git.vger.kernel.org archive mirror
 help / color / mirror / Atom feed
From: Lessley Dennington <lessleydennington@gmail.com>
To: Elijah Newren <newren@gmail.com>, Victoria Dye <vdye@github.com>
Cc: Elijah Newren via GitGitGadget <gitgitgadget@gmail.com>,
	Git Mailing List <git@vger.kernel.org>,
	Derrick Stolee <stolee@gmail.com>
Subject: Re: [PATCH 3/6] sparse-checkout: enable `set` to initialize sparse-checkout mode
Date: Tue, 7 Dec 2021 14:36:08 -0600	[thread overview]
Message-ID: <6e09ab19-7ffb-e58e-7b08-6e560b421c06@gmail.com> (raw)
In-Reply-To: <CABPp-BF-7DRmE+DPZOFuCYwSRdD8yk4pPsYpFN4=5fALEkqATg@mail.gmail.com>



On 12/7/21 9:04 AM, Elijah Newren wrote:
> On Sat, Dec 4, 2021 at 1:40 PM Victoria Dye <vdye@github.com> wrote:
>>
>> Elijah Newren via GitGitGadget wrote:
>>> From: Elijah Newren <newren@gmail.com>
>>>
>>> The previously suggested workflow:
>>>    git sparse-checkout init ...
>>>    git sparse-checkout set ...
>>>
>>> Suffered from three problems:
>>>    1) It would delete nearly all files in the first step, then
>>>       restore them in the second.  That was poor performance and
>>>       forced unnecessary rebuilds.
>>>    2) The two-step process resulted in two progress bars, which
>>>       was suboptimal from a UI point of view for wrappers that
>>>       invoked both of these commands but only exposed a single
>>>       command to their end users.
>>>    3) With cone mode, the first step would delete nearly all
>>>       ignored files everywhere, because everything was considered
>>>       to be outside of the specified sparsity paths.  (The user was
>>>       not allowed to specify any sparsity paths in the `init` step.)
>>>
>>> Avoid these problems by teaching `set` to understand the extra
>>> parameters that `init` takes and performing any necessary initialization
>>> if not already in a sparse checkout.
>>>
>>
>> I really like this change! It always seemed weird that `set` would
>> implicitly `init`, but without any of the options in `init`.
>>
>>> Signed-off-by: Elijah Newren <newren@gmail.com>
>>> ---
>>>   builtin/sparse-checkout.c | 52 ++++++++++++++++++++++++++++++++++++++-
>>>   1 file changed, 51 insertions(+), 1 deletion(-)
>>>
>>> diff --git a/builtin/sparse-checkout.c b/builtin/sparse-checkout.c
>>> index e252b82136e..cf6a6c6c3a7 100644
>>> --- a/builtin/sparse-checkout.c
>>> +++ b/builtin/sparse-checkout.c
>>> @@ -682,17 +682,26 @@ static int sparse_checkout_add(int argc, const char **argv, const char *prefix)
>>>   }
>>>
>>>   static char const * const builtin_sparse_checkout_set_usage[] = {
>>> -     N_("git sparse-checkout set (--stdin | <patterns>)"),
>>> +     N_("git sparse-checkout set [--cone] [--[no-]sparse-index] (--stdin | <patterns>)"),
>>
>> Since `cone` is an `OPT_BOOL`, shouldn't it appear in the usage string as
>> `--[no-]cone`? Without a `--no-cone` option, it's not clear how a user would
>> disable cone mode (since `set` preserves the existing cone mode setting if
>> `--cone` isn't given).
> 
> Yeah, fair point.  When copying from init I probably should have double checked.
> 
> Also, it makes me wonder if we should just make cone mode the default...
> 
That was actually my first thought when reviewing this change! It feels
like this would be a great moment to take the leap on that, but I can
also see it warranting a separate series.

>>>        NULL
>>>   };
>>>
>>>   static struct sparse_checkout_set_opts {
>>> +     int cone_mode;
>>> +     int sparse_index;
>>>        int use_stdin;
>>>   } set_opts;
>>>
>>>   static int sparse_checkout_set(int argc, const char **argv, const char *prefix)
>>>   {
>>> +     int mode, record_mode;
>>> +     const char *default_patterns[] = {"/*", "!/*/"};
>>> +
>>>        static struct option builtin_sparse_checkout_set_options[] = {
>>> +             OPT_BOOL(0, "cone", &set_opts.cone_mode,
>>> +                      N_("initialize the sparse-checkout in cone mode")),
>>> +             OPT_BOOL(0, "sparse-index", &set_opts.sparse_index,
>>> +                      N_("toggle the use of a sparse index")),
>>>                OPT_BOOL(0, "stdin", &set_opts.use_stdin,
>>>                         N_("read patterns from standard in")),
>>
>> I know this isn't part of this patch, but is `stdin` really a bool (i.e.
>> someone might specify `--no-stdin`)? If not, I think `OPT_SET_INT` might be
>> more appropriate.
> 
> Good point.  OPT_SET_INT() wouldn't fix it, though; you need to use
> OPT_BOOL_F or OPT_SET_INT_F and pass PARSE_OPT_NONEG as a flag.  I can
> include that.
> 
>>>                OPT_END(),
>>> @@ -700,11 +709,52 @@ static int sparse_checkout_set(int argc, const char **argv, const char *prefix)
>>>
>>>        repo_read_index(the_repository);
>>>
>>> +     set_opts.cone_mode = -1;
>>> +     set_opts.sparse_index = -1;
>>> +
>>>        argc = parse_options(argc, argv, prefix,
>>>                             builtin_sparse_checkout_set_options,
>>>                             builtin_sparse_checkout_set_usage,
>>>                             PARSE_OPT_KEEP_UNKNOWN);
>>>
>>> +     /* Determine if we need to record the mode; ensure sparse checkout on */
>>> +     record_mode = (set_opts.cone_mode != -1) || !core_apply_sparse_checkout;
>>> +     core_apply_sparse_checkout = 1;
>>> +
>>> +     /* If not specified, use previous definition of cone mode */
>>> +     if (set_opts.cone_mode == -1 && core_apply_sparse_checkout)
>>
>> I *think* this is supposed go before the `core_apply_sparse_checkout = 1;`
>> above it (if the intention is to only use the pre-existing cone mode setting
>> when already in a sparse checkout). If not, the `core_apply_sparse_checkout`
>> part of the condition is redundant and can be removed entirely.
> 
> Eek.  I had it there originally, then moved it later while doing
> various changes.  This definitely should be later; thanks for
> catching.
> 
>>> +             set_opts.cone_mode = core_sparse_checkout_cone;
>>> +
>>> +     /* Set cone/non-cone mode appropriately */
>>> +     if (set_opts.cone_mode == 1) {
>>> +             mode = MODE_CONE_PATTERNS;
>>> +             core_sparse_checkout_cone = 1;
>>> +     } else {
>>> +             mode = MODE_ALL_PATTERNS;
>>> +     }
>>> +     if (record_mode && set_config(mode))
>>> +             return 1;
>>> +
>>> +     /* Set sparse-index/non-sparse-index mode if specified */
>>> +     if (set_opts.sparse_index >= 0) {
>>> +             if (set_sparse_index_config(the_repository, set_opts.sparse_index) < 0)
>>> +                     die(_("failed to modify sparse-index config"));
>>> +
>>> +             /* force an index rewrite */
>>> +             repo_read_index(the_repository);
>>> +             the_repository->index->updated_workdir = 1;
>>> +     }
>>> +
>>> +     /*
>>> +      * Cone mode automatically specifies the toplevel directory.  For
>>> +      * non-cone mode, if nothing is specified, manually select just the
>>> +      * top-level directory (much as 'init' would do).
>>> +      */
>>> +     if (!core_sparse_checkout_cone && argc == 0) {
>>> +             argv = default_patterns;
>>> +             argc = 2;
>>> +     }
>>> +
>>>        return modify_pattern_list(argc, argv, set_opts.use_stdin, REPLACE);
>>>   }

  reply	other threads:[~2021-12-07 20:36 UTC|newest]

Thread overview: 63+ messages / expand[flat|nested]  mbox.gz  Atom feed  top
2021-12-04 20:00 [PATCH 0/6] sparse-checkout: make set subsume init Elijah Newren via GitGitGadget
2021-12-04 20:00 ` [PATCH 1/6] sparse-checkout: pass use_stdin as a parameter instead of as a global Elijah Newren via GitGitGadget
2021-12-04 20:00 ` [PATCH 2/6] sparse-checkout: break apart functions for sparse_checkout_(set|add) Elijah Newren via GitGitGadget
2021-12-04 20:00 ` [PATCH 3/6] sparse-checkout: enable `set` to initialize sparse-checkout mode Elijah Newren via GitGitGadget
2021-12-04 21:40   ` Victoria Dye
2021-12-07 17:04     ` Elijah Newren
2021-12-07 20:36       ` Lessley Dennington [this message]
2021-12-07 16:42   ` Derrick Stolee
2021-12-07 18:18     ` Elijah Newren
2021-12-04 20:00 ` [PATCH 4/6] git-sparse-checkout.txt: update to document that set handles init Elijah Newren via GitGitGadget
2021-12-04 21:48   ` Victoria Dye
2021-12-07 16:45     ` Derrick Stolee
2021-12-07 17:06     ` Elijah Newren
2021-12-07 20:42   ` Lessley Dennington
2021-12-07 21:13     ` Elijah Newren
2021-12-07 22:36       ` Lessley Dennington
2021-12-04 20:00 ` [PATCH 5/6] Documentation: clarify/correct a few sparsity related statements Elijah Newren via GitGitGadget
2021-12-04 20:00 ` [PATCH 6/6] clone: avoid using deprecated `sparse-checkout init` Elijah Newren via GitGitGadget
2021-12-07 16:48 ` [PATCH 0/6] sparse-checkout: make set subsume init Derrick Stolee
2021-12-07 20:20 ` [PATCH v2 00/10] " Elijah Newren via GitGitGadget
2021-12-07 20:20   ` [PATCH v2 01/10] sparse-checkout: pass use_stdin as a parameter instead of as a global Elijah Newren via GitGitGadget
2021-12-07 20:20   ` [PATCH v2 02/10] sparse-checkout: break apart functions for sparse_checkout_(set|add) Elijah Newren via GitGitGadget
2021-12-07 20:20   ` [PATCH v2 03/10] sparse-checkout: add sanity-checks on initial sparsity state Elijah Newren via GitGitGadget
2021-12-07 20:20   ` [PATCH v2 04/10] sparse-checkout: disallow --no-stdin as an argument to set Elijah Newren via GitGitGadget
2021-12-07 20:20   ` [PATCH v2 05/10] sparse-checkout: split out code for tweaking settings config Elijah Newren via GitGitGadget
2021-12-07 20:20   ` [PATCH v2 06/10] sparse-checkout: enable `set` to initialize sparse-checkout mode Elijah Newren via GitGitGadget
2021-12-07 20:20   ` [PATCH v2 07/10] sparse-checkout: enable reapply to take --[no-]{cone,sparse-index} Elijah Newren via GitGitGadget
2021-12-07 20:20   ` [PATCH v2 08/10] git-sparse-checkout.txt: update to document init/set/reapply changes Elijah Newren via GitGitGadget
2021-12-07 20:20   ` [PATCH v2 09/10] Documentation: clarify/correct a few sparsity related statements Elijah Newren via GitGitGadget
2021-12-07 20:20   ` [PATCH v2 10/10] clone: avoid using deprecated `sparse-checkout init` Elijah Newren via GitGitGadget
2021-12-07 21:05   ` [PATCH v2 00/10] sparse-checkout: make set subsume init Derrick Stolee
2021-12-08 14:16   ` Victoria Dye
2021-12-10  3:56   ` [PATCH v3 " Elijah Newren via GitGitGadget
2021-12-10  3:56     ` [PATCH v3 01/10] sparse-checkout: pass use_stdin as a parameter instead of as a global Elijah Newren via GitGitGadget
2021-12-10  3:56     ` [PATCH v3 02/10] sparse-checkout: break apart functions for sparse_checkout_(set|add) Elijah Newren via GitGitGadget
2021-12-10  3:56     ` [PATCH v3 03/10] sparse-checkout: add sanity-checks on initial sparsity state Elijah Newren via GitGitGadget
2021-12-13 18:11       ` Junio C Hamano
2021-12-14  2:25         ` Elijah Newren
2021-12-10  3:56     ` [PATCH v3 04/10] sparse-checkout: disallow --no-stdin as an argument to set Elijah Newren via GitGitGadget
2021-12-10  3:56     ` [PATCH v3 05/10] sparse-checkout: split out code for tweaking settings config Elijah Newren via GitGitGadget
2021-12-10  3:56     ` [PATCH v3 06/10] sparse-checkout: enable `set` to initialize sparse-checkout mode Elijah Newren via GitGitGadget
2021-12-10  3:56     ` [PATCH v3 07/10] sparse-checkout: enable reapply to take --[no-]{cone,sparse-index} Elijah Newren via GitGitGadget
2021-12-13 18:23       ` Junio C Hamano
2021-12-14  2:39         ` Elijah Newren
2021-12-10  3:56     ` [PATCH v3 08/10] git-sparse-checkout.txt: update to document init/set/reapply changes Elijah Newren via GitGitGadget
2021-12-10  3:56     ` [PATCH v3 09/10] Documentation: clarify/correct a few sparsity related statements Elijah Newren via GitGitGadget
2021-12-10  3:56     ` [PATCH v3 10/10] clone: avoid using deprecated `sparse-checkout init` Elijah Newren via GitGitGadget
2021-12-14  4:09     ` [PATCH v4 00/10] sparse-checkout: make set subsume init Elijah Newren via GitGitGadget
2021-12-14  4:09       ` [PATCH v4 01/10] sparse-checkout: pass use_stdin as a parameter instead of as a global Elijah Newren via GitGitGadget
2021-12-14  4:09       ` [PATCH v4 02/10] sparse-checkout: break apart functions for sparse_checkout_(set|add) Elijah Newren via GitGitGadget
2021-12-14  4:09       ` [PATCH v4 03/10] sparse-checkout: add sanity-checks on initial sparsity state Elijah Newren via GitGitGadget
2021-12-14  4:09       ` [PATCH v4 04/10] sparse-checkout: disallow --no-stdin as an argument to set Elijah Newren via GitGitGadget
2021-12-14  4:09       ` [PATCH v4 05/10] sparse-checkout: split out code for tweaking settings config Elijah Newren via GitGitGadget
2021-12-14  4:09       ` [PATCH v4 06/10] sparse-checkout: enable `set` to initialize sparse-checkout mode Elijah Newren via GitGitGadget
2021-12-14  4:09       ` [PATCH v4 07/10] sparse-checkout: enable reapply to take --[no-]{cone,sparse-index} Elijah Newren via GitGitGadget
2021-12-23  0:47         ` Jiang Xin
2021-12-23 17:09           ` Elijah Newren
2021-12-23 19:09             ` Junio C Hamano
2021-12-14  4:09       ` [PATCH v4 08/10] git-sparse-checkout.txt: update to document init/set/reapply changes Elijah Newren via GitGitGadget
2021-12-14  4:09       ` [PATCH v4 09/10] Documentation: clarify/correct a few sparsity related statements Elijah Newren via GitGitGadget
2021-12-14  7:38         ` Bagas Sanjaya
2021-12-14  7:48           ` Elijah Newren
2021-12-14  4:09       ` [PATCH v4 10/10] clone: avoid using deprecated `sparse-checkout init` Elijah Newren via GitGitGadget

Reply instructions:

You may reply publicly to this message via plain-text email
using any one of the following methods:

* Save the following mbox file, import it into your mail client,
  and reply-to-all from there: mbox

  Avoid top-posting and favor interleaved quoting:
  https://en.wikipedia.org/wiki/Posting_style#Interleaved_style

* Reply using the --to, --cc, and --in-reply-to
  switches of git-send-email(1):

  git send-email \
    --in-reply-to=6e09ab19-7ffb-e58e-7b08-6e560b421c06@gmail.com \
    --to=lessleydennington@gmail.com \
    --cc=git@vger.kernel.org \
    --cc=gitgitgadget@gmail.com \
    --cc=newren@gmail.com \
    --cc=stolee@gmail.com \
    --cc=vdye@github.com \
    /path/to/YOUR_REPLY

  https://kernel.org/pub/software/scm/git/docs/git-send-email.html

* If your mail client supports setting the In-Reply-To header
  via mailto: links, try the mailto: link
Be sure your reply has a Subject: header at the top and a blank line before the message body.
This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox;
as well as URLs for NNTP newsgroup(s).