Re: Self-inflicted "abort" in a newbie attempt at read-only exploration of a cloned repository?
On 06/04/18 07:56 PM, Jeff King wrote: On Thu, Apr 05, 2018 at 04:18:23PM -0700, Bryan Turner wrote: The documentation for --work-tree says: --work-tree= Set the path to the working tree. It can be an absolute path or a path relative to the current working directory. This can also be controlled by setting the GIT_WORK_TREE environment variable and the core.worktree configuration variable (see core.worktree in git-config(1) for a more detailed discussion). So passing --work-tree tells Git where to store your _files_, but it's still using the same .git directory. If your goal is to have worktrees for various versions, that implies the git worktree [1] command might be more along the lines of what you're looking for. An invocation based on above might look like this: $ git -C linux-stable/ worktree add $PWD/tmp/ checkout linux-4.15.y Everything you've said here is completely accurate. But the original report does make me wonder if we've set up users for failure by overloading the term "worktree". Clearly it means two very different things in: git --work-tree=foo and git worktree add foo I'm not sure what to do about it at this point, though. :( The documentation for 'git worktree add' was adequate for my problem solving process. My difficulty occurred because I used an earlier GIT version. The terminology overloading might not be a big issue. Regards, - Thierry
Re: Self-inflicted "abort" in a newbie attempt at read-only exploration of a cloned repository?
On Thu, Apr 05, 2018 at 04:18:23PM -0700, Bryan Turner wrote: > The documentation for --work-tree says: > > --work-tree= > > Set the path to the working tree. It can be an absolute path or a path > relative to the current working directory. This can also be controlled > by setting the GIT_WORK_TREE environment variable and the > core.worktree configuration variable (see core.worktree in > git-config(1) for a more detailed discussion). > > So passing --work-tree tells Git where to store your _files_, but it's > still using the same .git directory. > > If your goal is to have worktrees for various versions, that implies > the git worktree [1] command might be more along the lines of what > you're looking for. An invocation based on above might look like this: > $ git -C linux-stable/ worktree add $PWD/tmp/ checkout linux-4.15.y Everything you've said here is completely accurate. But the original report does make me wonder if we've set up users for failure by overloading the term "worktree". Clearly it means two very different things in: git --work-tree=foo and git worktree add foo I'm not sure what to do about it at this point, though. :( -Peff
Re: Self-inflicted "abort" in a newbie attempt at read-only exploration of a cloned repository?
On 05/04/18 11:34 PM, Bryan Turner wrote: On Thu, Apr 5, 2018 at 4:18 PM, Bryan Turnerwrote: So passing --work-tree tells Git where to store your _files_, but it's still using the same .git directory. If your goal is to have worktrees for various versions, that implies the git worktree [1] command might be more along the lines of what you're looking for. An invocation based on above might look like this: $ git -C linux-stable/ worktree add $PWD/tmp/ checkout linux-4.15.y Apologies, I didn't mean to have the "checkout" in that. $ git -C linux-stable/ worktree add $PWD/tmp/ linux-4.15.y That should leave linux-4.16.y checked out in linux-stable, while creating a full work tree in $PWD/tmp that has 4.15.y checked out. Note that worktree is a newer git command. 2.17 has it, but old versions like 2.1 won't. [1] https://git-scm.com/docs/git-worktree Hope this helps! For sure, it helps! Thanks. - Thierry
Re: Self-inflicted "abort" in a newbie attempt at read-only exploration of a cloned repository?
On Thu, Apr 5, 2018 at 4:18 PM, Bryan Turnerwrote: > On Thu, Apr 5, 2018 at 12:42 PM, Thierry Moreau > wrote: >> Dear GIT enthusiasts! >> >> This ends up with a "git checkout" command aborting. A bit frustrating at >> the early stage of GIT learning curve. >> >> My first goal is to clone repositories locally in order to explore the >> various linux kernel versions, with the rich GIT metadata. >> >> Thus, I clone: >> >> $ git clone --branch linux-4.16.y >> https://git.kernel.org/pub/scm/linux/kernel/git/stable/linux-stable.git >> linux-stable >> $ git -C linux-stable/ branch >> * linux-4.16.y >> >> So far so good. Then, I want to extract an earlier kernel version into a tmp >> dir: >> >> $ mkdir tmp >> $ git -C linux-stable/ --work-tree $PWD/tmp/ checkout linux-4.15.y >> $ git -C linux-stable/ branch >> * linux-4.15.y >> linux-4.16.y > > The documentation for --work-tree says: > > --work-tree= > > Set the path to the working tree. It can be an absolute path or a path > relative to the current working directory. This can also be controlled > by setting the GIT_WORK_TREE environment variable and the > core.worktree configuration variable (see core.worktree in > git-config(1) for a more detailed discussion). > > So passing --work-tree tells Git where to store your _files_, but it's > still using the same .git directory. > > If your goal is to have worktrees for various versions, that implies > the git worktree [1] command might be more along the lines of what > you're looking for. An invocation based on above might look like this: > $ git -C linux-stable/ worktree add $PWD/tmp/ checkout linux-4.15.y Apologies, I didn't mean to have the "checkout" in that. $ git -C linux-stable/ worktree add $PWD/tmp/ linux-4.15.y > > That should leave linux-4.16.y checked out in linux-stable, while > creating a full work tree in $PWD/tmp that has 4.15.y checked out. > > Note that worktree is a newer git command. 2.17 has it, but old > versions like 2.1 won't. > > [1] https://git-scm.com/docs/git-worktree > > Hope this helps! > Bryan
Re: Self-inflicted "abort" in a newbie attempt at read-only exploration of a cloned repository?
On Thu, Apr 5, 2018 at 12:42 PM, Thierry Moreauwrote: > Dear GIT enthusiasts! > > This ends up with a "git checkout" command aborting. A bit frustrating at > the early stage of GIT learning curve. > > My first goal is to clone repositories locally in order to explore the > various linux kernel versions, with the rich GIT metadata. > > Thus, I clone: > > $ git clone --branch linux-4.16.y > https://git.kernel.org/pub/scm/linux/kernel/git/stable/linux-stable.git > linux-stable > $ git -C linux-stable/ branch > * linux-4.16.y > > So far so good. Then, I want to extract an earlier kernel version into a tmp > dir: > > $ mkdir tmp > $ git -C linux-stable/ --work-tree $PWD/tmp/ checkout linux-4.15.y > $ git -C linux-stable/ branch > * linux-4.15.y > linux-4.16.y The documentation for --work-tree says: --work-tree= Set the path to the working tree. It can be an absolute path or a path relative to the current working directory. This can also be controlled by setting the GIT_WORK_TREE environment variable and the core.worktree configuration variable (see core.worktree in git-config(1) for a more detailed discussion). So passing --work-tree tells Git where to store your _files_, but it's still using the same .git directory. If your goal is to have worktrees for various versions, that implies the git worktree [1] command might be more along the lines of what you're looking for. An invocation based on above might look like this: $ git -C linux-stable/ worktree add $PWD/tmp/ checkout linux-4.15.y That should leave linux-4.16.y checked out in linux-stable, while creating a full work tree in $PWD/tmp that has 4.15.y checked out. Note that worktree is a newer git command. 2.17 has it, but old versions like 2.1 won't. [1] https://git-scm.com/docs/git-worktree Hope this helps! Bryan
Self-inflicted "abort" in a newbie attempt at read-only exploration of a cloned repository?
Dear GIT enthusiasts! This ends up with a "git checkout" command aborting. A bit frustrating at the early stage of GIT learning curve. My first goal is to clone repositories locally in order to explore the various linux kernel versions, with the rich GIT metadata. Thus, I clone: $ git clone --branch linux-4.16.y https://git.kernel.org/pub/scm/linux/kernel/git/stable/linux-stable.git linux-stable $ git -C linux-stable/ branch * linux-4.16.y So far so good. Then, I want to extract an earlier kernel version into a tmp dir: $ mkdir tmp $ git -C linux-stable/ --work-tree $PWD/tmp/ checkout linux-4.15.y $ git -C linux-stable/ branch * linux-4.15.y linux-4.16.y I got my extracted 4.15 version but the source repository (index? ...?) has somehow changed. Let me try something silly: $ git -C linux-stable/ --work-tree $PWD/tmp/ checkout linux-4.14.y $ git -C linux-stable/ branch * linux-4.14.y linux-4.15.y linux-4.16.y I indeed switched my extracted version from 4.15 to 4.14, but I am puzzled that the local source repository (linux-stable) is modified. Then I try to bring it back closer to its original state, just to keep things tidy: $ git -C linux-stable/ checkout linux-4.16.y And this command aborts, both with Git versions 2.01 and 2.17. Here is the truncated command output: error: Your local changes to the following files would be overwritten by checkout: .gitignore .mailmap Documentation/00-INDEX Documentation/ABI/obsolete/sysfs-gpio Documentation/ABI/stable/sysfs-bus-vmbus Documentation/ABI/stable/sysfs-devices [... ...] Documentation/devicetree/bindings/arm/mediatek/mediatek,vencsys.txt Documentation/devicetree/bindings/arm/omap/crossbar.txt Documentation/devicetree/bindings/arm/omap/ctrl.txt Documentation/devicetree/bindings/arm/realtek.txt Documentation/devicetree/bindings/arm/rock error: The following untracked working tree files would be overwritten by checkout: Documentation/RCU/Design/Memory-Ordering/Tree-RCU-Diagram.html Documentation/RCU/Design/Memory-Ordering/Tree-RCU-Memory-Ordering.html Documentation/RCU/Design/Memory-Ordering/TreeRCU-callback-invocation.svg Documentation/RCU/Design/Memory-Ordering/TreeRCU-callback-registry.svg Documentation/RCU/Design/Memory-Ordering/TreeRCU-dyntick.svg [... ...] arch/riscv/include/asm/smp.h arch/riscv/include/asm/spinlock.h arch/riscv/include/asm/spinlock_types.h arch/riscv/include/asm/string.h arch/riscv/include/ Aborting Questions: = Is there a GIT tutorial that begins with my stated goal of an extract-only usage of cloned GIT repositories? (Maybe the root cause is my reluctance to learn the more involved GIT usages.) Does the above reproducible abort deserve attention? Any suggestion for my stated goal? Thanks in advance, - Thierry Moreau