skipping submodule 'repos/ocaml'
Skipping submodule 'repos/php'
Skipping submodule 'repos/python'
Skipping submodule 'repos/ruby'
Skipping submodule 'repos/rust'
Skipping submodule 'repos/scala'
Skipping submodule 'repos/swift'
Skipping submodule 'repos/typescript'

On Tue, Jun 15, 2021 at 5:51 PM brian m. carlson
[off-list ref] wrote:

On 2021-06-16 at 00:16:06, Rose Kunkel wrote:

quoted

# What did you do before the bug happened? (Steps to reproduce your issue)
1. Clone a git repository that sets `update = none` in .gitmodules:
$ git clone --recurse-submodules https://github.com/ubolonton/tree-sitter-langs

2. Perform a hard reset:
$ cd tree-sitter-langs
$ git reset --hard

# What did you expect to happen? (Expected behavior)
The reset should succeed and do nothing.

I think we're in agreement on this.  This should be a fresh clone and so
a hard reset should change nothing.

quoted

# What happened instead? (Actual behavior)
The reset command fails with

fatal: not a git repository: ../../.git/modules/repos/agda
fatal: could not reset submodule index

Hmmm, I can't reproduce this behavior.  What I see is this:

  $ git reset --hard
  HEAD is now at 5d362ce Release 0.10.0

I'm running git version 2.32.0.272.g935e593368 on Debian sid (with the
experimental packages).

Can you try the clone and run a "git status" command in the repository
to see if anything is modified after your clone?  Are the submodules
checked out when you perform the clone?  In my case, I see lines like
this:

  Skipping submodule 'repos/agda'

If you're seeing something different, then that might contribute to the
different behavior we're seeing.
--
brian m. carlson (he/him or they/them)
Toronto, Ontario, CA

Potentially relevant: `git config --global --list` shows

status.showstash=true
status.submodulesummary=true
submodule.recurse=true
user.name=Rose Kunkel
user.email=rose@rosekunkel.me
pull.rebase=false
init.defaultbranch=main

On Tue, Jun 15, 2021 at 5:57 PM Rose Kunkel [off-list ref] wrote:

Running `git status` in the resulting repository gives

On branch master
Your branch is up to date with 'origin/master'.

nothing to commit, working tree clean

This is the output from the clone command:

Cloning into 'tree-sitter-langs'...
remote: Enumerating objects: 609, done.
remote: Counting objects: 100% (83/83), done.
remote: Compressing objects: 100% (52/52), done.
remote: Total 609 (delta 40), reused 58 (delta 24), pack-reused 526
Receiving objects: 100% (609/609), 117.17 KiB | 1.05 MiB/s, done.
Resolving deltas: 100% (322/322), done.
Submodule 'repos/agda'
(https://github.com/tree-sitter/tree-sitter-agda) registered for path
'repos/agda'
Submodule 'repos/bash'
(https://github.com/tree-sitter/tree-sitter-bash) registered for path
'repos/bash'
Submodule 'repos/c' (https://github.com/tree-sitter/tree-sitter-c)
registered for path 'repos/c'
Submodule 'repos/c-sharp'
(https://github.com/tree-sitter/tree-sitter-c-sharp) registered for
path 'repos/c-sharp'
Submodule 'repos/cpp' (https://github.com/tree-sitter/tree-sitter-cpp)
registered for path 'repos/cpp'
Submodule 'repos/css' (https://github.com/tree-sitter/tree-sitter-css)
registered for path 'repos/css'
Submodule 'repos/elm' (https://github.com/razzeee/tree-sitter-elm)
registered for path 'repos/elm'
Submodule 'repos/fluent'
(https://github.com/tree-sitter/tree-sitter-fluent) registered for
path 'repos/fluent'
Submodule 'repos/go' (https://github.com/tree-sitter/tree-sitter-go)
registered for path 'repos/go'
Submodule 'repos/html'
(https://github.com/tree-sitter/tree-sitter-html) registered for path
'repos/html'
Submodule 'repos/janet-simple'
(https://codeberg.org/sogaiu/tree-sitter-janet-simple) registered for
path 'repos/janet-simple'
Submodule 'repos/java'
(https://github.com/tree-sitter/tree-sitter-java) registered for path
'repos/java'
Submodule 'repos/javascript'
(https://github.com/tree-sitter/tree-sitter-javascript) registered for
path 'repos/javascript'
Submodule 'repos/jsdoc'
(https://github.com/tree-sitter/tree-sitter-jsdoc) registered for path
'repos/jsdoc'
Submodule 'repos/json'
(https://github.com/tree-sitter/tree-sitter-json) registered for path
'repos/json'
Submodule 'repos/julia'
(https://github.com/tree-sitter/tree-sitter-julia) registered for path
'repos/julia'
Submodule 'repos/ocaml'
(https://github.com/tree-sitter/tree-sitter-ocaml) registered for path
'repos/ocaml'
Submodule 'repos/php' (https://github.com/tree-sitter/tree-sitter-php)
registered for path 'repos/php'
Submodule 'repos/python'
(https://github.com/tree-sitter/tree-sitter-python) registered for
path 'repos/python'
Submodule 'repos/ruby'
(https://github.com/tree-sitter/tree-sitter-ruby) registered for path
'repos/ruby'
Submodule 'repos/rust'
(https://github.com/tree-sitter/tree-sitter-rust) registered for path
'repos/rust'
Submodule 'repos/scala'
(https://github.com/tree-sitter/tree-sitter-scala) registered for path
'repos/scala'
Submodule 'repos/swift'
(https://github.com/tree-sitter/tree-sitter-swift) registered for path
'repos/swift'
Submodule 'repos/typescript'
(https://github.com/tree-sitter/tree-sitter-typescript) registered for
path 'repos/typescript'
Skipping submodule 'repos/agda'
Skipping submodule 'repos/bash'
Skipping submodule 'repos/c'
Skipping submodule 'repos/c-sharp'
Skipping submodule 'repos/cpp'
Skipping submodule 'repos/css'
Skipping submodule 'repos/elm'
Skipping submodule 'repos/fluent'
Skipping submodule 'repos/go'
Skipping submodule 'repos/html'
Skipping submodule 'repos/janet-simple'
Skipping submodule 'repos/java'
Skipping submodule 'repos/javascript'
Skipping submodule 'repos/jsdoc'
Skipping submodule 'repos/json'
Skipping submodule 'repos/julia'
Skipping submodule 'repos/ocaml'
Skipping submodule 'repos/php'
Skipping submodule 'repos/python'
Skipping submodule 'repos/ruby'
Skipping submodule 'repos/rust'
Skipping submodule 'repos/scala'
Skipping submodule 'repos/swift'
Skipping submodule 'repos/typescript'

On Tue, Jun 15, 2021 at 5:51 PM brian m. carlson
[off-list ref] wrote:

quoted

On 2021-06-16 at 00:16:06, Rose Kunkel wrote:

quoted

# What did you do before the bug happened? (Steps to reproduce your issue)
1. Clone a git repository that sets `update = none` in .gitmodules:
$ git clone --recurse-submodules https://github.com/ubolonton/tree-sitter-langs

2. Perform a hard reset:
$ cd tree-sitter-langs
$ git reset --hard

# What did you expect to happen? (Expected behavior)
The reset should succeed and do nothing.

I think we're in agreement on this.  This should be a fresh clone and so
a hard reset should change nothing.

quoted

# What happened instead? (Actual behavior)
The reset command fails with

fatal: not a git repository: ../../.git/modules/repos/agda
fatal: could not reset submodule index

Hmmm, I can't reproduce this behavior.  What I see is this:

  $ git reset --hard
  HEAD is now at 5d362ce Release 0.10.0

I'm running git version 2.32.0.272.g935e593368 on Debian sid (with the
experimental packages).

Can you try the clone and run a "git status" command in the repository
to see if anything is modified after your clone?  Are the submodules
checked out when you perform the clone?  In my case, I see lines like
this:

  Skipping submodule 'repos/agda'

If you're seeing something different, then that might contribute to the
different behavior we're seeing.
--
brian m. carlson (he/him or they/them)
Toronto, Ontario, CA

If I set `submodule.recurse = false` in .git/config, I get the
behavior you're seeing.

On Tue, Jun 15, 2021 at 6:03 PM Rose Kunkel [off-list ref] wrote:

Potentially relevant: `git config --global --list` shows

status.showstash=true
status.submodulesummary=true
submodule.recurse=true
user.name=Rose Kunkel
user.email=rose@rosekunkel.me
pull.rebase=false
init.defaultbranch=main

On Tue, Jun 15, 2021 at 5:57 PM Rose Kunkel [off-list ref] wrote:

quoted

Running `git status` in the resulting repository gives

On branch master
Your branch is up to date with 'origin/master'.

nothing to commit, working tree clean

This is the output from the clone command:

Cloning into 'tree-sitter-langs'...
remote: Enumerating objects: 609, done.
remote: Counting objects: 100% (83/83), done.
remote: Compressing objects: 100% (52/52), done.
remote: Total 609 (delta 40), reused 58 (delta 24), pack-reused 526
Receiving objects: 100% (609/609), 117.17 KiB | 1.05 MiB/s, done.
Resolving deltas: 100% (322/322), done.
Submodule 'repos/agda'
(https://github.com/tree-sitter/tree-sitter-agda) registered for path
'repos/agda'
Submodule 'repos/bash'
(https://github.com/tree-sitter/tree-sitter-bash) registered for path
'repos/bash'
Submodule 'repos/c' (https://github.com/tree-sitter/tree-sitter-c)
registered for path 'repos/c'
Submodule 'repos/c-sharp'
(https://github.com/tree-sitter/tree-sitter-c-sharp) registered for
path 'repos/c-sharp'
Submodule 'repos/cpp' (https://github.com/tree-sitter/tree-sitter-cpp)
registered for path 'repos/cpp'
Submodule 'repos/css' (https://github.com/tree-sitter/tree-sitter-css)
registered for path 'repos/css'
Submodule 'repos/elm' (https://github.com/razzeee/tree-sitter-elm)
registered for path 'repos/elm'
Submodule 'repos/fluent'
(https://github.com/tree-sitter/tree-sitter-fluent) registered for
path 'repos/fluent'
Submodule 'repos/go' (https://github.com/tree-sitter/tree-sitter-go)
registered for path 'repos/go'
Submodule 'repos/html'
(https://github.com/tree-sitter/tree-sitter-html) registered for path
'repos/html'
Submodule 'repos/janet-simple'
(https://codeberg.org/sogaiu/tree-sitter-janet-simple) registered for
path 'repos/janet-simple'
Submodule 'repos/java'
(https://github.com/tree-sitter/tree-sitter-java) registered for path
'repos/java'
Submodule 'repos/javascript'
(https://github.com/tree-sitter/tree-sitter-javascript) registered for
path 'repos/javascript'
Submodule 'repos/jsdoc'
(https://github.com/tree-sitter/tree-sitter-jsdoc) registered for path
'repos/jsdoc'
Submodule 'repos/json'
(https://github.com/tree-sitter/tree-sitter-json) registered for path
'repos/json'
Submodule 'repos/julia'
(https://github.com/tree-sitter/tree-sitter-julia) registered for path
'repos/julia'
Submodule 'repos/ocaml'
(https://github.com/tree-sitter/tree-sitter-ocaml) registered for path
'repos/ocaml'
Submodule 'repos/php' (https://github.com/tree-sitter/tree-sitter-php)
registered for path 'repos/php'
Submodule 'repos/python'
(https://github.com/tree-sitter/tree-sitter-python) registered for
path 'repos/python'
Submodule 'repos/ruby'
(https://github.com/tree-sitter/tree-sitter-ruby) registered for path
'repos/ruby'
Submodule 'repos/rust'
(https://github.com/tree-sitter/tree-sitter-rust) registered for path
'repos/rust'
Submodule 'repos/scala'
(https://github.com/tree-sitter/tree-sitter-scala) registered for path
'repos/scala'
Submodule 'repos/swift'
(https://github.com/tree-sitter/tree-sitter-swift) registered for path
'repos/swift'
Submodule 'repos/typescript'
(https://github.com/tree-sitter/tree-sitter-typescript) registered for
path 'repos/typescript'
Skipping submodule 'repos/agda'
Skipping submodule 'repos/bash'
Skipping submodule 'repos/c'
Skipping submodule 'repos/c-sharp'
Skipping submodule 'repos/cpp'
Skipping submodule 'repos/css'
Skipping submodule 'repos/elm'
Skipping submodule 'repos/fluent'
Skipping submodule 'repos/go'
Skipping submodule 'repos/html'
Skipping submodule 'repos/janet-simple'
Skipping submodule 'repos/java'
Skipping submodule 'repos/javascript'
Skipping submodule 'repos/jsdoc'
Skipping submodule 'repos/json'
Skipping submodule 'repos/julia'
Skipping submodule 'repos/ocaml'
Skipping submodule 'repos/php'
Skipping submodule 'repos/python'
Skipping submodule 'repos/ruby'
Skipping submodule 'repos/rust'
Skipping submodule 'repos/scala'
Skipping submodule 'repos/swift'
Skipping submodule 'repos/typescript'

On Tue, Jun 15, 2021 at 5:51 PM brian m. carlson
[off-list ref] wrote:

quoted

On 2021-06-16 at 00:16:06, Rose Kunkel wrote:

quoted

# What did you do before the bug happened? (Steps to reproduce your issue)
1. Clone a git repository that sets `update = none` in .gitmodules:
$ git clone --recurse-submodules https://github.com/ubolonton/tree-sitter-langs

2. Perform a hard reset:
$ cd tree-sitter-langs
$ git reset --hard

# What did you expect to happen? (Expected behavior)
The reset should succeed and do nothing.

I think we're in agreement on this.  This should be a fresh clone and so
a hard reset should change nothing.

quoted

# What happened instead? (Actual behavior)
The reset command fails with

fatal: not a git repository: ../../.git/modules/repos/agda
fatal: could not reset submodule index

Hmmm, I can't reproduce this behavior.  What I see is this:

  $ git reset --hard
  HEAD is now at 5d362ce Release 0.10.0

I'm running git version 2.32.0.272.g935e593368 on Debian sid (with the
experimental packages).

Can you try the clone and run a "git status" command in the repository
to see if anything is modified after your clone?  Are the submodules
checked out when you perform the clone?  In my case, I see lines like
this:

  Skipping submodule 'repos/agda'

If you're seeing something different, then that might contribute to the
different behavior we're seeing.
--
brian m. carlson (he/him or they/them)
Toronto, Ontario, CA

On 2021-06-16 at 01:03:40, Rose Kunkel wrote:

Potentially relevant: `git config --global --list` shows

status.showstash=true
status.submodulesummary=true
submodule.recurse=true

Thanks for this additional information.  This line is the critical
piece.  Now I get this:

  $ git reset --hard
  fatal: not a git repository: ../../.git/modules/repos/agda
  fatal: could not reset submodule index

Predictably, "git -c submodules.recurse=true reset --hard" also results
in the same thing.

The --recurse-submodules option for git reset says this (emphasis mine):

  When the working tree is updated, using --recurse-submodules will also
  recursively reset the working tree of all *active* submodules
  according to the commit recorded in the superproject, also setting the
  submodules' HEAD to be detached at that commit.

On my system, .git/config has this:

  [submodule]
          active = .

So these submodules are active, but they probably should not be, since
we haven't checked anything out (or, for that matter, cloned any data)
and it wouldn't make sense to try to operate on them automatically with
submodules.recurse or --recurse-submodules.

My gut tells me that we should probably mark submodules with update=none
set on a clone as inactive.  Of course, this is a tricky area that I'm
not super familiar with, so opinions or thoughts are welcome.

If folks think this is a good way forward, I'll look into writing a
patch, probably tomorrow evening since it's starting to get late here.
-- 
brian m. carlson (he/him or they/them)
Toronto, Ontario, CA

That sounds reasonable to me.

I do think it's pretty unintuitive that `update = none` means that
submodules never get initialized, even with an explicit `git submodule
init` command. If this is intended behavior, it should be better
documented. If not, fixing that would also fix this bug.

On Tue, Jun 15, 2021 at 6:25 PM brian m. carlson
[off-list ref] wrote:

On 2021-06-16 at 01:03:40, Rose Kunkel wrote:

quoted

Potentially relevant: `git config --global --list` shows

status.showstash=true
status.submodulesummary=true
submodule.recurse=true

Thanks for this additional information.  This line is the critical
piece.  Now I get this:

  $ git reset --hard
  fatal: not a git repository: ../../.git/modules/repos/agda
  fatal: could not reset submodule index

Predictably, "git -c submodules.recurse=true reset --hard" also results
in the same thing.

The --recurse-submodules option for git reset says this (emphasis mine):

  When the working tree is updated, using --recurse-submodules will also
  recursively reset the working tree of all *active* submodules
  according to the commit recorded in the superproject, also setting the
  submodules' HEAD to be detached at that commit.

On my system, .git/config has this:

  [submodule]
          active = .

So these submodules are active, but they probably should not be, since
we haven't checked anything out (or, for that matter, cloned any data)
and it wouldn't make sense to try to operate on them automatically with
submodules.recurse or --recurse-submodules.

My gut tells me that we should probably mark submodules with update=none
set on a clone as inactive.  Of course, this is a tricky area that I'm
not super familiar with, so opinions or thoughts are welcome.

If folks think this is a good way forward, I'll look into writing a
patch, probably tomorrow evening since it's starting to get late here.
--
brian m. carlson (he/him or they/them)
Toronto, Ontario, CA

Actually, nevermind that last comment. I misunderstood what `git
submodule init` does. I thought it cloned missing submodules, but
apparently that's done by `git submodule update`, so the behavior does
make sense.

On Tue, Jun 15, 2021 at 6:39 PM Rose Kunkel [off-list ref] wrote:

That sounds reasonable to me.

I do think it's pretty unintuitive that `update = none` means that
submodules never get initialized, even with an explicit `git submodule
init` command. If this is intended behavior, it should be better
documented. If not, fixing that would also fix this bug.

On Tue, Jun 15, 2021 at 6:25 PM brian m. carlson
[off-list ref] wrote:

quoted

On 2021-06-16 at 01:03:40, Rose Kunkel wrote:

quoted

Potentially relevant: `git config --global --list` shows

status.showstash=true
status.submodulesummary=true
submodule.recurse=true

Thanks for this additional information.  This line is the critical
piece.  Now I get this:

  $ git reset --hard
  fatal: not a git repository: ../../.git/modules/repos/agda
  fatal: could not reset submodule index

Predictably, "git -c submodules.recurse=true reset --hard" also results
in the same thing.

The --recurse-submodules option for git reset says this (emphasis mine):

  When the working tree is updated, using --recurse-submodules will also
  recursively reset the working tree of all *active* submodules
  according to the commit recorded in the superproject, also setting the
  submodules' HEAD to be detached at that commit.

On my system, .git/config has this:

  [submodule]
          active = .

So these submodules are active, but they probably should not be, since
we haven't checked anything out (or, for that matter, cloned any data)
and it wouldn't make sense to try to operate on them automatically with
submodules.recurse or --recurse-submodules.

My gut tells me that we should probably mark submodules with update=none
set on a clone as inactive.  Of course, this is a tricky area that I'm
not super familiar with, so opinions or thoughts are welcome.

If folks think this is a good way forward, I'll look into writing a
patch, probably tomorrow evening since it's starting to get late here.
--
brian m. carlson (he/him or they/them)
Toronto, Ontario, CA