Google Summer of Code 2018 with Git

Proposal

You can check out my proposal at this link.

Blogs posts:

• Week 1 (introduction)
• Week 2
• Week 3
• Week 4
• Week 5
• Week 6
• Week 7
• Week 8
• Week 9
• Week 10
• Week 10 (update)
• Week 11
• Week 12
• Week 13
• Week 14

Additional:

• About GSoC (halfway blogpost)

Patches

git stash

This is what I have been working on during this summer. It is a series consisting of 26 patches. At the moment, they are on the mailing list, waiting to get reviewed. A part of them were already reviewed (up to and including "stash: convert store to builtin").

• Github pull request

V7:

• Github tag
• Mailing list

Link to every patch (mailing list)

• 01/26: sha1-name.c: added 'get_oidf', which acts like 'get_oid'
• 02/26: stash: improve option parsing test coverage *
• 03/26: stash: update test cases conform to coding guidelines
• 04/26: stash: renamed test cases to be more descriptive
• 05/26: stash: convert apply to builtin *
• 06/26: stash: convert drop and clear *
• 07/26: stash: convert branch *
• 08/26: stash: convert pop *
• 09/26: stash: implement the "list" command in the builtin
• 10/26: stash: convert show to builtin
• 11/26: stash: change git stash show usage text and documentation
• 12/26: stash: refactor show_stash() to use the diff API
• 13/26: stash: update git stash show documentation
• 14/26: stash: convert store to builtin
• 15/26: stash: convert create
• 16/26: stash: replace spawning a "read-tree" process
• 17/26: stash: avoid spawning a "diff-index" process
• 18/26: stash: convert push to builtin
• 19/26: stash: make push to be quiet
• 20/26: stash: add tests for git stash push -q
• 21/26: stash: replace spawning git ls-files child process
• 22/26: stash: convert save to builtin
• 23/26: stash: convert stash--helper.c into stash.c
• 24/26: stash: optimize get_untracked_files() and check_changes()
• 25/26: stash: replace all write-tree child processes with API calls
• 26/26: stash: replace all "git apply"

_{* The project was originally started by Joel Teichroeb. During the bonding period, he agreed to give me custody of his patches (the ones marked with *). I independently modified the original patches and developed new ones.}

Status: waiting to get reviewed

V6:

GitHub tag

Link to every patch (mailing list):

First series of patches:

• 1/4: sha1-name.c: added 'get_oidf', which acts like 'get_oid'
• 2/4: stash: improve option parsing test coverage
• 3/4: stash: update test cases conform to coding guidelines
• 4/4: stash: renamed test cases to be more descriptive

Second series of patches:

• 1/4: stash: convert apply to builtin
• 2/4: stash: convert drop and clear
• 3/4: stash: convert branch
• 4/4: stash: convert pop

Third series of patches:

• 1/6: stash: implement the "list" command in the builtin
• 2/6: stash: convert show to builtin
• 3/6: stash: change git stash show usage text and documentation
• 4/6: stash: refactor show_stash() to use the diff API
• 5/6: stash: update git stash show documentation
• 6/6: stash: convert store to builtin

GET_OID_GENTLY

When I started working on git stash, I did not forsee that get_oid() might die in case there is an empty reflog. The purpose of this patch is to replace spawning git rev-parse --verify --quiet.

GitHub pull request

GitHub tag

Mailing list

Link to every patch (mailing list)

• 1/6: sha1-name: Add GET_OID_GENTLY flag
• 2/6: tree-walk: Add three new gentle helpers
• 3/6: refs.c: Teach read_ref_at() to accept GET_OID_GENTLY flag
• 4/6: sha1-name: Teach get_oid_basic() to be gentle
• 5/6: sha1-name: Teach get_oid_with_context[_1]()
• 6/6: sha1-name: Add gentle alternative for get_oid()

Status: not merged

TODO:

• One thing which was not included in the schedule, but might be a welcomed feature, is to introduce a new subcommand: git stash commit which would act like git stash pop and git commit.

• Add the possibility to see untracked files in a stash. There has been a discussion about this on the mailing list.

• Running git stash show -p and trying to split a hunk does not work properly in all the cases. There has been a discussion about this on the mailing list

Initial file:

 Text1.
 Text2.
 Text3.
 Text4.
 Text5.

Case 1:

 Text1.foo
 Text2.
 Text3.
 Text4.bar
 Text5.

Case 2:

 Text1.foo
 Text2.
 Text3.
 Text4.
 Text5.bar

Running git stash push -p, splitting the hunk, accepting one and refusing the other one:

• in the first case an error messaged is falsely (to some degree) displayed. The stash was created, but the changes were not removed from the index.

• in the second case everything works fine.

So, the problem occurs only when there are 2 or less lines between the edited lines.

Results

Running time on:

Linux (Ubuntu 17.10)

• git test suite (t3903-stash.sh, t3904-stash-patch.sh, t3905-stash-include-untracked.sh and t3906-stash-submodule.sh)

        t3903-stash.sh:
        ** SHELL: 12,69s user 9,95s system 109% cpu 20,730 total
        **     C:  2,67s user 2,84s system 105% cpu  5,206 total

        t3904-stash-patch.sh:
        ** SHELL: 1,43s user 0,94s system 106% cpu 2,242 total
        **     C: 1,01s user 0,58s system 104% cpu 1,530 total

        t3905-stash-include-untracked.sh
        ** SHELL: 2,22s user 1,73s system 110% cpu 3,569 total
        **     C: 0,59s user 0,57s system 106% cpu 1,085 total

        t3906-stash-submodule.sh
        ** SHELL: 2,89s user 2,99s system 106% cpu 5,527 total
        **     C: 2,21s user 2,61s system 105% cpu 4,568 total

        TOTAL:
        ** SHELL: 19.23s user 15.61s system
        **     C:  6.48s user  6.60s system

• a git repository with 4000 files: 1000 not changed, 1000 staged files, 1000 unstaged files, 1000 untracked. In this case I ran some of the most used commands:

   git stash push:
  
   ** SHELL: 0,12s user 0,21s system 101% cpu 0,329 total
   **     C: 0,06s user 0,13s system 105% cpu 0,185 total
  
   git stash push -u:
  
   ** SHELL: 0,18s user 0,27s system  108% cpu 0,401 total
   **     C: 0,09s user 0,19s system  103% cpu 0,267 total
  
   git stash pop:
  
   ** SHELL: 0,16s user 0,26s system 103% cpu 0,399 total
   **     C: 0,13s user 0,19s system 102% cpu 0,308 total
  

Windows 8.1:

• git test suite (t3903-stash.sh, t3904-stash-patch.sh, t3905-stash-include-untracked.sh and t3906-stash-submodule.sh)

        TOTAL:
        ** SHELL: 17m20.407s
        **    C:   5m 6.967s

        =================================

        t3903-stash.sh:
        ** SHELL:
        real    11m28.043s
        user    0m11.042s
        sys     0m24.278s

        ** C:
        real    1m56.922s
        user    0m11.725s
        sys     0m25.462s

        ================================

        t3904-stash-patch.sh:
        ** SHELL:
        real    0m35.436s
        user    0m1.117s
        sys     0m2.669s

        ** C:
        real    0m32.015s
        user    0m2.346s
        sys     0m5.151s

        ================================

        t3905-stash-include-untracked.sh
        ** SHELL:
        real    2m11.608s
        user    0m2.482s
        sys     0m4.737s

        ** C:
        real    0m24.749s
        user    0m2.171s
        sys     0m5.060s

        =================================

        t3906-stash-submodule.sh
        ** SHELL:
        real    3m5.032s
        user    0m7.563s
        sys     0m21.697s

        ** C:
        real    2m13.281s
        user    0m8.644s
        sys     0m20.799s