Rebase to v2.24.0-rc0 #2360
Merged
Rebase to v2.24.0-rc0 #2360
Conversation
This file contains hidden or bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
|
Note: during the latest Git IRC standup, the Git maintainer indicated that v2.23.1 will not happen before v2.24.0 (enough patches accumulated a bugfix release, but the -rc0..-rc3 time is as good a time as any to focus on the upcoming "big" release). Therefore, Git for Windows' |
These fixes were necessary for Sverre Rabbelier's remote-hg to work, but for some magic reason they are not necessary for the current remote-hg. Makes you wonder how that one gets away with it. Signed-off-by: Johannes Schindelin <[email protected]>
Turns out that it don't work so well on Vista, see git-for-windows#1742 for details. According to https://devblogs.microsoft.com/oldnewthing/?p=8873, it *should* work on Windows Vista and later. But apparently there are issues on Windows Vista when pipes are involved. Given that Windows Vista is past its end of life (official support ended on April 11th, 2017), let's not spend *too* much time on this issue and just disable the file handle inheritance restriction on any Windows version earlier than Windows 7. Signed-off-by: Johannes Schindelin <[email protected]>
…er/register_rename_src diffcore-rename: speed up register_rename_src
This topic branch allows us to specify absolute paths without the drive prefix e.g. when cloning. Example: C:\Users\me> git clone https://github.com/git/git \upstream-git This will clone into a new directory C:\upstream-git, in line with how Windows interprets absolute paths. Signed-off-by: Johannes Schindelin <[email protected]>
Fix fetching from UNC paths. Signed-off-by: Johannes Schindelin <[email protected]>
When compiling Git with a runtime prefix (so that it can be installed into any location, finding its libexec/ directory relative to the location of the `git` executable), it is convenient to provide "absolute" Unix-y paths e.g. for http.sslCAInfo, and have those absolute paths be resolved relative to the runtime prefix. This patch makes it so for Windows. It is up for discussion whether we want this for other platforms, too, as long as building with RUNTIME_PREFIX. Signed-off-by: Johannes Schindelin <[email protected]>
Add a helper function to start GDB that was already attached to the current process
This topic branch avoids spawning `gzip` when asking `git archive` to create `.tar.gz` files. Signed-off-by: Johannes Schindelin <[email protected]>
This is hardly the first conversion of a Git command that is implemented as a script to a built-in. So far, the most successful strategy for such conversions has been to add a built-in helper and call that for more and more functionality from the script, as more and more parts are converted. With the interactive add, we choose a different strategy. The sole reason for this is that on Windows (where such a conversion has the most benefits in terms of speed and robustness) we face the very specific problem that a `system()` call in Perl seems to close `stdin` in the parent process when the spawned process consumes even one character from `stdin`. And that just does not work for us here, as it would stop the main loop as soon as any interactive command was performed by the helper. Which is almost all of the commands in `git add -i`. It is almost as if Perl told us once again that it does not want us to use it on Windows. Instead, we follow the opposite route where we start with a bare-bones version of the built-in interactive add, guarded by the new `add.interactive.useBuiltin` config variable, and then add more and more functionality to it, until it is feature complete. At this point, the built-in version of `git add -i` only states that it cannot do anything yet ;-) Signed-off-by: Johannes Schindelin <[email protected]>
Make the diffstat interface (namely, the diffstat_t struct and compute_diffstat) no longer be internal to diff.c and allow it to be used by other parts of git. This is helpful for code that may want to easily extract information from files using the diff machinery, while flushing it differently from how the show_* functions used by diff_flush() do it. One example is the builtin implementation of git-add--interactive's status. Signed-off-by: Daniel Ferreira <[email protected]> Signed-off-by: Slavica Đukić <[email protected]> Signed-off-by: Johannes Schindelin <[email protected]>
This implements the `status` command of `git add -i`. The data structures introduced in this commit will be extended later, as needed. At this point, we re-implement only part of the `list_and_choose()` function of the Perl script `git-add--interactive.perl` and call it `list()`. It does not yet color anything, or do columns, or allow user input. Over the course of the next commits, we will introduce a `list_and_choose()` function that uses `list()` to display the list of options and let the user choose one or more of the displayed items. This will be used to implement the main loop of the built-in `git add -i`, at which point the new `status` command can actually be used. Signed-off-by: Daniel Ferreira <[email protected]> Signed-off-by: Slavica Đukić <[email protected]> Signed-off-by: Johannes Schindelin <[email protected]>
For simplicity, we only implemented the `status` command without colors. This patch starts adding color, matching what the Perl script `git-add--interactive.perl` does. Original-Patch-By: Daniel Ferreira <[email protected]> Signed-off-by: Slavica Đukić <[email protected]> Signed-off-by: Johannes Schindelin <[email protected]>
The reason why we did not start with the main loop to begin with is that it is the first user of `list_and_choose()`, which uses the `list()` function that we conveniently introduced for use by the `status` command. Apart from the "and choose" part, there are more differences between the way the `status` command calls the `list_and_choose()` function in the Perl version of `git add -i` compared to the other callers of said function. The most important ones: - The list is not only shown, but the user is also asked to make a choice, possibly selecting multiple entries. - The list of items is prefixed with a marker indicating what items have been selected, if multi-selection is allowed. - Initially, for each item a unique prefix (if there exists any within the given parameters) is determined, and shown in the list, and accepted as a shortcut for the selection. These features will be implemented later, except the part where the user can choose a command. At this stage, though, the built-in `git add -i` still only supports the `status` command, with the remaining commands to follow over the course of the next commits. In addition, we also modify `list()` to support displaying the commands in columns, even if there is currently only one. The Perl script `git-add--interactive.perl` mixed the purposes of the "list" and the "and choose" part into the same function. In the C version, we will keep them separate instead, calling the `list()` function from the `list_and_choose()` function. Note that we only have a prompt ending in a single ">" at this stage; later commits will add commands that display a double ">>" to indicate that the user is in a different loop than the main one. Signed-off-by: Johannes Schindelin <[email protected]>
Just like in the Perl script `git-add--interactive.perl`, for each command a unique prefix is determined (if there exists any within the given parameters), and shown in the list, and accepted as a shortcut for the command. To determine the unique prefixes, as well as to look up the command in question, we use a copy of the list and sort it. While this might seem like overkill for a single command, it will make much more sense when all the commands are implemented, and when we reuse the same logic to present a list of files to edit, with convenient unique prefixes. At the start of the development of this patch series, a dedicated data structure was introduced that imitated the Trie that the Perl version implements. However, this was deemed overkill, and we now simply sort the list before determining the length of the unique prefixes by looking at each item's neighbor. As a bonus, we now use the same sorted list to perform a binary search using the user-provided prefix as search key. Original-patch-by: Slavica Đukić <[email protected]> Helped-by: SZEDER Gábor <[email protected]> Signed-off-by: Johannes Schindelin <[email protected]>
With this change, we print out the same colored help text that the Perl-based `git add -i` prints in the main loop when question mark is entered. Signed-off-by: Johannes Schindelin <[email protected]>
The error messages as well as the unique prefixes are colored in `git add -i` by default; We need to do the same in the built-in version. Signed-off-by: Slavica Đukić <[email protected]> Signed-off-by: Johannes Schindelin <[email protected]>
This imitates the code to show the help text from the Perl script `git-add--interactive.perl` in the built-in version. To make sure that it renders exactly like the Perl version of `git add -i`, we also add a test case for that to `t3701-add-interactive.sh`. Signed-off-by: Slavica Đukić <[email protected]> Signed-off-by: Johannes Schindelin <[email protected]>
In the `update` command of `git add -i`, we are primarily interested in the list of modified files that have worktree (i.e. unstaged) changes. At the same time, we need to determine _also_ the staged changes, to be able to produce the full added/deleted information. The Perl script version of `git add -i` has a parameter of the `list_modified()` function for that matter. In C, we can be a lot more precise, using an `enum`. The C implementation of the filter also has an easier time to avoid unnecessary work, simply by using an adaptive order of the `diff-index` and `diff-files` phases, and then skipping files in the second phase when they have not been seen in the first phase. Signed-off-by: Johannes Schindelin <[email protected]>
The `upgrade`, `revert` and `add-untracked` commands allow selecting multiple entries. Let's extend the `list_and_choose()` function to accommodate those use cases. Signed-off-by: Johannes Schindelin <[email protected]>
After `status` and `help`, it is now turn to port the `update` command to C, the second command that is shown in the main loop menu of `git add -i`. This `git add -i` command is the first one which lets the user choose a subset of a list of files, and as such, this patch lays the groundwork for the other commands of that category: - It teaches the `print_file_item()` function to show a unique prefix if we found any (the code to find it had been added already in the previous patch where we colored the unique prefixes of the main loop commands, but that patch uses the `print_command_item()` function to display the menu items). - This patch also adds the help text that is shown when the user input to select items from the shown list could not be parsed. - As `get_modified_files()` clears the list of files, it now has to take care of clearing the _full_ `prefix_item_list` lest the `sorted` and `selected` fields go stale and inconsistent. Signed-off-by: Johannes Schindelin <[email protected]>
This is a relatively straight-forward port from the Perl version, with the notable exception that we imitate `git reset -- <paths>` in the C version rather than the convoluted `git ls-tree HEAD -- <paths> | git update-index --index-info` followed by `git update-index --force-remove -- <paths>` for the missed ones. While at it, we fix the pretty obvious bug where the `revert` command offers to unstage files that do not have staged changes. Signed-off-by: Johannes Schindelin <[email protected]>
This is yet another command, ported to C. It builds nicely on the support functions introduced for other commands, with the notable difference that only names are displayed for untracked files, no file type or diff summary. Signed-off-by: Johannes Schindelin <[email protected]>
Well, it is not a full implementation yet. In the interest of making this easy to review (and easy to keep bugs out), we still hand off to the Perl script to do the actual work. The `patch` functionality actually makes up for more than half of the 1,800+ lines of `git-add--interactive.perl`. It will be ported from Perl to C incrementally, later. Signed-off-by: Johannes Schindelin <[email protected]>
It is not only laziness that we simply spawn `git diff -p --cached` here: this command needs to use the pager, and the pager needs to exit when the diff is done. Currently we do not have any way to make that happen if we run the diff in-process. So let's just spawn. Signed-off-by: Johannes Schindelin <[email protected]>
We do not really want to `exit()` here, of course, as this is safely libified code. Signed-off-by: Johannes Schindelin <[email protected]>
In this developer's workflows, it often happens that a hunk needs to be edited in a way that adds lines, and even reduces the context Let's add a regression test for this. Note that just like the preceding test case, the new test case is *not* handled gracefully by the current `git add -p`. It will be handled correctly by the upcoming built-in `git add -p`, though. Signed-off-by: Johannes Schindelin <[email protected]>
The TTY prerequisite is a rather heavy one: it not only requires Perl to work, but also the IO/Pty.pm module (with native support, and it requires pseudo terminals, too). In particular, test cases marked with the TTY prerequisite would be skipped in Git for Windows' SDK. In the case of `git add -p`, we do not actually need that big a hammer, as we do not want to test any functionality that requires a pseudo terminal; all we want is to talk the interactive add command to use color, even when being called from within the test suite. And we found exactly such a trick earlier already: when we added a test case to verify that the main loop of `git add -i` is colored appropriately. Let's use that trick instead of the TTY prerequisite. While at it, we avoid the pipes, as we do not want a SIGPIPE to break the regression test cases (which will be much more likely when we do not run everything through Perl because that is inherently slower). Signed-off-by: Johannes Schindelin <[email protected]>
The `git add -p` command offers different prompts for regular diff hunks vs mode change pseudo hunks vs diffs deleting files. Let's cover this in the regresion test suite, in preparation for re-implementing `git add -p` in C. For the mode change prompt, we use a trick that lets this test case pass even on systems without executable bit, i.e. where `core.filemode = false` (such as Windows): we first add the file to the index with `git add --chmod=+x`, and then call `git add -p` with `core.filemode` forced to `true`. The file on disk has no executable bit set, therefore we will see a mode change. Signed-off-by: Johannes Schindelin <[email protected]>
fetch-pack.c: enable fscache for stats under .git/objects
…t_flush checkout.c: enable fscache for checkout again Signed-off-by: Johannes Schindelin <[email protected]>
…_index Enable the filesystem cache (fscache) in refresh_index().
…-gfw fscache: use FindFirstFileExW to avoid retrieving the short name
…ter-status-gfw status: disable and free fscache at the end of the status command
…e-gfw fscache: add GIT_TEST_FSCACHE support
…ter-add-gfw At the end of the add command, disable and free the fscache
…ics-gfw fscache: add fscache hit statistics
This brings substantial wins in performance because the FSCache is now per-thread, being merged to the primary thread only at the end, so we do not have to lock (except while merging). Signed-off-by: Johannes Schindelin <[email protected]>
…safe-enable-gfw fscache: make fscache_enable() thread safe
…DirectoryFile-gfw fscache: teach fscache to use NtQueryDirectoryFile
When updating the skip-worktree bits in the index to align with new values in a sparse-checkout file, Git scans the entire working directory with lstat() calls. In a sparse-checkout, many of these lstat() calls are for paths that do not exist. Enable the fscache feature during this scan. In a local test of a repo with ~2.2 million paths, updating the index with `git read-tree -m -u HEAD` with a sparse-checkout file containing only `/.gitattributes` improved from 2-3 minutes to 15-20 seconds. More work could be done to stop running lstat() calls when recursing into directories that are known to not exist.
We already avoid traversing NTFS junction points in `git clean -dfx`. With this topic branch, we do that when the FSCache is enabled, too. Signed-off-by: Johannes Schindelin <[email protected]>
This was pull request git-for-windows#1645 from ZCube/master Support windows container. Signed-off-by: Johannes Schindelin <[email protected]>
Specify symlink type in .gitattributes
Signed-off-by: Johannes Schindelin <[email protected]>
Handle Ctrl+C in Git Bash nicely Signed-off-by: Johannes Schindelin <[email protected]>
This branch allows third-party tools to call `git status --no-lock-index` to avoid lock contention with the interactive Git usage of the actual human user. Signed-off-by: Johannes Schindelin <[email protected]>
…ored-directory-gracefully Phase out `--show-ignored-directory` gracefully
Add a README.md for GitHub goodness. Signed-off-by: Johannes Schindelin <[email protected]>
dscho
force-pushed
the
prepare-for-v2.24.0
branch
from
07b1c24 to
51b78e6
Compare
1 task
|
I managed to release https://github.com/git-for-windows/git/releases/tag/v2.24.0-rc0.windows.1 last night, now it is time to merge. |
Merged
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
Add this suggestion to a batch that can be applied as a single commit.
This suggestion is invalid because no changes were made to the code.
Suggestions cannot be applied while the pull request is closed.
Suggestions cannot be applied while viewing a subset of changes.
Only one suggestion per line can be applied in a batch.
Add this suggestion to a batch that can be applied as a single commit.
Applying suggestions on deleted lines is not supported.
You must change the existing code in this line in order to create a valid suggestion.
Outdated suggestions cannot be applied.
This suggestion has been applied or marked resolved.
Suggestions cannot be applied from pending reviews.
Suggestions cannot be applied on multi-line comments.
Suggestions cannot be applied while the pull request is queued to merge.
Suggestion cannot be applied right now. Please check back later.
This is a work-in-progress:
phase-out-*and thereadmebranches to the tip of the thicket.range-diffrelative tomastermakes sense.add-i/add-ipatches to the latest iteration.vcxprojinstead of viaMSVC=1).My plan was to make sure that every single commit compiles cleanly, but since I already spent a dozen hours today on getting things into a good shape, I think I'll skip this opportunity, deferring to the
-rc1rebase.