Support untracked cache with '--untracked-files=all' if configured

From: Tao Klerks <[email protected]>

Untracked cache was originally designed to only work with '-untracked-files=normal', but this is a concern for UI tooling that wants to see "all" on a frequent basis, and the conditions that prevented applicability to the "all" mode no longer seem to apply.

The disqualification of untracked cache is a particular problem on Windows with the defaulted fscache, where the introduction of fsmonitor can make the first and costly directory-iteration happen in untracked file detection, single-threaded, rather than in preload-index on multiple threads. Improving the performance of a "normal" 'git status' run with fsmonitor can make
'git status --untracked-files=all' perform much worse.

Here we align the supported directory flags for the stored untracked cache data with the git config. If a user specifies an '--untracked-files=' commandline parameter that does not align with their 'status.showuntrackedfiles' config value, then the untracked cache will be ignored - as it is for other unsupported situations like when a pathspec is specified.

If the previously stored flags no longer match the current configuration, but the currently-applicable flags do match the current configuration, then the previously stored untracked cache data is discarded.

For most users there will be no change in behavior. Users who need '--untracked-files=all' to perform well will have the option of setting "status.showuntrackedfiles" to "all".

Users who need '--untracked-files=all' to perform well for their tooling AND prefer to avoid the verbosity of "all" when running git status explicitly... are out of luck for now.

Users who set "status.showuntrackedfiles" to "all" and yet most frequently explicitly call
'git status --untracked-files=normal' (and use the untracked cache) are the only users who will be explicitly disadvantaged by this change. These should be vanishingly rare, if there are any at all.

---
Support untracked cache with '--untracked-files=all' if configured

This patch aims to make it possible for users of -uall, either by preference or by need (eg UI tooling), to benefit from the untracked cache when they set their 'status.showuntrackedfiles' config setting to 'all'. This is very important for large repos in Windows.

More detail on the change and context in the commit message, I assume repeating a verbose message here is discouraged.

These changes result from a couple of conversations, [email protected] and CABPp-BFiwzzUgiTj_zu+vF5x20L0=1cf25cHwk7KZQj2YkVzXw@mail.gmail.com>.

The test suite passes, but as a n00b I do have questions:

* Is there any additional validatiln that could/should be done to confirm that "-uall" untracked data can be cached safely? It looks that way to me, both from following the code, discussing briefly with Elijah Newren in the thread above, and manual testing, but I'm not sure what edge-cases to be looking out for
* Is it OK to check the repo configuration in the body of processing? It seems to be a rare pattern
* Should I be adding test cases? To the existing t7603 untracked cache tests, or elsewhere?

Published-As: https://github.com/gitgitgadget/git/releases/tag/pr-985%2FTaoK%2Ftaok-untracked-cache-with-uall-v1
Fetch-It-Via: git fetch https://github.com/gitgitgadget/git pr-985/TaoK/taok-untracked-cache-with-uall-v1
Pull-Request: https://github.com/gitgitgadget/git/pull/985

 dir.c | 86 ++++++++++++++++++++++++++++++++++++++++++++++-------------
 1 file changed, 68 insertions(+), 18 deletions(-)

index ebe5ec046e..49fc13c111 100644
@@ -2668,13 +2668,33 @@ static void set_untracked_ident(struct untracked_cache *uc)
 	strbuf_addch(&uc->ident, 0);
 }
 
-static void new_untracked_cache(struct index_state *istate)
+static unsigned configured_default_dir_flags(struct index_state *istate)
+{
+	/* This logic is coordinated with the setting of these flags in
+	 * wt-status.c#wt_status_collect_untracked(), and the evaluation
+	 * of the config setting in commit.c#git_status_config()
+	 */
+	char *status_untracked_files_config_value;
+	int config_outcome = repo_config_get_string(istate->repo,
+								"status.showuntrackedfiles",
+								&status_untracked_files_config_value);
+	if (!config_outcome && !strcmp(status_untracked_files_config_value, "all")) {
+		return 0;
+	} else {
+		/*
+		 * The default, if "all" is not set, is "normal" - leading us here.
+		 * If the value is "none" then it really doesn't matter.
+		 */
+		return DIR_SHOW_OTHER_DIRECTORIES | DIR_HIDE_EMPTY_DIRECTORIES;
+	}
+}
+
+static void new_untracked_cache(struct index_state *istate, unsigned flags)
 {
 	struct untracked_cache *uc = xcalloc(1, sizeof(*uc));
 	strbuf_init(&uc->ident, 100);
 	uc->exclude_per_dir = ".gitignore";
-	/* should be the same flags used by git-status */
-	uc->dir_flags = DIR_SHOW_OTHER_DIRECTORIES | DIR_HIDE_EMPTY_DIRECTORIES;
+	uc->dir_flags = flags;
 	set_untracked_ident(uc);
 	istate->untracked = uc;
 	istate->cache_changed |= UNTRACKED_CHANGED;

@@ -2683,11 +2703,13 @@ static void new_untracked_cache(struct index_state *istate)
 void add_untracked_cache(struct index_state *istate)
 {
 	if (!istate->untracked) {
-		new_untracked_cache(istate);
+		new_untracked_cache(istate,
+				configured_default_dir_flags(istate));
 	} else {
 		if (!ident_in_untracked(istate->untracked)) {
 			free_untracked_cache(istate->untracked);
-			new_untracked_cache(istate);
+			new_untracked_cache(istate,
+					configured_default_dir_flags(istate));
 		}
 	}
 }

@@ -2703,10 +2725,12 @@ void remove_untracked_cache(struct index_state *istate)
 
 static struct untracked_cache_dir *validate_untracked_cache(struct dir_struct *dir,
 						      int base_len,
-						      const struct pathspec *pathspec)
+						      const struct pathspec *pathspec,
+							  struct index_state *istate)
 {
 	struct untracked_cache_dir *root;
 	static int untracked_cache_disabled = -1;
+	unsigned configured_dir_flags;
 
 	if (!dir->untracked)
 		return NULL;

@@ -2734,17 +2758,9 @@ static struct untracked_cache_dir *validate_untracked_cache(struct dir_struct *d
 	if (base_len || (pathspec && pathspec->nr))
 		return NULL;
 
-	/* Different set of flags may produce different results */
-	if (dir->flags != dir->untracked->dir_flags ||
-	    /*
-	     * See treat_directory(), case index_nonexistent. Without
-	     * this flag, we may need to also cache .git file content
-	     * for the resolve_gitlink_ref() call, which we don't.
-	     */
-	    !(dir->flags & DIR_SHOW_OTHER_DIRECTORIES) ||
-	    /* We don't support collecting ignore files */
-	    (dir->flags & (DIR_SHOW_IGNORED | DIR_SHOW_IGNORED_TOO |
-			   DIR_COLLECT_IGNORED)))
+	/* We don't support collecting ignore files */
+	if (dir->flags & (DIR_SHOW_IGNORED | DIR_SHOW_IGNORED_TOO |
+			DIR_COLLECT_IGNORED))
 		return NULL;
 
 	/*

@@ -2767,6 +2783,40 @@ static struct untracked_cache_dir *validate_untracked_cache(struct dir_struct *d
 		return NULL;
 	}
 
+	/* We don't support using or preparing the untracked cache if
+	 * the current effective flags don't match the configured
+	 * flags.
+	 */
+	configured_dir_flags = configured_default_dir_flags(istate);
+	if (dir->flags != configured_dir_flags)
+		return NULL;
+
+	/* If the untracked structure we received does not have the same flags
+	 * as configured, but the configured flags do match the effective flags,
+	 * then we need to reset / create a new "untracked" structure to match
+	 * the new config.
+	 * Keeping the saved and used untracked cache in-line with the
+	 * configuration provides an opportunity for frequent users of
+	 * "git status -uall" to leverage the untracked cache by aligning their
+	 * configuration (setting "status.showuntrackedfiles" to "all" or
+	 * "normal" as appropriate), where previously this option was
+	 * incompatible with untracked cache and *consistently* caused
+	 * surprisingly bad performance (with fscache and fsmonitor enabled) on
+	 * Windows.
+	 *
+	 * IMPROVEMENT OPPORTUNITY: If we reworked the untracked cache storage
+	 * to not be as bound up with the desired output in a given run,
+	 * and instead iterated through and stored enough information to
+	 * correctly serve both "modes", then users could get peak performance
+	 * with or without '-uall' regardless of their
+	 * "status.showuntrackedfiles" config.
+	 */
+	if (dir->flags != dir->untracked->dir_flags) {
+		free_untracked_cache(istate->untracked);
+		new_untracked_cache(istate, configured_dir_flags);
+		dir->untracked = istate->untracked;
+	}
+
 	if (!dir->untracked->root)
 		FLEX_ALLOC_STR(dir->untracked->root, name, "");
 

@@ -2838,7 +2888,7 @@ int read_directory(struct dir_struct *dir, struct index_state *istate,
 		return dir->nr;
 	}
 
-	untracked = validate_untracked_cache(dir, len, pathspec);
+	untracked = validate_untracked_cache(dir, len, pathspec, istate);
 	if (!untracked)
 		/*
 		 * make sure untracked cache code path is disabled,

base-commit: 670b81a890388c60b7032a4f5b879f2ece8c4558

gitgitgadget

From: Tao Klerks <[email protected]>

Untracked cache was originally designed to only work with '-untracked-files=normal', but this is a concern for UI tooling that wants to see "all" on a frequent basis, and the conditions that prevented applicability to the "all" mode no longer seem to apply.

The disqualification of untracked cache is a particular problem on Windows with the defaulted fscache, where the introduction of fsmonitor can make the first and costly directory-iteration happen in untracked file detection, single-threaded, rather than in preload-index on multiple threads. Improving the performance of a "normal" 'git status' run with fsmonitor can make
'git status --untracked-files=all' perform much worse.

In this change, align the supported directory flags for the stored untracked cache data with the git config. If a user specifies an '--untracked-files=' commandline parameter that does not align with their 'status.showuntrackedfiles' config value, then the untracked cache will be ignored - as it is for other unsupported situations like when a pathspec is specified.

If the previously stored flags no longer match the current configuration, but the currently-applicable flags do match the current configuration, then the previously stored untracked cache data is discarded.

For most users there will be no change in behavior. Users who need '--untracked-files=all' to perform well will have the option of setting "status.showuntrackedfiles" to "all".

Users who need '--untracked-files=all' to perform well for their tooling AND prefer to avoid the verbosity of "all" when running git status explicitly without options... are out of luck for now (no change).

Users who set "status.showuntrackedfiles" to "all" and yet most frequently explicitly call 'git status --untracked-files=normal' (and use the untracked cache) are the only ones who would be disadvantaged by this change. It seems unlikely there are any such users.

---
Support untracked cache with '--untracked-files=all' if configured

Resending this patch from a few months ago (with better standards compliance) - it hasn't seen any comments yet, I would dearly love some eyes on this as the change can make a big difference to hundreds of windows users in my environment (and I'd really rather not start distributing customized git builds!)

This patch aims to make it possible for users of the -uall flag to git status, either by preference or by need (eg UI tooling), to benefit from the untracked cache when they set their 'status.showuntrackedfiles' config setting to 'all'. This is very important for large repos in Windows.

More detail on the change and context in the commit message, I assume repeating a verbose message here is discouraged.

These changes result from a couple of conversations, [email protected] and CABPp-BFiwzzUgiTj_zu+vF5x20L0=1cf25cHwk7KZQj2YkVzXw@mail.gmail.com>.

The test suite passes, but as a n00b I do have questions:

* Is there any additional validation that could/should be done to confirm that "-uall" untracked data can be cached safely?

** It looks safe from following the code ** It seems from discussing briefly with Elijah Newren in the thread above that thare are no obvious red flags ** Manual testing, explicitly and implicitly through months of use, yields no issues

* Is it OK to check the repo configuration in the body of processing? It seems to be a rare pattern
* Can anyone think of a way to test this change?

Published-As: https://github.com/gitgitgadget/git/releases/tag/pr-985%2FTaoK%2Ftaok-untracked-cache-with-uall-v2
Fetch-It-Via: git fetch https://github.com/gitgitgadget/git pr-985/TaoK/taok-untracked-cache-with-uall-v2
Pull-Request: https://github.com/gitgitgadget/git/pull/985

Range-diff vs v1:

1: 2797efad9a4 ! 1: e2f1ad26c78 Support untracked cache with '--untracked-files=all' if configured
@@ Metadata
Author: Tao Klerks <[email protected]>

## Commit message ##
- Support untracked cache with '--untracked-files=all' if configured
+ untracked-cache: support '--untracked-files=all' if configured

Untracked cache was originally designed to only work with
- '-untracked-files=normal', but this is a concern for UI tooling
- that wants to see "all" on a frequent basis, and the conditions
- that prevented applicability to the "all" mode no longer seem to
- apply.
+ '-untracked-files=normal', but this is a concern for UI tooling that
+ wants to see "all" on a frequent basis, and the conditions that
+ prevented applicability to the "all" mode no longer seem to apply.

The disqualification of untracked cache is a particular problem on Windows with the defaulted fscache, where the introduction of
@@ Commit message
"normal" 'git status' run with fsmonitor can make
'git status --untracked-files=all' perform much worse.

- Here we align the supported directory flags for the stored
+ In this change, align the supported directory flags for the stored
untracked cache data with the git config. If a user specifies an
- '--untracked-files=' commandline parameter that does not align
- with their 'status.showuntrackedfiles' config value, then the
- untracked cache will be ignored - as it is for other unsupported
- situations like when a pathspec is specified.
+ '--untracked-files=' commandline parameter that does not align with
+ their 'status.showuntrackedfiles' config value, then the untracked
+ cache will be ignored - as it is for other unsupported situations like
+ when a pathspec is specified.

If the previously stored flags no longer match the current
- configuration, but the currently-applicable flags do match the
- current configuration, then the previously stored untracked cache
- data is discarded.
+ configuration, but the currently-applicable flags do match the current
+ configuration, then the previously stored untracked cache data is
+ discarded.

- For most users there will be no change in behavior. Users who
- need '--untracked-files=all' to perform well will have the option
- of setting "status.showuntrackedfiles" to "all".
+ For most users there will be no change in behavior. Users who need
+ '--untracked-files=all' to perform well will have the option of
+ setting "status.showuntrackedfiles" to "all".

Users who need '--untracked-files=all' to perform well for their tooling AND prefer to avoid the verbosity of "all" when running
- git status explicitly... are out of luck for now.
+ git status explicitly without options... are out of luck for now (no
+ change).

- Users who set "status.showuntrackedfiles" to "all" and yet
- most frequently explicitly call
- 'git status --untracked-files=normal' (and use the untracked
- cache) are the only users who will be explicitly disadvantaged
- by this change. These should be vanishingly rare, if there are
- any at all.
+ Users who set "status.showuntrackedfiles" to "all" and yet most
+ frequently explicitly call 'git status --untracked-files=normal' (and
+ use the untracked cache) are the only ones who would be disadvantaged
+ by this change. It seems unlikely there are any such users.

 dir.c | 86 ++++++++++++++++++++++++++++++++++++++++++++++-------------
 1 file changed, 68 insertions(+), 18 deletions(-)

index d91295f2bcd..e35331d3f71 100644
@@ -2746,13 +2746,33 @@ static void set_untracked_ident(struct untracked_cache *uc)
 	strbuf_addch(&uc->ident, 0);
 }
 
-static void new_untracked_cache(struct index_state *istate)
+static unsigned configured_default_dir_flags(struct index_state *istate)
+{
+	/* This logic is coordinated with the setting of these flags in
+	 * wt-status.c#wt_status_collect_untracked(), and the evaluation
+	 * of the config setting in commit.c#git_status_config()
+	 */
+	char *status_untracked_files_config_value;
+	int config_outcome = repo_config_get_string(istate->repo,
+								"status.showuntrackedfiles",
+								&status_untracked_files_config_value);
+	if (!config_outcome && !strcmp(status_untracked_files_config_value, "all")) {
+		return 0;
+	} else {
+		/*
+		 * The default, if "all" is not set, is "normal" - leading us here.
+		 * If the value is "none" then it really doesn't matter.
+		 */
+		return DIR_SHOW_OTHER_DIRECTORIES | DIR_HIDE_EMPTY_DIRECTORIES;
+	}
+}
+
+static void new_untracked_cache(struct index_state *istate, unsigned flags)
 {
 	struct untracked_cache *uc = xcalloc(1, sizeof(*uc));
 	strbuf_init(&uc->ident, 100);
 	uc->exclude_per_dir = ".gitignore";
-	/* should be the same flags used by git-status */
-	uc->dir_flags = DIR_SHOW_OTHER_DIRECTORIES | DIR_HIDE_EMPTY_DIRECTORIES;
+	uc->dir_flags = flags;
 	set_untracked_ident(uc);
 	istate->untracked = uc;
 	istate->cache_changed |= UNTRACKED_CHANGED;

@@ -2761,11 +2781,13 @@ static void new_untracked_cache(struct index_state *istate)
 void add_untracked_cache(struct index_state *istate)
 {
 	if (!istate->untracked) {
-		new_untracked_cache(istate);
+		new_untracked_cache(istate,
+				configured_default_dir_flags(istate));
 	} else {
 		if (!ident_in_untracked(istate->untracked)) {
 			free_untracked_cache(istate->untracked);
-			new_untracked_cache(istate);
+			new_untracked_cache(istate,
+					configured_default_dir_flags(istate));
 		}
 	}
 }

@@ -2781,10 +2803,12 @@ void remove_untracked_cache(struct index_state *istate)
 
 static struct untracked_cache_dir *validate_untracked_cache(struct dir_struct *dir,
 						      int base_len,
-						      const struct pathspec *pathspec)
+						      const struct pathspec *pathspec,
+							  struct index_state *istate)
 {
 	struct untracked_cache_dir *root;
 	static int untracked_cache_disabled = -1;
+	unsigned configured_dir_flags;
 
 	if (!dir->untracked)
 		return NULL;

@@ -2812,17 +2836,9 @@ static struct untracked_cache_dir *validate_untracked_cache(struct dir_struct *d
 	if (base_len || (pathspec && pathspec->nr))
 		return NULL;
 
-	/* Different set of flags may produce different results */
-	if (dir->flags != dir->untracked->dir_flags ||
-	    /*
-	     * See treat_directory(), case index_nonexistent. Without
-	     * this flag, we may need to also cache .git file content
-	     * for the resolve_gitlink_ref() call, which we don't.
-	     */
-	    !(dir->flags & DIR_SHOW_OTHER_DIRECTORIES) ||
-	    /* We don't support collecting ignore files */
-	    (dir->flags & (DIR_SHOW_IGNORED | DIR_SHOW_IGNORED_TOO |
-			   DIR_COLLECT_IGNORED)))
+	/* We don't support collecting ignore files */
+	if (dir->flags & (DIR_SHOW_IGNORED | DIR_SHOW_IGNORED_TOO |
+			DIR_COLLECT_IGNORED))
 		return NULL;
 
 	/*

@@ -2845,6 +2861,40 @@ static struct untracked_cache_dir *validate_untracked_cache(struct dir_struct *d
 		return NULL;
 	}
 
+	/* We don't support using or preparing the untracked cache if
+	 * the current effective flags don't match the configured
+	 * flags.
+	 */
+	configured_dir_flags = configured_default_dir_flags(istate);
+	if (dir->flags != configured_dir_flags)
+		return NULL;
+
+	/* If the untracked structure we received does not have the same flags
+	 * as configured, but the configured flags do match the effective flags,
+	 * then we need to reset / create a new "untracked" structure to match
+	 * the new config.
+	 * Keeping the saved and used untracked cache in-line with the
+	 * configuration provides an opportunity for frequent users of
+	 * "git status -uall" to leverage the untracked cache by aligning their
+	 * configuration (setting "status.showuntrackedfiles" to "all" or
+	 * "normal" as appropriate), where previously this option was
+	 * incompatible with untracked cache and *consistently* caused
+	 * surprisingly bad performance (with fscache and fsmonitor enabled) on
+	 * Windows.
+	 *
+	 * IMPROVEMENT OPPORTUNITY: If we reworked the untracked cache storage
+	 * to not be as bound up with the desired output in a given run,
+	 * and instead iterated through and stored enough information to
+	 * correctly serve both "modes", then users could get peak performance
+	 * with or without '-uall' regardless of their
+	 * "status.showuntrackedfiles" config.
+	 */
+	if (dir->flags != dir->untracked->dir_flags) {
+		free_untracked_cache(istate->untracked);
+		new_untracked_cache(istate, configured_dir_flags);
+		dir->untracked = istate->untracked;
+	}
+
 	if (!dir->untracked->root)
 		FLEX_ALLOC_STR(dir->untracked->root, name, "");
 

@@ -2916,7 +2966,7 @@ int read_directory(struct dir_struct *dir, struct index_state *istate,
 		return dir->nr;
 	}
 
-	untracked = validate_untracked_cache(dir, len, pathspec);
+	untracked = validate_untracked_cache(dir, len, pathspec, istate);
 	if (!untracked)
 		/*
 		 * make sure untracked cache code path is disabled,

base-commit: dab1b7905d0b295f1acef9785bb2b9cbb0fdec84

gitgitgadget

So, in short, the root of the problem is that untracked cache can be used to serve only one mode (between normal and all) of operation, and the real solution to that problem must come in a different form, i.e. allowing a single unified untracked cache to be usable in both modes, perhaps by maintaining all the untracked ones in the cache, but filter out upon output when the 'normal' mode is asked for?

Given how widely used we are these days, I am afraid that the days we can safely say "users with such a strange use pattern would not exist at all" is long gone.

I wouldn't call this the root of the problem I was trying to solve with this patch, but rather the root of the remaining problem, yes.

The challenge that I can't get my head around for this longer-term approach or objective, is *when* the untracked-folder nested files should be enumerated.

Currently, untracked folders are only recursed into if -uall is specified or configured - and that's a significant characteristic:
It's perfectly plausible that some users sometimes have huge deep untracked folder hierarchies that take a long time to explore, but git never needs to because they never specify -uall.

If we decided to serve both modes with one single untracked cache structure, then we would either need to always fully recurse, or implement some sort of "just-in-time" filling in of the recursive bits when someone specifies -uall for the first time.

Either way, I'm pretty sure it's beyond me to do that right. Hence this very-pragmatic approach that makes it *possible* to get good/normal performance with -uall.

Thanks, sorry, I thought I'd corrected them all but clearly missed some.

Fixed, thx.

Makes sense, fixed, thx.

Because we *do* now support using untracked cache with -uall... As long as the config matches the runtime flags (new check later).

Thx.

I've tightened this up a little with an inline call to configured_default_dir_flags(), getting rid of the variable, let's see if that makes more sense / is cleaner.

Sending new version with these changes.

From: Tao Klerks <[email protected]>

Untracked cache was originally designed to only work with '-untracked-files=normal', but this causes performance issues for UI tooling that wants to see "all" on a frequent basis. On the other hand, the conditions that prevented applicability to the "all" mode no longer seem to apply.

The disqualification of untracked cache has a particularly significant impact on Windows with the defaulted fscache, where the introduction of fsmonitor can make the first and costly directory-iteration happen in untracked file detection, single-threaded, rather than in preload-index on multiple threads. Improving the performance of a "normal" 'git status' run with fsmonitor can make
'git status --untracked-files=all' perform much worse.

To partially address this, align the supported directory flags for the stored untracked cache data with the git config. If a user specifies an '--untracked-files=' commandline parameter that does not align with their 'status.showuntrackedfiles' config value, then the untracked cache will be ignored - as it is for other unsupported situations like when a pathspec is specified.

If the previously stored flags no longer match the current configuration, but the currently-applicable flags do match the current configuration, then discard the previously stored untracked cache data.

For most users there will be no change in behavior. Users who need '--untracked-files=all' to perform well will now have the option of setting "status.showuntrackedfiles" to "all" for better / more consistent performance.

Users who need '--untracked-files=all' to perform well for their tooling AND prefer to avoid the verbosity of "all" when running git status explicitly without options... are out of luck for now (no change).

Users who have the "status.showuntrackedfiles" config set to "all" and yet frequently explicitly call
'git status --untracked-files=normal' (and use the untracked cache) are the only ones who will be disadvantaged by this change. Their "--untracked-files=normal" calls will, after this change, no longer use the untracked cache.

---
Support untracked cache with '--untracked-files=all' if configured

Resending this patch from a few months ago (with better standards compliance) - it hasn't seen any comments yet, I would dearly love some eyes on this as the change can make a big difference to hundreds of windows users in my environment (and I'd really rather not start distributing customized git builds!)

This patch aims to make it possible for users of the -uall flag to git status, either by preference or by need (eg UI tooling), to benefit from the untracked cache when they set their 'status.showuntrackedfiles' config setting to 'all'. This is very important for large repos in Windows.

More detail on the change and context in the commit message, I assume repeating a verbose message here is discouraged.

These changes result from a couple of conversations, [email protected] and CABPp-BFiwzzUgiTj_zu+vF5x20L0=1cf25cHwk7KZQj2YkVzXw@mail.gmail.com>.

The test suite passes, but as a n00b I do have questions:

* Is there any additional validation that could/should be done to confirm that "-uall" untracked data can be cached safely?
* It looks safe from following the code
* It seems from discussing briefly with Elijah Newren in the thread above that thare are no obvious red flags
* Manual testing, explicitly and implicitly through months of use, yields no issues
* Is it OK to check the repo configuration in the body of processing? It seems to be a rare pattern
* Can anyone think of a way to test this change?

Published-As: https://github.com/gitgitgadget/git/releases/tag/pr-985%2FTaoK%2Ftaok-untracked-cache-with-uall-v3
Fetch-It-Via: git fetch https://github.com/gitgitgadget/git pr-985/TaoK/taok-untracked-cache-with-uall-v3
Pull-Request: https://github.com/gitgitgadget/git/pull/985

Range-diff vs v2:

1: e2f1ad26c78 ! 1: 49cf90bfab5 untracked-cache: support '--untracked-files=all' if configured
@@ Commit message
untracked-cache: support '--untracked-files=all' if configured

Untracked cache was originally designed to only work with
- '-untracked-files=normal', but this is a concern for UI tooling that
- wants to see "all" on a frequent basis, and the conditions that
- prevented applicability to the "all" mode no longer seem to apply.
+ '-untracked-files=normal', but this causes performance issues for UI
+ tooling that wants to see "all" on a frequent basis. On the other hand,
+ the conditions that prevented applicability to the "all" mode no
+ longer seem to apply.

- The disqualification of untracked cache is a particular problem on
- Windows with the defaulted fscache, where the introduction of
- fsmonitor can make the first and costly directory-iteration happen
+ The disqualification of untracked cache has a particularly significant
+ impact on Windows with the defaulted fscache, where the introduction
+ of fsmonitor can make the first and costly directory-iteration happen
in untracked file detection, single-threaded, rather than in preload-index on multiple threads. Improving the performance of a "normal" 'git status' run with fsmonitor can make 'git status --untracked-files=all' perform much worse.

- In this change, align the supported directory flags for the stored
- untracked cache data with the git config. If a user specifies an
- '--untracked-files=' commandline parameter that does not align with
+ To partially address this, align the supported directory flags for the
+ stored untracked cache data with the git config. If a user specifies
+ an '--untracked-files=' commandline parameter that does not align with
their 'status.showuntrackedfiles' config value, then the untracked cache will be ignored - as it is for other unsupported situations like when a pathspec is specified.

If the previously stored flags no longer match the current configuration, but the currently-applicable flags do match the current
- configuration, then the previously stored untracked cache data is
- discarded.
+ configuration, then discard the previously stored untracked cache
+ data.

For most users there will be no change in behavior. Users who need
- '--untracked-files=all' to perform well will have the option of
- setting "status.showuntrackedfiles" to "all".
+ '--untracked-files=all' to perform well will now have the option of
+ setting "status.showuntrackedfiles" to "all" for better / more
+ consistent performance.

Users who need '--untracked-files=all' to perform well for their tooling AND prefer to avoid the verbosity of "all" when running git status explicitly without options... are out of luck for now (no change).

- Users who set "status.showuntrackedfiles" to "all" and yet most
- frequently explicitly call 'git status --untracked-files=normal' (and
- use the untracked cache) are the only ones who would be disadvantaged
- by this change. It seems unlikely there are any such users.
+ Users who have the "status.showuntrackedfiles" config set to "all"
+ and yet frequently explicitly call
+ 'git status --untracked-files=normal' (and use the untracked cache)
+ are the only ones who will be disadvantaged by this change. Their
+ "--untracked-files=normal" calls will, after this change, no longer
+ use the untracked cache.

 dir.c | 86 ++++++++++++++++++++++++++++++++++++++++++++++-------------
 1 file changed, 68 insertions(+), 18 deletions(-)

index d91295f2bcd..e35331d3f71 100644
@@ -2746,13 +2746,33 @@ static void set_untracked_ident(struct untracked_cache *uc)
 	strbuf_addch(&uc->ident, 0);
 }
 
-static void new_untracked_cache(struct index_state *istate)
+static unsigned configured_default_dir_flags(struct index_state *istate)
+{
+	/* This logic is coordinated with the setting of these flags in
+	 * wt-status.c#wt_status_collect_untracked(), and the evaluation
+	 * of the config setting in commit.c#git_status_config()
+	 */
+	char *status_untracked_files_config_value;
+	int config_outcome = repo_config_get_string(istate->repo,
+								"status.showuntrackedfiles",
+								&status_untracked_files_config_value);
+	if (!config_outcome && !strcmp(status_untracked_files_config_value, "all")) {
+		return 0;
+	} else {
+		/*
+		 * The default, if "all" is not set, is "normal" - leading us here.
+		 * If the value is "none" then it really doesn't matter.
+		 */
+		return DIR_SHOW_OTHER_DIRECTORIES | DIR_HIDE_EMPTY_DIRECTORIES;
+	}
+}
+
+static void new_untracked_cache(struct index_state *istate, unsigned flags)
 {
 	struct untracked_cache *uc = xcalloc(1, sizeof(*uc));
 	strbuf_init(&uc->ident, 100);
 	uc->exclude_per_dir = ".gitignore";
-	/* should be the same flags used by git-status */
-	uc->dir_flags = DIR_SHOW_OTHER_DIRECTORIES | DIR_HIDE_EMPTY_DIRECTORIES;
+	uc->dir_flags = flags;
 	set_untracked_ident(uc);
 	istate->untracked = uc;
 	istate->cache_changed |= UNTRACKED_CHANGED;

@@ -2761,11 +2781,13 @@ static void new_untracked_cache(struct index_state *istate)
 void add_untracked_cache(struct index_state *istate)
 {
 	if (!istate->untracked) {
-		new_untracked_cache(istate);
+		new_untracked_cache(istate,
+				configured_default_dir_flags(istate));
 	} else {
 		if (!ident_in_untracked(istate->untracked)) {
 			free_untracked_cache(istate->untracked);
-			new_untracked_cache(istate);
+			new_untracked_cache(istate,
+					configured_default_dir_flags(istate));
 		}
 	}
 }

@@ -2781,10 +2803,12 @@ void remove_untracked_cache(struct index_state *istate)
 
 static struct untracked_cache_dir *validate_untracked_cache(struct dir_struct *dir,
 						      int base_len,
-						      const struct pathspec *pathspec)
+						      const struct pathspec *pathspec,
+							  struct index_state *istate)
 {
 	struct untracked_cache_dir *root;
 	static int untracked_cache_disabled = -1;
+	unsigned configured_dir_flags;
 
 	if (!dir->untracked)
 		return NULL;

@@ -2812,17 +2836,9 @@ static struct untracked_cache_dir *validate_untracked_cache(struct dir_struct *d
 	if (base_len || (pathspec && pathspec->nr))
 		return NULL;
 
-	/* Different set of flags may produce different results */
-	if (dir->flags != dir->untracked->dir_flags ||
-	    /*
-	     * See treat_directory(), case index_nonexistent. Without
-	     * this flag, we may need to also cache .git file content
-	     * for the resolve_gitlink_ref() call, which we don't.
-	     */
-	    !(dir->flags & DIR_SHOW_OTHER_DIRECTORIES) ||
-	    /* We don't support collecting ignore files */
-	    (dir->flags & (DIR_SHOW_IGNORED | DIR_SHOW_IGNORED_TOO |
-			   DIR_COLLECT_IGNORED)))
+	/* We don't support collecting ignore files */
+	if (dir->flags & (DIR_SHOW_IGNORED | DIR_SHOW_IGNORED_TOO |
+			DIR_COLLECT_IGNORED))
 		return NULL;
 
 	/*

@@ -2845,6 +2861,40 @@ static struct untracked_cache_dir *validate_untracked_cache(struct dir_struct *d
 		return NULL;
 	}
 
+	/* We don't support using or preparing the untracked cache if
+	 * the current effective flags don't match the configured
+	 * flags.
+	 */
+	configured_dir_flags = configured_default_dir_flags(istate);
+	if (dir->flags != configured_dir_flags)
+		return NULL;
+
+	/* If the untracked structure we received does not have the same flags
+	 * as configured, but the configured flags do match the effective flags,
+	 * then we need to reset / create a new "untracked" structure to match
+	 * the new config.
+	 * Keeping the saved and used untracked cache in-line with the
+	 * configuration provides an opportunity for frequent users of
+	 * "git status -uall" to leverage the untracked cache by aligning their
+	 * configuration (setting "status.showuntrackedfiles" to "all" or
+	 * "normal" as appropriate), where previously this option was
+	 * incompatible with untracked cache and *consistently* caused
+	 * surprisingly bad performance (with fscache and fsmonitor enabled) on
+	 * Windows.
+	 *
+	 * IMPROVEMENT OPPORTUNITY: If we reworked the untracked cache storage
+	 * to not be as bound up with the desired output in a given run,
+	 * and instead iterated through and stored enough information to
+	 * correctly serve both "modes", then users could get peak performance
+	 * with or without '-uall' regardless of their
+	 * "status.showuntrackedfiles" config.
+	 */
+	if (dir->flags != dir->untracked->dir_flags) {
+		free_untracked_cache(istate->untracked);
+		new_untracked_cache(istate, configured_dir_flags);
+		dir->untracked = istate->untracked;
+	}
+
 	if (!dir->untracked->root)
 		FLEX_ALLOC_STR(dir->untracked->root, name, "");
 

@@ -2916,7 +2966,7 @@ int read_directory(struct dir_struct *dir, struct index_state *istate,
 		return dir->nr;
 	}
 
-	untracked = validate_untracked_cache(dir, len, pathspec);
+	untracked = validate_untracked_cache(dir, len, pathspec, istate);
 	if (!untracked)
 		/*
 		 * make sure untracked cache code path is disabled,

base-commit: dab1b7905d0b295f1acef9785bb2b9cbb0fdec84

gitgitgadget

Bad submission, my apologies - commit message is right but code changes are missing. New one coming soon.

From: Tao Klerks <[email protected]>

Untracked cache was originally designed to only work with '-untracked-files=normal', but this causes performance issues for UI tooling that wants to see "all" on a frequent basis. On the other hand, the conditions that prevented applicability to the "all" mode no longer seem to apply.

The disqualification of untracked cache has a particularly significant impact on Windows with the defaulted fscache, where the introduction of fsmonitor can make the first and costly directory-iteration happen in untracked file detection, single-threaded, rather than in preload-index on multiple threads. Improving the performance of a "normal" 'git status' run with fsmonitor can make
'git status --untracked-files=all' perform much worse.

To partially address this, align the supported directory flags for the stored untracked cache data with the git config. If a user specifies an '--untracked-files=' commandline parameter that does not align with their 'status.showuntrackedfiles' config value, then the untracked cache will be ignored - as it is for other unsupported situations like when a pathspec is specified.

If the previously stored flags no longer match the current configuration, but the currently-applicable flags do match the current configuration, then discard the previously stored untracked cache data.

For most users there will be no change in behavior. Users who need '--untracked-files=all' to perform well will now have the option of setting "status.showuntrackedfiles" to "all" for better / more consistent performance.

Users who need '--untracked-files=all' to perform well for their tooling AND prefer to avoid the verbosity of "all" when running git status explicitly without options... are out of luck for now (no change).

Users who have the "status.showuntrackedfiles" config set to "all" and yet frequently explicitly call
'git status --untracked-files=normal' (and use the untracked cache) are the only ones who will be disadvantaged by this change. Their "--untracked-files=normal" calls will, after this change, no longer use the untracked cache.

---
Support untracked cache with '--untracked-files=all' if configured

Resending this patch from a few months ago (with better standards compliance) - it hasn't seen any comments yet, I would dearly love some eyes on this as the change can make a big difference to hundreds of windows users in my environment (and I'd really rather not start distributing customized git builds!)

This patch aims to make it possible for users of the -uall flag to git status, either by preference or by need (eg UI tooling), to benefit from the untracked cache when they set their 'status.showuntrackedfiles' config setting to 'all'. This is very important for large repos in Windows.

More detail on the change and context in the commit message, I assume repeating a verbose message here is discouraged.

These changes result from a couple of conversations, [email protected] and CABPp-BFiwzzUgiTj_zu+vF5x20L0=1cf25cHwk7KZQj2YkVzXw@mail.gmail.com>.

The test suite passes, but as a n00b I do have questions:

* Is there any additional validation that could/should be done to confirm that "-uall" untracked data can be cached safely?
* It looks safe from following the code
* It seems from discussing briefly with Elijah Newren in the thread above that thare are no obvious red flags
* Manual testing, explicitly and implicitly through months of use, yields no issues
* Is it OK to check the repo configuration in the body of processing? It seems to be a rare pattern
* Can anyone think of a way to test this change?

Published-As: https://github.com/gitgitgadget/git/releases/tag/pr-985%2FTaoK%2Ftaok-untracked-cache-with-uall-v4
Fetch-It-Via: git fetch https://github.com/gitgitgadget/git pr-985/TaoK/taok-untracked-cache-with-uall-v4
Pull-Request: https://github.com/gitgitgadget/git/pull/985

Range-diff vs v3:

1: 49cf90bfab5 ! 1: 5da418e5c60 untracked-cache: support '--untracked-files=all' if configured

     @@ dir.c: static void set_untracked_ident(struct untracked_cache *uc)
       }
       
      -static void new_untracked_cache(struct index_state *istate)
     -+static unsigned configured_default_dir_flags(struct index_state *istate)
     ++static unsigned configured_default_dir_flags(struct repository *repo)
      +{
     -+	/* This logic is coordinated with the setting of these flags in
     ++	/*
     ++	 * This logic is coordinated with the setting of these flags in
      +	 * wt-status.c#wt_status_collect_untracked(), and the evaluation
      +	 * of the config setting in commit.c#git_status_config()
      +	 */
      +	char *status_untracked_files_config_value;
     -+	int config_outcome = repo_config_get_string(istate->repo,
     -+								"status.showuntrackedfiles",
     -+								&status_untracked_files_config_value);
     ++	int config_outcome = repo_config_get_string(repo,
     ++						    "status.showuntrackedfiles",
     ++						    &status_untracked_files_config_value);
      +	if (!config_outcome && !strcmp(status_untracked_files_config_value, "all")) {
      +		return 0;
      +	} else {
     @@ dir.c: static void new_untracked_cache(struct index_state *istate)
       	if (!istate->untracked) {
      -		new_untracked_cache(istate);
      +		new_untracked_cache(istate,
     -+				configured_default_dir_flags(istate));
     ++				    configured_default_dir_flags(istate->repo));
       	} else {
       		if (!ident_in_untracked(istate->untracked)) {
       			free_untracked_cache(istate->untracked);
      -			new_untracked_cache(istate);
      +			new_untracked_cache(istate,
     -+					configured_default_dir_flags(istate));
     ++					    configured_default_dir_flags(istate->repo));
       		}
       	}
       }
      @@ dir.c: void remove_untracked_cache(struct index_state *istate)
     + }
       
       static struct untracked_cache_dir *validate_untracked_cache(struct dir_struct *dir,
     - 						      int base_len,
     +-						      int base_len,
      -						      const struct pathspec *pathspec)
     -+						      const struct pathspec *pathspec,
     -+							  struct index_state *istate)
     ++							    int base_len,
     ++							    const struct pathspec *pathspec,
     ++							    struct index_state *istate)
       {
       	struct untracked_cache_dir *root;
       	static int untracked_cache_disabled = -1;
     -+	unsigned configured_dir_flags;
     - 
     - 	if (!dir->untracked)
     - 		return NULL;
      @@ dir.c: static struct untracked_cache_dir *validate_untracked_cache(struct dir_struct *d
       	if (base_len || (pathspec && pathspec->nr))
       		return NULL;
     @@ dir.c: static struct untracked_cache_dir *validate_untracked_cache(struct dir_st
       		return NULL;
       	}
       
     -+	/* We don't support using or preparing the untracked cache if
     ++	/*
     ++	 * We don't support using or preparing the untracked cache if
      +	 * the current effective flags don't match the configured
      +	 * flags.
      +	 */
     -+	configured_dir_flags = configured_default_dir_flags(istate);
     -+	if (dir->flags != configured_dir_flags)
     ++	if (dir->flags != configured_default_dir_flags(istate->repo))
      +		return NULL;
      +
     -+	/* If the untracked structure we received does not have the same flags
     ++	/*
     ++	 * If the untracked structure we received does not have the same flags
      +	 * as configured, but the configured flags do match the effective flags,
      +	 * then we need to reset / create a new "untracked" structure to match
      +	 * the new config.
     @@ dir.c: static struct untracked_cache_dir *validate_untracked_cache(struct dir_st
      +	 */
      +	if (dir->flags != dir->untracked->dir_flags) {
      +		free_untracked_cache(istate->untracked);
     -+		new_untracked_cache(istate, configured_dir_flags);
     ++		new_untracked_cache(istate, dir->flags);
      +		dir->untracked = istate->untracked;
      +	}
      +

 dir.c | 89 ++++++++++++++++++++++++++++++++++++++++++++++-------------
 1 file changed, 70 insertions(+), 19 deletions(-)

index d91295f2bcd..57a7d42482f 100644
@@ -2746,13 +2746,34 @@ static void set_untracked_ident(struct untracked_cache *uc)
 	strbuf_addch(&uc->ident, 0);
 }
 
-static void new_untracked_cache(struct index_state *istate)
+static unsigned configured_default_dir_flags(struct repository *repo)
+{
+	/*
+	 * This logic is coordinated with the setting of these flags in
+	 * wt-status.c#wt_status_collect_untracked(), and the evaluation
+	 * of the config setting in commit.c#git_status_config()
+	 */
+	char *status_untracked_files_config_value;
+	int config_outcome = repo_config_get_string(repo,
+						    "status.showuntrackedfiles",
+						    &status_untracked_files_config_value);
+	if (!config_outcome && !strcmp(status_untracked_files_config_value, "all")) {
+		return 0;
+	} else {
+		/*
+		 * The default, if "all" is not set, is "normal" - leading us here.
+		 * If the value is "none" then it really doesn't matter.
+		 */
+		return DIR_SHOW_OTHER_DIRECTORIES | DIR_HIDE_EMPTY_DIRECTORIES;
+	}
+}
+
+static void new_untracked_cache(struct index_state *istate, unsigned flags)
 {
 	struct untracked_cache *uc = xcalloc(1, sizeof(*uc));
 	strbuf_init(&uc->ident, 100);
 	uc->exclude_per_dir = ".gitignore";
-	/* should be the same flags used by git-status */
-	uc->dir_flags = DIR_SHOW_OTHER_DIRECTORIES | DIR_HIDE_EMPTY_DIRECTORIES;
+	uc->dir_flags = flags;
 	set_untracked_ident(uc);
 	istate->untracked = uc;
 	istate->cache_changed |= UNTRACKED_CHANGED;

@@ -2761,11 +2782,13 @@ static void new_untracked_cache(struct index_state *istate)
 void add_untracked_cache(struct index_state *istate)
 {
 	if (!istate->untracked) {
-		new_untracked_cache(istate);
+		new_untracked_cache(istate,
+				    configured_default_dir_flags(istate->repo));
 	} else {
 		if (!ident_in_untracked(istate->untracked)) {
 			free_untracked_cache(istate->untracked);
-			new_untracked_cache(istate);
+			new_untracked_cache(istate,
+					    configured_default_dir_flags(istate->repo));
 		}
 	}
 }

@@ -2780,8 +2803,9 @@ void remove_untracked_cache(struct index_state *istate)
 }
 
 static struct untracked_cache_dir *validate_untracked_cache(struct dir_struct *dir,
-						      int base_len,
-						      const struct pathspec *pathspec)
+							    int base_len,
+							    const struct pathspec *pathspec,
+							    struct index_state *istate)
 {
 	struct untracked_cache_dir *root;
 	static int untracked_cache_disabled = -1;

@@ -2812,17 +2836,9 @@ static struct untracked_cache_dir *validate_untracked_cache(struct dir_struct *d
 	if (base_len || (pathspec && pathspec->nr))
 		return NULL;
 
-	/* Different set of flags may produce different results */
-	if (dir->flags != dir->untracked->dir_flags ||
-	    /*
-	     * See treat_directory(), case index_nonexistent. Without
-	     * this flag, we may need to also cache .git file content
-	     * for the resolve_gitlink_ref() call, which we don't.
-	     */
-	    !(dir->flags & DIR_SHOW_OTHER_DIRECTORIES) ||
-	    /* We don't support collecting ignore files */
-	    (dir->flags & (DIR_SHOW_IGNORED | DIR_SHOW_IGNORED_TOO |
-			   DIR_COLLECT_IGNORED)))
+	/* We don't support collecting ignore files */
+	if (dir->flags & (DIR_SHOW_IGNORED | DIR_SHOW_IGNORED_TOO |
+			DIR_COLLECT_IGNORED))
 		return NULL;
 
 	/*

@@ -2845,6 +2861,41 @@ static struct untracked_cache_dir *validate_untracked_cache(struct dir_struct *d
 		return NULL;
 	}
 
+	/*
+	 * We don't support using or preparing the untracked cache if
+	 * the current effective flags don't match the configured
+	 * flags.
+	 */
+	if (dir->flags != configured_default_dir_flags(istate->repo))
+		return NULL;
+
+	/*
+	 * If the untracked structure we received does not have the same flags
+	 * as configured, but the configured flags do match the effective flags,
+	 * then we need to reset / create a new "untracked" structure to match
+	 * the new config.
+	 * Keeping the saved and used untracked cache in-line with the
+	 * configuration provides an opportunity for frequent users of
+	 * "git status -uall" to leverage the untracked cache by aligning their
+	 * configuration (setting "status.showuntrackedfiles" to "all" or
+	 * "normal" as appropriate), where previously this option was
+	 * incompatible with untracked cache and *consistently* caused
+	 * surprisingly bad performance (with fscache and fsmonitor enabled) on
+	 * Windows.
+	 *
+	 * IMPROVEMENT OPPORTUNITY: If we reworked the untracked cache storage
+	 * to not be as bound up with the desired output in a given run,
+	 * and instead iterated through and stored enough information to
+	 * correctly serve both "modes", then users could get peak performance
+	 * with or without '-uall' regardless of their
+	 * "status.showuntrackedfiles" config.
+	 */
+	if (dir->flags != dir->untracked->dir_flags) {
+		free_untracked_cache(istate->untracked);
+		new_untracked_cache(istate, dir->flags);
+		dir->untracked = istate->untracked;
+	}
+
 	if (!dir->untracked->root)
 		FLEX_ALLOC_STR(dir->untracked->root, name, "");
 

@@ -2916,7 +2967,7 @@ int read_directory(struct dir_struct *dir, struct index_state *istate,
 		return dir->nr;
 	}
 
-	untracked = validate_untracked_cache(dir, len, pathspec);
+	untracked = validate_untracked_cache(dir, len, pathspec, istate);
 	if (!untracked)
 		/*
 		 * make sure untracked cache code path is disabled,

base-commit: dab1b7905d0b295f1acef9785bb2b9cbb0fdec84

gitgitgadget

The extra variable does not bother me at all. Leaving the room for the caller to screw up and pass an incorrect configred_dir_flags is what disturbs me.

If this caller needs to call configred_default_dir_flags(istate), then we cannot avoid it. And the extra variable needed to call the function only once, store its result, and use that result twice, is perfectly a good thing to have. We don't want to see inline or anything tricky.

Thanks.

I haven't given this any substantial review, sorry.

Yes, at a glance it looks safe to me.

Yes, it's not very common, but it's fine.

Well, if we set "flags" to 0 in your new helper the existing tests will fail, so that's something at least.

index 57a7d42482f..22a27d7780f 100644
@@ -2746,34 +2746,33 @@ static void set_untracked_ident(struct untracked_cache *uc)
 	strbuf_addch(&uc->ident, 0);
 }
 
-static unsigned configured_default_dir_flags(struct repository *repo)
+static unsigned new_untracked_cache_flags(struct index_state *istate)
 {
+	struct repository *repo = istate->repo;
+	char *val;
+
 	/*
 	 * This logic is coordinated with the setting of these flags in
 	 * wt-status.c#wt_status_collect_untracked(), and the evaluation
 	 * of the config setting in commit.c#git_status_config()
 	 */
-	char *status_untracked_files_config_value;
-	int config_outcome = repo_config_get_string(repo,
-						    "status.showuntrackedfiles",
-						    &status_untracked_files_config_value);
-	if (!config_outcome && !strcmp(status_untracked_files_config_value, "all")) {
+	if (!repo_config_get_string(repo, "status.showuntrackedfiles", &val) &&
+	    !strcmp(val, "all"))
 		return 0;
-	} else {
-		/*
-		 * The default, if "all" is not set, is "normal" - leading us here.
-		 * If the value is "none" then it really doesn't matter.
-		 */
-		return DIR_SHOW_OTHER_DIRECTORIES | DIR_HIDE_EMPTY_DIRECTORIES;
-	}
+
+	/*
+	 * The default, if "all" is not set, is "normal" - leading us here.
+	 * If the value is "none" then it really doesn't matter.
+	 */
+	return DIR_SHOW_OTHER_DIRECTORIES | DIR_HIDE_EMPTY_DIRECTORIES;
 }
 
-static void new_untracked_cache(struct index_state *istate, unsigned flags)
+static void new_untracked_cache(struct index_state *istate, int flags)
 {
 	struct untracked_cache *uc = xcalloc(1, sizeof(*uc));
 	strbuf_init(&uc->ident, 100);
 	uc->exclude_per_dir = ".gitignore";
-	uc->dir_flags = flags;
+	uc->dir_flags = flags >= 0 ? flags : new_untracked_cache_flags(istate);
 	set_untracked_ident(uc);
 	istate->untracked = uc;
 	istate->cache_changed |= UNTRACKED_CHANGED;

@@ -2782,13 +2781,11 @@ static void new_untracked_cache(struct index_state *istate, unsigned flags)
 void add_untracked_cache(struct index_state *istate)
 {
 	if (!istate->untracked) {
-		new_untracked_cache(istate,
-				    configured_default_dir_flags(istate->repo));
+		new_untracked_cache(istate, -1);
 	} else {
 		if (!ident_in_untracked(istate->untracked)) {
 			free_untracked_cache(istate->untracked);
-			new_untracked_cache(istate,
-					    configured_default_dir_flags(istate->repo));
+			new_untracked_cache(istate, -1);
 		}
 	}
 }

@@ -2866,7 +2863,7 @@ static struct untracked_cache_dir *validate_untracked_cache(struct dir_struct *d
 	 * the current effective flags don't match the configured
 	 * flags.
 	 */
-	if (dir->flags != configured_default_dir_flags(istate->repo))
+	if (dir->flags != new_untracked_cache_flags(istate))
 		return NULL;
 
 	/*

Thank you, this looks much cleaner!

One question I have about this patch itself is why you changed the arg to new_untracked_cache_flags / configured_default_dir_flags from "repo" back to "istate" / whether that was intentional; I had understood Junio's recommendation that the function get the minimum it needed (the repo in this case) as very sensible... especially now that we need to pass in that state less often.

Another question is whether I should add a comment in new_untracked_cache_flags explaining the new meaning of "-1" for the flags argument, or whether this is sufficiently standard in this project or in C generally that I should leave it implied.

A third question is whether you have an opinion on the better approach to getting these changes merged: does it make more sense to argue that this is an improvement overall, and that the case of configured "all" and argument-provided "normal", which will now have worse performance, can and should be ignored, *or* will it be easier to get this reviewed/merged if the new behavior is made conditional upon a new configuration key?

Last question is about protocol/workflow in this project: When I squash your proposal in, should I add an extra
"Signed-Off-By: Ævar Arnfjörð Bjarmason <[email protected]>" line, or just include the changes with mine?? (I've read SubmittingPatches but don't really understand the whole signing off business; I do understand that as your changes improve mine directly, they should indeed be "squashed" in, rather than added as an extra commit)

From: Tao Klerks <[email protected]>

Untracked cache was originally designed to only work with '--untracked-files=normal', and it gets ignored when '--untracked-files=all' is specified instead.

Add explicit tests for this behavior.

---
 t/t7063-status-untracked-cache.sh | 30 ++++++++++++++++++++++++++++++
 1 file changed, 30 insertions(+)

index ca90ee805e7..c44b70b96e4 100755
@@ -190,6 +190,36 @@ test_expect_success 'untracked cache after second status' '
 	test_cmp ../dump.expect ../actual
 '
 
+cat >../status_uall.expect <<EOF &&
+A  done/one
+A  one
+A  two
+?? dthree/three
+?? dtwo/two
+?? three
+EOF
+
+# Bypassing the untracked cache here is not desirable, but it expected
+# in the current implementation
+test_expect_success 'untracked cache is bypassed with -uall' '
+	: >../trace.output &&
+	GIT_TRACE2_PERF="$TRASH_DIRECTORY/trace.output" \
+	git status -uall --porcelain >../actual &&
+	iuc status -uall --porcelain >../status.iuc &&
+	test_cmp ../status_uall.expect ../status.iuc &&
+	test_cmp ../status_uall.expect ../actual &&
+	get_relevant_traces ../trace.output ../trace.relevant &&
+	cat >../trace.expect <<EOF &&
+ ....path:
+EOF
+	test_cmp ../trace.expect ../trace.relevant
+'
+
+test_expect_success 'untracked cache remains after bypass' '
+	test-tool dump-untracked-cache >../actual &&
+	test_cmp ../dump.expect ../actual
+'
+
 test_expect_success 'modify in root directory, one dir invalidation' '
 	: >four &&
 	test-tool chmtime =-240 four &&

gitgitgadget

Make it possible for users of the -uall flag to git status, either by preference or by need (eg UI tooling), to benefit from the untracked cache when they set their 'status.showuntrackedfiles' config setting to 'all'. This is especially useful for large repos in Windows, where without untracked cache "git status" times can easily be 5x higher, and GUI tooling typically does use -uall.

In this fifth version, split the change into two patches - one to introduce tests of the current untracked-cache-skipping behavior when -uall is specified, and then new tests checking the new behavior with 'status.showuntrackedfiles=all'.

My two remaining questions with respect to this patchset are:

1. Does it make sense to do this as a simple enhancement as proposed here, or would people be more comfortable with a new configuration option, given the potential for worse performance under specific (and. I believe, vanishingly rare) circumstances?
2. If it makes sense to envision a future where a single untracked cache structure can support both -uall and -unormal, where should this possible future be alluded to, if anywhere?

Tao Klerks (2):
untracked-cache: test untracked-cache-bypassing behavior with -uall
untracked-cache: support '--untracked-files=all' if configured

 dir.c                             |  85 +++++++++++++++++------
 t/t7063-status-untracked-cache.sh | 108 ++++++++++++++++++++++++++++++
 2 files changed, 174 insertions(+), 19 deletions(-)

base-commit: abf474a5dd901f28013c52155411a48fd4c09922
Published-As: https://github.com/gitgitgadget/git/releases/tag/pr-985%2FTaoK%2Ftaok-untracked-cache-with-uall-v5
Fetch-It-Via: git fetch https://github.com/gitgitgadget/git pr-985/TaoK/taok-untracked-cache-with-uall-v5
Pull-Request: https://github.com/gitgitgadget/git/pull/985

Range-diff vs v4:

 -:  ----------- > 1:  98a4d8f35c5 untracked-cache: test untracked-cache-bypassing behavior with -uall
 1:  5da418e5c60 ! 2:  f60d2c6e36c untracked-cache: support '--untracked-files=all' if configured
     @@ Commit message
          untracked-cache: support '--untracked-files=all' if configured
      
          Untracked cache was originally designed to only work with
     -    '-untracked-files=normal', but this causes performance issues for UI
     +    '--untracked-files=normal', but this causes performance issues for UI
          tooling that wants to see "all" on a frequent basis. On the other hand,
          the conditions that prevented applicability to the "all" mode no
          longer seem to apply.
     @@ dir.c: static void set_untracked_ident(struct untracked_cache *uc)
       }
       
      -static void new_untracked_cache(struct index_state *istate)
     -+static unsigned configured_default_dir_flags(struct repository *repo)
     ++static unsigned new_untracked_cache_flags(struct index_state *istate)
      +{
     ++	struct repository *repo = istate->repo;
     ++	char *val;
     ++
      +	/*
      +	 * This logic is coordinated with the setting of these flags in
      +	 * wt-status.c#wt_status_collect_untracked(), and the evaluation
      +	 * of the config setting in commit.c#git_status_config()
      +	 */
     -+	char *status_untracked_files_config_value;
     -+	int config_outcome = repo_config_get_string(repo,
     -+						    "status.showuntrackedfiles",
     -+						    &status_untracked_files_config_value);
     -+	if (!config_outcome && !strcmp(status_untracked_files_config_value, "all")) {
     ++	if (!repo_config_get_string(repo, "status.showuntrackedfiles", &val) &&
     ++	    !strcmp(val, "all"))
      +		return 0;
     -+	} else {
     -+		/*
     -+		 * The default, if "all" is not set, is "normal" - leading us here.
     -+		 * If the value is "none" then it really doesn't matter.
     -+		 */
     -+		return DIR_SHOW_OTHER_DIRECTORIES | DIR_HIDE_EMPTY_DIRECTORIES;
     -+	}
     ++
     ++	/*
     ++	 * The default, if "all" is not set, is "normal" - leading us here.
     ++	 * If the value is "none" then it really doesn't matter.
     ++	 */
     ++	return DIR_SHOW_OTHER_DIRECTORIES | DIR_HIDE_EMPTY_DIRECTORIES;
      +}
      +
     -+static void new_untracked_cache(struct index_state *istate, unsigned flags)
     ++static void new_untracked_cache(struct index_state *istate, int flags)
       {
       	struct untracked_cache *uc = xcalloc(1, sizeof(*uc));
       	strbuf_init(&uc->ident, 100);
       	uc->exclude_per_dir = ".gitignore";
      -	/* should be the same flags used by git-status */
      -	uc->dir_flags = DIR_SHOW_OTHER_DIRECTORIES | DIR_HIDE_EMPTY_DIRECTORIES;
     -+	uc->dir_flags = flags;
     ++	uc->dir_flags = flags >= 0 ? flags : new_untracked_cache_flags(istate);
       	set_untracked_ident(uc);
       	istate->untracked = uc;
       	istate->cache_changed |= UNTRACKED_CHANGED;
     @@ dir.c: static void new_untracked_cache(struct index_state *istate)
       {
       	if (!istate->untracked) {
      -		new_untracked_cache(istate);
     -+		new_untracked_cache(istate,
     -+				    configured_default_dir_flags(istate->repo));
     ++		new_untracked_cache(istate, -1);
       	} else {
       		if (!ident_in_untracked(istate->untracked)) {
       			free_untracked_cache(istate->untracked);
      -			new_untracked_cache(istate);
     -+			new_untracked_cache(istate,
     -+					    configured_default_dir_flags(istate->repo));
     ++			new_untracked_cache(istate, -1);
       		}
       	}
       }
     @@ dir.c: void remove_untracked_cache(struct index_state *istate)
       
       static struct untracked_cache_dir *validate_untracked_cache(struct dir_struct *dir,
      -						      int base_len,
     --						      const struct pathspec *pathspec)
     +-						      const struct pathspec *pathspec,
     +-						      struct index_state *istate)
      +							    int base_len,
      +							    const struct pathspec *pathspec,
      +							    struct index_state *istate)
     @@ dir.c: static struct untracked_cache_dir *validate_untracked_cache(struct dir_st
      +	 * the current effective flags don't match the configured
      +	 * flags.
      +	 */
     -+	if (dir->flags != configured_default_dir_flags(istate->repo))
     ++	if (dir->flags != new_untracked_cache_flags(istate))
      +		return NULL;
      +
      +	/*
     @@ dir.c: static struct untracked_cache_dir *validate_untracked_cache(struct dir_st
      +		dir->untracked = istate->untracked;
      +	}
      +
     - 	if (!dir->untracked->root)
     + 	if (!dir->untracked->root) {
     + 		/* Untracked cache existed but is not initialized; fix that */
       		FLEX_ALLOC_STR(dir->untracked->root, name, "");
     +
     + ## t/t7063-status-untracked-cache.sh ##
     +@@ t/t7063-status-untracked-cache.sh: test_expect_success 'untracked cache remains after bypass' '
     + 	test_cmp ../dump.expect ../actual
     + '
       
     -@@ dir.c: int read_directory(struct dir_struct *dir, struct index_state *istate,
     - 		return dir->nr;
     - 	}
     - 
     --	untracked = validate_untracked_cache(dir, len, pathspec);
     -+	untracked = validate_untracked_cache(dir, len, pathspec, istate);
     - 	if (!untracked)
     - 		/*
     - 		 * make sure untracked cache code path is disabled,
     ++test_expect_success 'if -uall is configured, untracked cache gets populated by default' '
     ++	test_config status.showuntrackedfiles all &&
     ++	: >../trace.output &&
     ++	GIT_TRACE2_PERF="$TRASH_DIRECTORY/trace.output" \
     ++	git status --porcelain >../actual &&
     ++	iuc status --porcelain >../status.iuc &&
     ++	test_cmp ../status_uall.expect ../status.iuc &&
     ++	test_cmp ../status_uall.expect ../actual &&
     ++	get_relevant_traces ../trace.output ../trace.relevant &&
     ++	cat >../trace.expect <<EOF &&
     ++ ....path:
     ++ ....node-creation:3
     ++ ....gitignore-invalidation:1
     ++ ....directory-invalidation:0
     ++ ....opendir:4
     ++EOF
     ++	test_cmp ../trace.expect ../trace.relevant
     ++'
     ++
     ++cat >../dump_uall.expect <<EOF &&
     ++info/exclude $EMPTY_BLOB
     ++core.excludesfile $ZERO_OID
     ++exclude_per_dir .gitignore
     ++flags 00000000
     ++/ $ZERO_OID recurse valid
     ++three
     ++/done/ $ZERO_OID recurse valid
     ++/dthree/ $ZERO_OID recurse valid
     ++three
     ++/dtwo/ $ZERO_OID recurse valid
     ++two
     ++EOF
     ++
     ++test_expect_success 'if -uall was configured, untracked cache is populated' '
     ++	test-tool dump-untracked-cache >../actual &&
     ++	test_cmp ../dump_uall.expect ../actual
     ++'
     ++
     ++test_expect_success 'if -uall is configured, untracked cache is used by default' '
     ++	test_config status.showuntrackedfiles all &&
     ++	: >../trace.output &&
     ++	GIT_TRACE2_PERF="$TRASH_DIRECTORY/trace.output" \
     ++	git status --porcelain >../actual &&
     ++	iuc status --porcelain >../status.iuc &&
     ++	test_cmp ../status_uall.expect ../status.iuc &&
     ++	test_cmp ../status_uall.expect ../actual &&
     ++	get_relevant_traces ../trace.output ../trace.relevant &&
     ++	cat >../trace.expect <<EOF &&
     ++ ....path:
     ++ ....node-creation:0
     ++ ....gitignore-invalidation:0
     ++ ....directory-invalidation:0
     ++ ....opendir:0
     ++EOF
     ++	test_cmp ../trace.expect ../trace.relevant
     ++'
     ++
     ++# Bypassing the untracked cache here is not desirable, but it expected
     ++# in the current implementation
     ++test_expect_success 'if -uall is configured, untracked cache is bypassed with -unormal' '
     ++	test_config status.showuntrackedfiles all &&
     ++	: >../trace.output &&
     ++	GIT_TRACE2_PERF="$TRASH_DIRECTORY/trace.output" \
     ++	git status -unormal --porcelain >../actual &&
     ++	iuc status -unormal --porcelain >../status.iuc &&
     ++	test_cmp ../status.expect ../status.iuc &&
     ++	test_cmp ../status.expect ../actual &&
     ++	get_relevant_traces ../trace.output ../trace.relevant &&
     ++	cat >../trace.expect <<EOF &&
     ++ ....path:
     ++EOF
     ++	test_cmp ../trace.expect ../trace.relevant
     ++'
     ++
     ++test_expect_success 'repopulate untracked cache for -unormal' '
     ++	git status --porcelain
     ++'
     ++
     + test_expect_success 'modify in root directory, one dir invalidation' '
     + 	: >four &&
     + 	test-tool chmtime =-240 four &&

gitgitgadget

From: Tao Klerks <[email protected]>

Untracked cache was originally designed to only work with '--untracked-files=normal', but this causes performance issues for UI tooling that wants to see "all" on a frequent basis. On the other hand, the conditions that prevented applicability to the "all" mode no longer seem to apply.

The disqualification of untracked cache has a particularly significant impact on Windows with the defaulted fscache, where the introduction of fsmonitor can make the first and costly directory-iteration happen in untracked file detection, single-threaded, rather than in preload-index on multiple threads. Improving the performance of a "normal" 'git status' run with fsmonitor can make
'git status --untracked-files=all' perform much worse.

To partially address this, align the supported directory flags for the stored untracked cache data with the git config. If a user specifies an '--untracked-files=' commandline parameter that does not align with their 'status.showuntrackedfiles' config value, then the untracked cache will be ignored - as it is for other unsupported situations like when a pathspec is specified.

If the previously stored flags no longer match the current configuration, but the currently-applicable flags do match the current configuration, then discard the previously stored untracked cache data.

For most users there will be no change in behavior. Users who need '--untracked-files=all' to perform well will now have the option of setting "status.showuntrackedfiles" to "all" for better / more consistent performance.

Users who need '--untracked-files=all' to perform well for their tooling AND prefer to avoid the verbosity of "all" when running git status explicitly without options... are out of luck for now (no change).

Users who have the "status.showuntrackedfiles" config set to "all" and yet frequently explicitly call
'git status --untracked-files=normal' (and use the untracked cache) are the only ones who will be disadvantaged by this change. Their "--untracked-files=normal" calls will, after this change, no longer use the untracked cache.

---
 dir.c                             | 85 ++++++++++++++++++++++++-------
 t/t7063-status-untracked-cache.sh | 78 ++++++++++++++++++++++++++++
 2 files changed, 144 insertions(+), 19 deletions(-)

index f2b0f242101..c5b513ddd2a 100644
@@ -2747,13 +2747,33 @@ static void set_untracked_ident(struct untracked_cache *uc)
 	strbuf_addch(&uc->ident, 0);
 }
 
-static void new_untracked_cache(struct index_state *istate)
+static unsigned new_untracked_cache_flags(struct index_state *istate)
+{
+	struct repository *repo = istate->repo;
+	char *val;
+
+	/*
+	 * This logic is coordinated with the setting of these flags in
+	 * wt-status.c#wt_status_collect_untracked(), and the evaluation
+	 * of the config setting in commit.c#git_status_config()
+	 */
+	if (!repo_config_get_string(repo, "status.showuntrackedfiles", &val) &&
+	    !strcmp(val, "all"))
+		return 0;
+
+	/*
+	 * The default, if "all" is not set, is "normal" - leading us here.
+	 * If the value is "none" then it really doesn't matter.
+	 */
+	return DIR_SHOW_OTHER_DIRECTORIES | DIR_HIDE_EMPTY_DIRECTORIES;
+}
+
+static void new_untracked_cache(struct index_state *istate, int flags)
 {
 	struct untracked_cache *uc = xcalloc(1, sizeof(*uc));
 	strbuf_init(&uc->ident, 100);
 	uc->exclude_per_dir = ".gitignore";
-	/* should be the same flags used by git-status */
-	uc->dir_flags = DIR_SHOW_OTHER_DIRECTORIES | DIR_HIDE_EMPTY_DIRECTORIES;
+	uc->dir_flags = flags >= 0 ? flags : new_untracked_cache_flags(istate);
 	set_untracked_ident(uc);
 	istate->untracked = uc;
 	istate->cache_changed |= UNTRACKED_CHANGED;

@@ -2762,11 +2782,11 @@ static void new_untracked_cache(struct index_state *istate)
 void add_untracked_cache(struct index_state *istate)
 {
 	if (!istate->untracked) {
-		new_untracked_cache(istate);
+		new_untracked_cache(istate, -1);
 	} else {
 		if (!ident_in_untracked(istate->untracked)) {
 			free_untracked_cache(istate->untracked);
-			new_untracked_cache(istate);
+			new_untracked_cache(istate, -1);
 		}
 	}
 }

@@ -2781,9 +2801,9 @@ void remove_untracked_cache(struct index_state *istate)
 }
 
 static struct untracked_cache_dir *validate_untracked_cache(struct dir_struct *dir,
-						      int base_len,
-						      const struct pathspec *pathspec,
-						      struct index_state *istate)
+							    int base_len,
+							    const struct pathspec *pathspec,
+							    struct index_state *istate)
 {
 	struct untracked_cache_dir *root;
 	static int untracked_cache_disabled = -1;

@@ -2814,17 +2834,9 @@ static struct untracked_cache_dir *validate_untracked_cache(struct dir_struct *d
 	if (base_len || (pathspec && pathspec->nr))
 		return NULL;
 
-	/* Different set of flags may produce different results */
-	if (dir->flags != dir->untracked->dir_flags ||
-	    /*
-	     * See treat_directory(), case index_nonexistent. Without
-	     * this flag, we may need to also cache .git file content
-	     * for the resolve_gitlink_ref() call, which we don't.
-	     */
-	    !(dir->flags & DIR_SHOW_OTHER_DIRECTORIES) ||
-	    /* We don't support collecting ignore files */
-	    (dir->flags & (DIR_SHOW_IGNORED | DIR_SHOW_IGNORED_TOO |
-			   DIR_COLLECT_IGNORED)))
+	/* We don't support collecting ignore files */
+	if (dir->flags & (DIR_SHOW_IGNORED | DIR_SHOW_IGNORED_TOO |
+			DIR_COLLECT_IGNORED))
 		return NULL;
 
 	/*

@@ -2847,6 +2859,41 @@ static struct untracked_cache_dir *validate_untracked_cache(struct dir_struct *d
 		return NULL;
 	}
 
+	/*
+	 * We don't support using or preparing the untracked cache if
+	 * the current effective flags don't match the configured
+	 * flags.
+	 */
+	if (dir->flags != new_untracked_cache_flags(istate))
+		return NULL;
+
+	/*
+	 * If the untracked structure we received does not have the same flags
+	 * as configured, but the configured flags do match the effective flags,
+	 * then we need to reset / create a new "untracked" structure to match
+	 * the new config.
+	 * Keeping the saved and used untracked cache in-line with the
+	 * configuration provides an opportunity for frequent users of
+	 * "git status -uall" to leverage the untracked cache by aligning their
+	 * configuration (setting "status.showuntrackedfiles" to "all" or
+	 * "normal" as appropriate), where previously this option was
+	 * incompatible with untracked cache and *consistently* caused
+	 * surprisingly bad performance (with fscache and fsmonitor enabled) on
+	 * Windows.
+	 *
+	 * IMPROVEMENT OPPORTUNITY: If we reworked the untracked cache storage
+	 * to not be as bound up with the desired output in a given run,
+	 * and instead iterated through and stored enough information to
+	 * correctly serve both "modes", then users could get peak performance
+	 * with or without '-uall' regardless of their
+	 * "status.showuntrackedfiles" config.
+	 */
+	if (dir->flags != dir->untracked->dir_flags) {
+		free_untracked_cache(istate->untracked);
+		new_untracked_cache(istate, dir->flags);
+		dir->untracked = istate->untracked;
+	}
+
 	if (!dir->untracked->root) {
 		/* Untracked cache existed but is not initialized; fix that */
 		FLEX_ALLOC_STR(dir->untracked->root, name, "");

gitgitgadget

There is a slight leap in logic flow before ", but this causes" that forces readers read the above twice and guess what is missing. I am guessing that

... was designed to only work with "--untracked-files=normal", and is bypassed when "--untracked-files=all" request, but this causes ...

is what you meant to say.

The claim in the last sentence "... no longer seem to apply" is thrown at the readers without much rationale. Either justify it more solidly or discard the claim?

The last sentence is a bit hard to parse and very hard to reason about. Do you mean to say "--untracked-files=all is slower when the untracked cache is in use, so the performance of 'git status' may be improved for '--untracked-files=normal' when the untracked cache is used, it hurts when 'git status --untracked-files=all' is run"?

Let me follow what actually happens in the patch aloud.

Sorry, I'm not completely sure whether my comment was misleading, or whether I'm misunderstanding your feedback.

The test added here does not test "desirable" behavior from an end-user functional perspective, but it does test behavior that is working "as-designed" as of many years ago.

The intent in adding this test is to ensure that if/when someone changes this behavior down the line, they be forced to understand the implications (eg: the current untracked cache structure does not store the same data for a -uall run as for a -unormal run, and so using the data from one in another would lead to output correctness errors).

Importantly, the behavior that this test is exercising *is not changed* by the 2/2 "improvement" patch; this is why I separated it out - it's a change that I feel we should make either way, and could even be its own independent patch. It's in this series as it adds relevant context and I am later adding similar tests for the new/additional behavior.

Given the (I believe) misunderstanding, a better comment would be something like:

# Bypassing the untracked cache here is not desirable from an

# end-user perspective, but is expected in the current implementation.
# The untracked cache data stored for a -unormal run cannot be
# correctly used in a -uall run - it would yield incorrect output.

Does that make more sense?

Is it "as-designed", or just "left buggy"?

I somehow had an impression that you meant the latter, and it would be our aspirational goal to eventually fix it. And for such case, it would be better to write the test body to show what the command should do, which would make the test fail with today's Git, and mark it as test_expect_failure (which is not an ideal mechanism to prepare for a future fix, but that is what we have now and should use, until a better alternative being worked on is finished).

But if it is "as-designed, some users may find it suboptimal or confusing or with any other negative adjectives, but it is too late to change now and more importantly it is unthinkable to change it because existing tools and users do depend on the current behaviour", then what you did is perfectly fine.

That, too.

Thanks.

the

Yep, will fix.

I can add references to previous conversation threads around the topic, and add the word "experimentally".

Something like that. The case where untracked cache bypassing (or disabling) has a huge impact is on windows, and involves fscache and fsmonitor:
* Untracked cache saves some work under qualifying conditions
* If qualifying conditions are not met (eg when -uall), we do the work anyway as though untracked cache were disabled.
* fscache optimises directory-iteration on windows, with an N-thread preload strategy. This is enabled by default, and improves windows "git status" performance transparently from "abominably slow" to just "slow".
* fsmonitor, when combined with untracked cache, eliminates recursive directory-iteration in favor of using a persistent in-memory model of what's changed on disk (to compare to the index and untracked cache)
* When fsmonitor is enabled and used, the fscache preload strategy is (obviously) disabled
* When fsmonitor is enabled and untracked cache is disabled or bypassed, the otherwise-ubiquitous "fscache preload" does not happen, and the untracked file search (directory iteration) takes... forever.

In other words: In windows, where fscache behavior is critical for large repos and defaulted-on, when fsmonitor is also enabled which generally further improves things but disqualifies fscache preload, the bypassing of untracked cache causes a huge performance impact. A "-uall" run can take a minute, where a "-unormal" run takes only a second, because the enumeration of untracked files without using the untracked cache data means a full directory iteration, not optimized by fscache. On the other hand, removing fsmonitor and untracked cache altogether, and leaving just the fscache optimization to do its job consistently, results in consistent middle-of-the-road performance in the order of 10 seconds in this example.

Enabling fsmonitor in this situation causes -uall to perform pathologically badly, because of untracked cache bypassing (interacting with the fscache-preload-bypassing of fsmonitor). However, the benefits of dropping status time from 10 seconds to 1 second make fsmonitor very attractive. On the other hand, the pathological performance with -uall is a problem if you use GUIs that *require* -uall. Therefore, fixing untracked cache to be compatible with -uall is critical to large repos on windows where the ideal experience would be fscache working, fsmonitor working, untracked cache working, and GUIs able to use -uall without the user waiting a minute or more and sometimes timing out.

This problem sounds like an edge-case when I describe the chain of interactions, but is actually very common among large-repo windows users. Windows users are typically GUI users. GUIs typically use -uall, as git's default "hide contents of untracked folders" behavior is obviously limiting.

This patch doesn't actually change anything I've described by default
- but it makes it *possible* to get good/consistent performance with -uall (with fscache, with fsmonitor, with untracked cache), *if* you configure status.showuntrackedfiles=all.

Now, coming back to the paragraph in question; maybe, instead of trying to summarize this windows-fscache-fsmonitor-untrackedcache interaction, I should be a little more vague. Would this work?

--
When 'git status' runs without using the untracked cache, on a large repo, on windows, with fsmonitor, it can run very slowly. This can make GUIs that need to use "-uall" (and therefore currently bypass untracked cache) unusable when fsmonitor is enabled, on such large repos.
--

[...]

Yes - the proposal, in this patch, is that the untracked cache usability/compatibility be aligned with the configuration. We are adding an extra meaning to the "status.showuntrackedfiles" configuration, such that it not only indicates what you want to have happen when you don't specify "-u" to git status, but it *also* indicates what kind of status output the untracked cache should be targeting/supporting.

Yep - although the word "guess" may be slightly misleading here. The proposed semantics are such that if the existing untracked cache flags are inconsistent with the config, we *discard* the no-longer-consistent-with-config untracked cache; so it's not so much a guess as a mandate.

As I noted above, we later consider that an untracked cache with flags that are inconsistent with the current config is an invalid untracked cache, and discard it; so setting the flags based on config is the right thing (the only useful thing) to do.

Will fix; I think there was a previous version in which these lines actually changed.

The logic in the current patch is:
* If the configuration doesn't match the request, ignore untracked cache completely (exit the UC logic)
* If the configuration doesn't match the current untracked cache data, discard the existing untracked cache data (and later recreate it if everything else works out)

I think your proposal is something like:
* If the current untracked cache data doesn't match the request ** If the current untracked cache data doesn't match the configuration, then discard the existing untracked cache (and later recreate it if everything else works out)
** Otherwise, ignore untracked cache completely (exit the UC logic)
* (otherwise, if current untracked cache data *does* match the request, assume it's good enough to keep using for now - this is the new behavior)

I find the latter slightly harder to understand and explain, but I'm reasonably certain it addresses the case you identified above without compromising correctness in any other cases - it can only have a positive impact on performance. I'll have a go at making that change today.

Heh, the answer is in-between: there certainly is an aspiration that someone figure out how to have the single untracked cache structure safely support both -uall and -unormal modes with the same data... but it's not clear that this is possible, and it's certainly not planned.

I don't feel comfortable saying "we should fix this", and that's not the intent of this test - its intent is to say "this is how it is (currently) designed to work. If you accidentally change it, be sure that you are changing the design, not just accidentally changing behavior and potentially/probably introducing a bug".

I'll try with this updated comment, using the term "current design" to make the distinction:

# Bypassing the untracked cache here is not desirable from an

# end-user perspective, but is expected in the current design.
# The untracked cache data stored for a -unormal run cannot be
# correctly used in a -uall run - it would yield incorrect output.

Make it possible for users of the -uall flag to git status, either by preference or by need (eg UI tooling), to benefit from the untracked cache when they set their 'status.showuntrackedfiles' config setting to 'all'. This is especially useful for large repos in Windows, where without untracked cache "git status" times can easily be 5x higher, and GUI tooling typically does use -uall.

In this sixth version, clarify the main commit message and some of the code comments, and adjust the logic slightly such that an existing untracked cache structure, when consistent with the requested flags in the current run, can be reused even though the current config would set/store/use the other set of flags on a new untracked cache structure.

I'm comfortable with this patch as-is, but am still interested in any thoughts as to whether it makes sense and is likely to be accepted to do this as a simple enhancement as proposed here, or whether people be more comfortable with a new configuration option, given the potential for worse performance under specific (and, I believe, vanishingly rare) circumstances.

Tao Klerks (2):
untracked-cache: test untracked-cache-bypassing behavior with -uall
untracked-cache: support '--untracked-files=all' if configured

 dir.c                             |  88 ++++++++++++++++++-----
 t/t7063-status-untracked-cache.sh | 113 ++++++++++++++++++++++++++++++
 2 files changed, 185 insertions(+), 16 deletions(-)

base-commit: 805e0a68082a217f0112db9ee86a022227a9c81b
Published-As: https://github.com/gitgitgadget/git/releases/tag/pr-985%2FTaoK%2Ftaok-untracked-cache-with-uall-v6
Fetch-It-Via: git fetch https://github.com/gitgitgadget/git pr-985/TaoK/taok-untracked-cache-with-uall-v6
Pull-Request: https://github.com/gitgitgadget/git/pull/985

Range-diff vs v5:

1: 98a4d8f35c5 ! 1: f27f018493a untracked-cache: test untracked-cache-bypassing behavior with -uall
@@ Commit message
'--untracked-files=normal', and it gets ignored when '--untracked-files=all' is specified instead.

- Add explicit tests for this behavior.
+ Add explicit tests for this known as-designed behavior.

@@ t/t7063-status-untracked-cache.sh: test_expect_success 'untracked cache after se
+?? three
+EOF
+
-+# Bypassing the untracked cache here is not desirable, but it expected
-+# in the current implementation
++# Bypassing the untracked cache here is not desirable from an
++# end-user perspective, but is expected in the current design.
++# The untracked cache data stored for a -unormal run cannot be
++# correctly used in a -uall run - it would yield incorrect output.
+test_expect_success 'untracked cache is bypassed with -uall' '
+ : >../trace.output &&
+ GIT_TRACE2_PERF="$TRASH_DIRECTORY/trace.output" \
2: f60d2c6e36c ! 2: 0095ec49c5f untracked-cache: support '--untracked-files=all' if configured
@@ Commit message
untracked-cache: support '--untracked-files=all' if configured

Untracked cache was originally designed to only work with
- '--untracked-files=normal', but this causes performance issues for UI
- tooling that wants to see "all" on a frequent basis. On the other hand,
- the conditions that prevented applicability to the "all" mode no
- longer seem to apply.
+ "--untracked-files=normal", and is bypassed when
+ "--untracked-files=all" is requested, but this causes performance
+ issues for UI tooling that wants to see "all" on a frequent basis.

- The disqualification of untracked cache has a particularly significant
- impact on Windows with the defaulted fscache, where the introduction
- of fsmonitor can make the first and costly directory-iteration happen
- in untracked file detection, single-threaded, rather than in
- preload-index on multiple threads. Improving the performance of a
- "normal" 'git status' run with fsmonitor can make
- 'git status --untracked-files=all' perform much worse.
+ On the other hand, the conditions that altogether prevented
+ applicability to the "all" mode no longer seem to apply, after
+ several major refactors in recent years; this possibility was
+ discussed in
+ [email protected] and
+ CABPp-BFiwzzUgiTj_zu+vF5x20L0=1cf25cHwk7KZQj2YkVzXw@mail.gmail.com,
+ and somewhat confirmed experimentally by several users using a
+ version of this patch to use untracked cache with -uall for about a
+ year.
+
+ When 'git status' runs without using the untracked cache, on a large
+ repo, on windows, with fsmonitor, it can run very slowly. This can
+ make GUIs that need to use "-uall" (and therefore currently bypass
+ untracked cache) unusable when fsmonitor is enabled, on such large
+ repos.

To partially address this, align the supported directory flags for the stored untracked cache data with the git config. If a user specifies
@@ dir.c: static void new_untracked_cache(struct index_state *istate)
}
}
}
-@@ dir.c: void remove_untracked_cache(struct index_state *istate)
- }
-
- static struct untracked_cache_dir *validate_untracked_cache(struct dir_struct *dir,
-- int base_len,
-- const struct pathspec *pathspec,
-- struct index_state *istate)
-+ int base_len,
-+ const struct pathspec *pathspec,
-+ struct index_state *istate)
- {
- struct untracked_cache_dir *root;
- static int untracked_cache_disabled = -1;
@@ dir.c: static struct untracked_cache_dir *validate_untracked_cache(struct dir_struct *d
if (base_len || (pathspec && pathspec->nr))
return NULL;
@@ dir.c: static struct untracked_cache_dir *validate_untracked_cache(struct dir_st
return NULL;
}

-+ /*
-+ * We don't support using or preparing the untracked cache if
-+ * the current effective flags don't match the configured
-+ * flags.
-+ */
-+ if (dir->flags != new_untracked_cache_flags(istate))
-+ return NULL;
-+
+ /*
+ * If the untracked structure we received does not have the same flags
-+ * as configured, but the configured flags do match the effective flags,
-+ * then we need to reset / create a new "untracked" structure to match
-+ * the new config.
-+ * Keeping the saved and used untracked cache in-line with the
-+ * configuration provides an opportunity for frequent users of
-+ * "git status -uall" to leverage the untracked cache by aligning their
-+ * configuration (setting "status.showuntrackedfiles" to "all" or
-+ * "normal" as appropriate), where previously this option was
-+ * incompatible with untracked cache and *consistently* caused
-+ * surprisingly bad performance (with fscache and fsmonitor enabled) on
-+ * Windows.
-+ *
-+ * IMPROVEMENT OPPORTUNITY: If we reworked the untracked cache storage
-+ * to not be as bound up with the desired output in a given run,
-+ * and instead iterated through and stored enough information to
-+ * correctly serve both "modes", then users could get peak performance
-+ * with or without '-uall' regardless of their
-+ * "status.showuntrackedfiles" config.
++ * as requested in this run, we're going to need to either discard the
++ * existing structure (and potentially later recreate), or bypass the
++ * untracked cache mechanism for this run.
+ */
+ if (dir->flags != dir->untracked->dir_flags) {
-+ free_untracked_cache(istate->untracked);
-+ new_untracked_cache(istate, dir->flags);
-+ dir->untracked = istate->untracked;
++ /*
++ * If the untracked structure we received does not have the same flags
++ * as configured, then we need to reset / create a new "untracked"
++ * structure to match the new config.
++ *
++ * Keeping the saved and used untracked cache consistent with the
++ * configuration provides an opportunity for frequent users of
++ * "git status -uall" to leverage the untracked cache by aligning their
++ * configuration - setting "status.showuntrackedfiles" to "all" or
++ * "normal" as appropriate.
++ *
++ * Previously using -uall (or setting "status.showuntrackedfiles" to
++ * "all") was incompatible with untracked cache and *consistently*
++ * caused surprisingly bad performance (with fscache and fsmonitor
++ * enabled) on Windows.
++ *
++ * IMPROVEMENT OPPORTUNITY: If we reworked the untracked cache storage
++ * to not be as bound up with the desired output in a given run,
++ * and instead iterated through and stored enough information to
++ * correctly serve both "modes", then users could get peak performance
++ * with or without '-uall' regardless of their
++ * "status.showuntrackedfiles" config.
++ */
++ if (dir->untracked->dir_flags != new_untracked_cache_flags(istate)) {
++ free_untracked_cache(istate->untracked);
++ new_untracked_cache(istate, dir->flags);
++ dir->untracked = istate->untracked;
++ }
++ else {
++ /*
++ * Current untracked cache data is consistent with config, but not
++ * usable in this request/run; just bypass untracked cache.
++ */
++ return NULL;
++ }
+ }
+
if (!dir->untracked->root) {
@@ t/t7063-status-untracked-cache.sh: test_expect_success 'untracked cache remains
+ test_cmp ../trace.expect ../trace.relevant
+'
+
-+# Bypassing the untracked cache here is not desirable, but it expected -+# in the current implementation
++# Bypassing the untracked cache here is not desirable from an
++# end-user perspective, but is expected in the current design.
++# The untracked cache data stored for a -all run cannot be
++# correctly used in a -unormal run - it would yield incorrect
++# output.
+test_expect_success 'if -uall is configured, untracked cache is bypassed with -unormal' '
+ test_config status.showuntrackedfiles all &&
+ : >../trace.output &&

gitgitgadget

From: Tao Klerks <[email protected]>

Untracked cache was originally designed to only work with '--untracked-files=normal', and it gets ignored when '--untracked-files=all' is specified instead.

Add explicit tests for this known as-designed behavior.

---
 t/t7063-status-untracked-cache.sh | 32 +++++++++++++++++++++++++++++++
 1 file changed, 32 insertions(+)

index ca90ee805e7..b89be8dc6d6 100755
@@ -190,6 +190,38 @@ test_expect_success 'untracked cache after second status' '
 	test_cmp ../dump.expect ../actual
 '
 
+cat >../status_uall.expect <<EOF &&
+A  done/one
+A  one
+A  two
+?? dthree/three
+?? dtwo/two
+?? three
+EOF
+
+# Bypassing the untracked cache here is not desirable from an
+# end-user perspective, but is expected in the current design.
+# The untracked cache data stored for a -unormal run cannot be
+# correctly used in a -uall run - it would yield incorrect output.
+test_expect_success 'untracked cache is bypassed with -uall' '
+	: >../trace.output &&
+	GIT_TRACE2_PERF="$TRASH_DIRECTORY/trace.output" \
+	git status -uall --porcelain >../actual &&
+	iuc status -uall --porcelain >../status.iuc &&
+	test_cmp ../status_uall.expect ../status.iuc &&
+	test_cmp ../status_uall.expect ../actual &&
+	get_relevant_traces ../trace.output ../trace.relevant &&
+	cat >../trace.expect <<EOF &&
+ ....path:
+EOF
+	test_cmp ../trace.expect ../trace.relevant
+'
+
+test_expect_success 'untracked cache remains after bypass' '
+	test-tool dump-untracked-cache >../actual &&
+	test_cmp ../dump.expect ../actual
+'
+
 test_expect_success 'modify in root directory, one dir invalidation' '
 	: >four &&
 	test-tool chmtime =-240 four &&

gitgitgadget

From: Tao Klerks <[email protected]>

Untracked cache was originally designed to only work with "--untracked-files=normal", and is bypassed when "--untracked-files=all" is requested, but this causes performance issues for UI tooling that wants to see "all" on a frequent basis.

On the other hand, the conditions that altogether prevented applicability to the "all" mode no longer seem to apply, after several major refactors in recent years; this possibility was discussed in [email protected] and CABPp-BFiwzzUgiTj_zu+vF5x20L0=1cf25cHwk7KZQj2YkVzXw@mail.gmail.com, and somewhat confirmed experimentally by several users using a version of this patch to use untracked cache with -uall for about a year.

When 'git status' runs without using the untracked cache, on a large repo, on windows, with fsmonitor, it can run very slowly. This can make GUIs that need to use "-uall" (and therefore currently bypass untracked cache) unusable when fsmonitor is enabled, on such large repos.

To partially address this, align the supported directory flags for the stored untracked cache data with the git config. If a user specifies an '--untracked-files=' commandline parameter that does not align with their 'status.showuntrackedfiles' config value, then the untracked cache will be ignored - as it is for other unsupported situations like when a pathspec is specified.

If the previously stored flags no longer match the current configuration, but the currently-applicable flags do match the current configuration, then discard the previously stored untracked cache data.

For most users there will be no change in behavior. Users who need '--untracked-files=all' to perform well will now have the option of setting "status.showuntrackedfiles" to "all" for better / more consistent performance.

Users who need '--untracked-files=all' to perform well for their tooling AND prefer to avoid the verbosity of "all" when running git status explicitly without options... are out of luck for now (no change).

Users who have the "status.showuntrackedfiles" config set to "all" and yet frequently explicitly call
'git status --untracked-files=normal' (and use the untracked cache) are the only ones who will be disadvantaged by this change. Their "--untracked-files=normal" calls will, after this change, no longer use the untracked cache.

---
 dir.c                             | 88 +++++++++++++++++++++++++------
 t/t7063-status-untracked-cache.sh | 81 ++++++++++++++++++++++++++++
 2 files changed, 153 insertions(+), 16 deletions(-)

index f2b0f242101..26c4d141ab9 100644
@@ -2747,13 +2747,33 @@ static void set_untracked_ident(struct untracked_cache *uc)
 	strbuf_addch(&uc->ident, 0);
 }
 
-static void new_untracked_cache(struct index_state *istate)
+static unsigned new_untracked_cache_flags(struct index_state *istate)
+{
+	struct repository *repo = istate->repo;
+	char *val;
+
+	/*
+	 * This logic is coordinated with the setting of these flags in
+	 * wt-status.c#wt_status_collect_untracked(), and the evaluation
+	 * of the config setting in commit.c#git_status_config()
+	 */
+	if (!repo_config_get_string(repo, "status.showuntrackedfiles", &val) &&
+	    !strcmp(val, "all"))
+		return 0;
+
+	/*
+	 * The default, if "all" is not set, is "normal" - leading us here.
+	 * If the value is "none" then it really doesn't matter.
+	 */
+	return DIR_SHOW_OTHER_DIRECTORIES | DIR_HIDE_EMPTY_DIRECTORIES;
+}
+
+static void new_untracked_cache(struct index_state *istate, int flags)
 {
 	struct untracked_cache *uc = xcalloc(1, sizeof(*uc));
 	strbuf_init(&uc->ident, 100);
 	uc->exclude_per_dir = ".gitignore";
-	/* should be the same flags used by git-status */
-	uc->dir_flags = DIR_SHOW_OTHER_DIRECTORIES | DIR_HIDE_EMPTY_DIRECTORIES;
+	uc->dir_flags = flags >= 0 ? flags : new_untracked_cache_flags(istate);
 	set_untracked_ident(uc);
 	istate->untracked = uc;
 	istate->cache_changed |= UNTRACKED_CHANGED;

@@ -2762,11 +2782,11 @@ static void new_untracked_cache(struct index_state *istate)
 void add_untracked_cache(struct index_state *istate)
 {
 	if (!istate->untracked) {
-		new_untracked_cache(istate);
+		new_untracked_cache(istate, -1);
 	} else {
 		if (!ident_in_untracked(istate->untracked)) {
 			free_untracked_cache(istate->untracked);
-			new_untracked_cache(istate);
+			new_untracked_cache(istate, -1);
 		}
 	}
 }

@@ -2814,17 +2834,9 @@ static struct untracked_cache_dir *validate_untracked_cache(struct dir_struct *d
 	if (base_len || (pathspec && pathspec->nr))
 		return NULL;
 
-	/* Different set of flags may produce different results */
-	if (dir->flags != dir->untracked->dir_flags ||
-	    /*
-	     * See treat_directory(), case index_nonexistent. Without
-	     * this flag, we may need to also cache .git file content
-	     * for the resolve_gitlink_ref() call, which we don't.
-	     */
-	    !(dir->flags & DIR_SHOW_OTHER_DIRECTORIES) ||
-	    /* We don't support collecting ignore files */
-	    (dir->flags & (DIR_SHOW_IGNORED | DIR_SHOW_IGNORED_TOO |
-			   DIR_COLLECT_IGNORED)))
+	/* We don't support collecting ignore files */
+	if (dir->flags & (DIR_SHOW_IGNORED | DIR_SHOW_IGNORED_TOO |
+			DIR_COLLECT_IGNORED))
 		return NULL;
 
 	/*

@@ -2847,6 +2859,50 @@ static struct untracked_cache_dir *validate_untracked_cache(struct dir_struct *d
 		return NULL;
 	}
 
+	/*
+	 * If the untracked structure we received does not have the same flags
+	 * as requested in this run, we're going to need to either discard the
+	 * existing structure (and potentially later recreate), or bypass the
+	 * untracked cache mechanism for this run.
+	 */
+	if (dir->flags != dir->untracked->dir_flags) {
+		/*
+		 * If the untracked structure we received does not have the same flags
+		 * as configured, then we need to reset / create a new "untracked"
+		 * structure to match the new config.
+		 *
+		 * Keeping the saved and used untracked cache consistent with the
+		 * configuration provides an opportunity for frequent users of
+		 * "git status -uall" to leverage the untracked cache by aligning their
+		 * configuration - setting "status.showuntrackedfiles" to "all" or
+		 * "normal" as appropriate.
+		 *
+		 * Previously using -uall (or setting "status.showuntrackedfiles" to
+		 * "all") was incompatible with untracked cache and *consistently*
+		 * caused surprisingly bad performance (with fscache and fsmonitor
+		 * enabled) on Windows.
+		 *
+		 * IMPROVEMENT OPPORTUNITY: If we reworked the untracked cache storage
+		 * to not be as bound up with the desired output in a given run,
+		 * and instead iterated through and stored enough information to
+		 * correctly serve both "modes", then users could get peak performance
+		 * with or without '-uall' regardless of their
+		 * "status.showuntrackedfiles" config.
+		 */
+		if (dir->untracked->dir_flags != new_untracked_cache_flags(istate)) {
+			free_untracked_cache(istate->untracked);
+			new_untracked_cache(istate, dir->flags);
+			dir->untracked = istate->untracked;
+		}
+		else {
+			/*
+			 * Current untracked cache data is consistent with config, but not
+			 * usable in this request/run; just bypass untracked cache.
+			 */
+			return NULL;
+		}
+	}
+
 	if (!dir->untracked->root) {
 		/* Untracked cache existed but is not initialized; fix that */
 		FLEX_ALLOC_STR(dir->untracked->root, name, "");

gitgitgadget