The rebase to v2.52.0 #820
Conversation
While performing a fetch with a virtual file system we know that there will be missing objects and we don't want to download them just because of the reachability of the commits. We also don't want to download a pack file with commits, trees, and blobs since these will be downloaded on demand. This flag will skip the first connectivity check and by returning zero will skip the upload pack. It will also skip the second connectivity check but continue to update the branches to the latest commit ids. Signed-off-by: Kevin Willford <[email protected]>
Ensure all filters and EOL conversions are blocked when running under GVFS so that our projected file sizes will match the actual file size when it is hydrated on the local machine. Signed-off-by: Ben Peart <[email protected]>
This adds hard-coded call to GVFS.hooks.exe before and after each Git command runs. To make sure that this is only called on repositories cloned with GVFS, we test for the tell-tale .gvfs. 2021-10-30: Recent movement of find_hook() to hook.c required moving these changes out of run-command.c to hook.c. 2025-11-06: The `warn_on_auto_comment_char` hack is so ugly that it forces us to pile similarly ugly code on top because that hack _expects_ that the config has not been read when `cmd_commit()`, `cmd_revert()`, `cmd_cherry_pick()`, `cmd_merge()`, or `cmd_rebase()` set that flag. But with the `pre_command()` hook already run, that assumption is incorrect. Signed-off-by: Ben Peart <[email protected]> Signed-off-by: Johannes Schindelin <[email protected]>
The idea is to allow blob objects to be missing from the local repository, and to load them lazily on demand. After discussing this idea on the mailing list, we will rename the feature to "lazy clone" and work more on this. Signed-off-by: Ben Peart <[email protected]> Signed-off-by: Johannes Schindelin <[email protected]>
Suggested by Ben Peart. Signed-off-by: Johannes Schindelin <[email protected]>
Hydrate missing loose objects in check_and_freshen() when running virtualized. Add test cases to verify read-object hook works when running virtualized. This hook is called in check_and_freshen() rather than check_and_freshen_local() to make the hook work also with alternates. Helped-by: Kevin Willford <[email protected]> Signed-off-by: Ben Peart <[email protected]>
Verify that the core.hooksPath configuration is repsected by the pre-command hook. Original regression test was written by Alejandro Pauly. Signed-off-by: Matthew John Cheetham <[email protected]>
If we are going to write an object there is no use in calling the read object hook to get an object from a potentially remote source. We would rather just write out the object and avoid the potential round trip for an object that doesn't exist. This change adds a flag to the check_and_freshen() and freshen_loose_object() functions' signatures so that the hook is bypassed when the functions are called before writing loose objects. The check for a local object is still performed so we don't overwrite something that has already been written to one of the objects directories. Based on a patch by Kevin Willford. Signed-off-by: Johannes Schindelin <[email protected]>
Signed-off-by: Alejandro Pauly <[email protected]>
When using the sparse-checkout feature, the file might not be on disk because the skip-worktree bit is on. This used to be a bug in the (hence deleted) `recursive` strategy. Let's ensure that this bug does not resurface. Signed-off-by: Kevin Willford <[email protected]> Signed-off-by: Johannes Schindelin <[email protected]>
The 'git worktree' command was marked as BLOCK_ON_GVFS_REPO because it does not interact well with the virtual filesystem of VFS for Git. When a Scalar clone uses the GVFS protocol, it enables the GVFS_BLOCK_COMMANDS flag, since commands like 'git gc' do not work well with the GVFS protocol. However, 'git worktree' works just fine with the GVFS protocol since it isn't doing anything special. It copies the sparse-checkout from the current worktree, so it does not have performance issues. This is a highly requested option. The solution is to stop using the BLOCK_ON_GVFS_REPO option and instead add a special-case check in cmd_worktree() specifically for a particular bit of the 'core_gvfs' global variable (loaded by very early config reading) that corresponds to the virtual filesystem. The bit that most closely resembled this behavior was non-obviously named, but does provide a signal that we are in a Scalar clone and not a VFS for Git clone. The error message is copied from git.c, so it will have the same output as before if a user runs this in a VFS for Git clone. Signed-off-by: Derrick Stolee <[email protected]>
When using the sparse-checkout feature git should not write to the working directory for files with the skip-worktree bit on. With the skip-worktree bit on the file may or may not be in the working directory and if it is not we don't want or need to create it by calling checkout_entry. There are two callers of checkout_target. Both of which check that the file does not exist before calling checkout_target. load_current which make a call to lstat right before calling checkout_target and check_preimage which will only run checkout_taret it stat_ret is less than zero. It sets stat_ret to zero and only if !stat->cached will it lstat the file and set stat_ret to something other than zero. This patch checks if skip-worktree bit is on in checkout_target and just returns so that the entry doesn't not end up in the working directory. This is so that apply will not create a file in the working directory, then update the index but not keep the working directory up to date with the changes that happened in the index. Signed-off-by: Kevin Willford <[email protected]>
Signed-off-by: Kevin Willford <[email protected]>
Add the ability to block built-in commands based on if the `core.gvfs` setting has the `GVFS_USE_VIRTUAL_FILESYSTEM` bit set. This allows us to selectively block commands that use the GVFS protocol, but don't use VFS for Git (for example repos cloned via `scalar clone` against Azure DevOps). Signed-off-by: Matthew John Cheetham <[email protected]>
As of 9e59b38 (object-file: emit corruption errors when detected, 2022-12-14), Git will loudly complain about corrupt objects. That is fine, as long as the idea isn't to re-download locally-corrupted objects. But that's exactly what we want to do in VFS for Git via the `read-object` hook, as per the `GitCorruptObjectTests` code added in microsoft/VFSForGit@2db0c030eb25 (New features: [...] - GVFS can now recover from corrupted git object files [...] , 2018-02-16). So let's support precisely that, and add a regression test that ensures that re-downloading corrupt objects via the `read-object` hook works. While at it, avoid the XOR operator to flip the bits, when we actually want to make sure that they are turned off: Use the AND-NOT operator for that purpose. Helped-by: Matthew John Cheetham <[email protected]> Helped-by: Derrick Stolee <[email protected]> Signed-off-by: Johannes Schindelin <[email protected]>
Loosen the blocking of the `repack` command from all "GVFS repos" (those that have `core.gvfs` set) to only those that actually use the virtual file system (VFS for Git only). This allows for `repack` to be used in Scalar clones. Signed-off-by: Matthew John Cheetham <[email protected]>
String formatting can be a performance issue when there are hundreds of thousands of trees. Change to stop using the strbuf_addf and just add the strings or characters individually. There are a limited number of modes so added a switch for the known ones and a default case if something comes through that are not a known one for git. In one scenario regarding a huge worktree, this reduces the time required for a `git checkout <branch>` from 44 seconds to 38 seconds, i.e. it is a non-negligible performance improvement. Signed-off-by: Kevin Willford <[email protected]>
Loosen the blocking of the `fsck` command from all "GVFS repos" (those that have `core.gvfs` set) to only those that actually use the virtual file system (VFS for Git only). This allows for `fsck` to be used in Scalar clones. Signed-off-by: Matthew John Cheetham <[email protected]>
The following commands and options are not currently supported when working in a GVFS repo. Add code to detect and block these commands from executing. 1) fsck 2) gc 4) prune 5) repack 6) submodule 8) update-index --split-index 9) update-index --index-version (other than 4) 10) update-index --[no-]skip-worktree 11) worktree Signed-off-by: Ben Peart <[email protected]> Signed-off-by: Johannes Schindelin <[email protected]>
Loosen the blocking of the `prune` command from all "GVFS repos" (those that have `core.gvfs` set) to only those that actually use the virtual file system (VFS for Git only). This allows for `prune` to be used in Scalar clones. Signed-off-by: Matthew John Cheetham <[email protected]>
In earlier versions of `microsoft/git`, we found a user who had set `core.gvfs = false` in their global config. This should not have been necessary, but it also should not have caused a problem. However, it did. The reason was that `gvfs_load_config_value()` was called from `config.c` when reading config key/value pairs from all the config files. The local config should override the global config, and this is done by `config.c` reading the global config first then reading the local config. However, our logic only allowed writing the `core_gvfs` variable once. In v2.51.0, we had to adapt to upstream changes that changed way the `core.gvfs` config value is read, and the special handling is no longer necessary, yet we still want the test case that ensures that this bug does not experience a regression. Signed-off-by: Derrick Stolee <[email protected]> Signed-off-by: Johannes Schindelin <[email protected]>
Replace the special casing of the `worktree` command being blocked on VFS-enabled repos with the new `BLOCK_ON_VFS_ENABLED` flag. Signed-off-by: Matthew John Cheetham <[email protected]>
Emit a warning message when the `gvfs.sharedCache` option is set that the `repack` command will not perform repacking on the shared cache. In the future we can teach `repack` to operate on the shared cache, at which point we can drop this commit. Signed-off-by: Matthew John Cheetham <[email protected]>
Teach STATUS to optionally serialize the results of a status computation to a file. Teach STATUS to optionally read an existing serialization file and simply print the results, rather than actually scanning. This is intended for immediate status results on extremely large repos and assumes the use of a service/daemon to maintain a fresh current status snapshot. 2021-10-30: packet_read() changed its prototype in ec9a37d (pkt-line.[ch]: remove unused packet_read_line_buf(), 2021-10-14). 2021-10-30: sscanf() now does an extra check that "%d" goes into an "int" and complains about "uint32_t". Replacing with "%u" fixes the compile-time error. 2021-10-30: string_list_init() was removed by abf897b (string-list.[ch]: remove string_list_init() compatibility function, 2021-09-28), so we need to initialize manually. Signed-off-by: Jeff Hostetler <[email protected]> Signed-off-by: Derrick Stolee <[email protected]>
Teach status serialization to take an optional pathname on
the command line to direct that cache data be written there
rather than to stdout. When used this way, normal status
results will still be written to stdout.
When no path is given, only binary serialization data is
written to stdout.
Usage:
git status --serialize[=<path>]
Signed-off-by: Jeff Hostetler <[email protected]>
Teach status deserialize code to reject status cache when printing in porcelain V2 and there are unresolved conflicts in the cache file. A follow-on task might extend the cache format to include this additiona data. See code for longer explanation. Signed-off-by: Jeff Hostetler <[email protected]>
On index load, clear/set the skip worktree bits based on the virtual file system data. Use virtual file system data to update skip-worktree bit in unpack-trees. Use virtual file system data to exclude files and folders not explicitly requested. Update 2022-04-05: disable the "present-despite-SKIP_WORKTREE" file removal behavior when 'core.virtualfilesystem' is enabled. Signed-off-by: Ben Peart <[email protected]>
In some instances, CodeQL's web UI on github.com leaves questions unanswered. For example, in some alerts it is really necessary to follow the entire "taint flow" to understand why something might be an issue. The alerts for the `cpp/uncontrolled-allocation-size` rule, for example, are all false positives, and only when inspecting the exact flow does it become obvious that one alert wants to point out that the size of a binary patch hunk, which is specified in the patch, is then used to determine how much memory to allocate, which may potentially run out of memory (and is hence just Git doing what it is asked to, and does not need to be changed). To help with those issues, publish the `.sarif` file as part of every workflow run; This allows downloading that file and inspecting it e.g. with the SARIF viewer extension in VS Code (for details, see https://marketplace.visualstudio.com/items?itemName=MS-SarifVSCode.sarif-viewer). Signed-off-by: Johannes Schindelin <[email protected]>
CodeQL points out that `branch_get()` can return NULL values. Signed-off-by: Johannes Schindelin <[email protected]>
The code is a bit too hard to reason about for CodeQL to figure out whether the `fill_commit_graph_info()` function is at all called after `write_commit_graph()` returns (and hence whether `topo_levels` goes out of context before it is used again). The Git project insists that this is correct (and does not want to make the code more obviously correct), so let's silence CodeQL's complaints in this instance. Signed-off-by: Johannes Schindelin <[email protected]>
A couple of CodeQL's queries are opinionated in a way that is obviously not shared by Git's source code's state, and apparently intentionally so. For example, the "For loop variable changed in body" query as well as the "No trivial switch statements" one result in too many results that are apparently intentional in Git's source code. Let's not worry about those, then. Also, Git has plenty of instances where variables shadow other variables. Other valid yet not quite critical issues identified by CodeQL include complex conditionals and nested switch statements spanning several pages. We probably want to address these issues at some stage, but they are not as critical as other problems pointed out by CodeQL, so let's silence those queries for now and take care of them at a later stage. Signed-off-by: Johannes Schindelin <[email protected]>
Signed-off-by: Johannes Schindelin <[email protected]>
Signed-off-by: Johannes Schindelin <[email protected]>
…ray past end Signed-off-by: Johannes Schindelin <[email protected]>
Signed-off-by: Johannes Schindelin <[email protected]>
A few places where CodeQL thinks that variables might be uninitialized. Signed-off-by: Johannes Schindelin <[email protected]>
…oes NUL-terminate correctly Signed-off-by: Johannes Schindelin <[email protected]>
These patches implement some defensive programming to address complaints some static analyzers might have. Signed-off-by: Johannes Schindelin <[email protected]>
Let's exclude GitWeb from being scanned; It is not distributed by us. Signed-off-by: Johannes Schindelin <[email protected]>
CodeQL pointed out a couple of issues, which are addressed in this patch series. Signed-off-by: Johannes Schindelin <[email protected]>
This patch series has been long in the making, ever since Johannes Nicolai and myself spiked this in November/December 2020. Signed-off-by: Johannes Schindelin <[email protected]>
The microsoft/git fork includes pre- and post-command hooks, with the
initial intention of using these for VFS for Git. In that environment,
these are important hooks to avoid concurrent issues when the
virtualization is incomplete.
However, in the Office monorepo the post-command hook is used in a
different way. A custom hook is used to update the sparse-checkout, if
necessary. To avoid this hook from being incredibly slow on every Git
command, this hook checks for the existence of a "sentinel file" that is
written by a custom post-index-change hook and no-ops if that file does
not exist.
However, even this "no-op" is 200ms due to the use of two scripts (one
simple script in .git/hooks/ does some environment checking and then
calls a script from the working directory which actually contains the
logic).
Add a new config option, 'postCommand.strategy', that will allow for
multiple possible strategies in the future. For now, the one we are
adding is 'post-index-change' which states that we should write a
sentinel file instead of running the 'post-index-change' hook and then
skip the 'post-command' hook if the proper sentinel file doesn't exist.
(If it does exist, then delete it and run the hook.)
---
This fork contains changes specific to monorepo scenarios. If you are an
external contributor, then please detail your reason for submitting to
this fork:
* [ ] This is an early version of work already under review upstream.
* [ ] This change only applies to interactions with Azure DevOps and the
GVFS Protocol.
* [ ] This change only applies to the virtualization hook and VFS for
Git.
* [x] This change only applies to custom bits in the microsoft/git fork.
This patch series has been long in the making, ever since Johannes Nicolai and myself spiked this in November/December 2020. Signed-off-by: Johannes Schindelin <[email protected]>
Range-diff relative to v2.52.0-rc2
|
After all the -rc rebases, this is the real rebase.