aboutsummaryrefslogtreecommitdiff
path: root/lisp
AgeCommit message (Collapse)Author
2025-05-01Release version 4.3.3v4.3.3Jonas Bernoulli
2025-05-01Bump dependenciesJonas Bernoulli
2025-04-28magit-{branch-unshelve,insert-shelved-branches}: List newest firstJonas Bernoulli
2025-04-28magit-credential-cache-daemon-socket: CosmeticsJonas Bernoulli
2025-04-28magit-mouse-set-point: New commandJonas Bernoulli
2025-04-26magit-diff-paint-hunk: Only refine if actually paintingJonas Bernoulli
2025-04-25magit-rebase--todo: Fix order of valueJonas Bernoulli
The order was reversed in the recent [1: 8e72767262]. 1: 2025-04-15 8e7276726242ed7040a9c721a247a6d52397da95 Speed-up inserting rebase sequences
2025-04-25make: Set lexical-binding in magit-version.elJonas Bernoulli
Closes #5361.
2025-04-22magit--merge-absorb: Add MESSAGE argumentJonas Bernoulli
`forge-pullreq-state-set-merged' is going to use this.
2025-04-22magit-merge-dissolve: Avoid duplicating workJonas Bernoulli
2025-04-22magit-get-local-upstream-branch: New functionJonas Bernoulli
2025-04-19magit-diff-visit-file: Remove inaccurate sentence from documentationRichard Sent
Co-authored-by: Jonas Bernoulli <jonas@bernoul.li>
2025-04-18magit-blame-styles: Support %h to truncate hashesJonas Bernoulli
Closes #4057.
2025-04-16magit-insert-shelved-branches: Fix margin highlightingJonas Bernoulli
2025-04-16magit-insert-shelved-branches: Indent branchesJonas Bernoulli
As for other ref lists.
2025-04-16magit-section-make-overlay: Remove kludge for bad user settingJonas Bernoulli
We strongly discourage setting `magit-section-keep-region-overlay' to t. If a user did that anyway that can lead to conflicts between the involved faces. Before we tried to automatically avoid such a potential conflict by only using the `:foreground' of the `magit-*-heading-selection' faces, which is very confusing for users who want to enable this and also use another face attribute.
2025-04-16magit-section-update-highlight: Always unhighlight sectionsJonas Bernoulli
Start doing so even after long-lines shortcuts kick in and even if the new current section is the root section. In both cases sections might be highlighted that should no longer be highlighted.
2025-04-16magit-section--maybe-enable-long-lines-shortcuts: New functionJonas Bernoulli
Replacing `magit-section--enable-long-lines-shortcuts'.
2025-04-16magit-diff-paint-hunk: CosmeticsJonas Bernoulli
2025-04-16magit-wip-purge: New commandJonas Bernoulli
2025-04-16magit-insert-shelved-branches: New functionJonas Bernoulli
2025-04-16Prefix names of shelved branches with the tip's committer dateJonas Bernoulli
Remove that prefix when unshelving, but keep supporting unshelving branches that were shelved without adding such a prefix.
2025-04-16magit-file-dispatch: CosmeticsJonas Bernoulli
2025-04-16magit-dired-log: If no files are marked, show log for directoryJonas Bernoulli
That way we can use it instead of `magit-log-buffer-file' in `magit-file-dispatch', without losing the ability to quickly show the log for the visited directory.
2025-04-16magit-dired-log: Do not limit to tracked filesJonas Bernoulli
If a file was tracked in the past, isn't being tracked anymore, but still/again exists for some reason, we might still be interested in its history.
2025-04-16magit-dired-{stage,unstage}: New commandsJonas Bernoulli
2025-04-16magit-dired.el: New libraryJonas Bernoulli
2025-04-15git-rebase-kill-line: If the line is commented then uncomment itJonas Bernoulli
2025-04-15magit-rebase-insert-merge-sequence: Insert sections for all stepsJonas Bernoulli
2025-04-15Speed-up inserting rebase sequencesJonas Bernoulli
2025-04-15magit-insert-rebase-sequence: Rename internal variableJonas Bernoulli
The two backends are now named "merge" and "apply"; no longer "interactive" and "am", see Git commit [1: 10cdb9f38a]. 1: 2020-02-16 10cdb9f38adaa16aad64403bc931405564523985 rebase: rename the two primary rebase backends
2025-04-15magit-git-dir: Fix obsolescence declarationJonas Bernoulli
2025-04-15Use ## moreJonas Bernoulli
2025-04-15magit-section-show-level: Support acting on selected sectionsX4lldux
2025-04-15magit-section-show-level: Improve docstringJonas Bernoulli
2025-04-15magit-process-section: Define earlier to avoid warningJonas Bernoulli
The `process' slot is added by this class and used in `magit-start-process'.
2025-04-01Release version 4.3.2v4.3.2Jonas Bernoulli
2025-04-01Bump dependenciesJonas Bernoulli
2025-03-31magit-{stage,unstage}-file: Rename to magit-{stage,unstage}-filesJonas Bernoulli
Also no longer allow the FILES argument to be a single file instead of a list of one or more files.
2025-03-31magit-{stage,unstage}-buffer-file: Rename to magit-file-{stage,unstage}Jonas Bernoulli
For consistency with the other `magit-file-*' commands.
2025-03-31magit-{stage,unstage}-buffer-file: Move definitionsJonas Bernoulli
Move them next to the definitions of the related `magit-file-*' commands.
2025-03-31magit-file-untrack: Move definitionJonas Bernoulli
Order definitions the same as these commands are ordered in the menu.
2025-03-31magit-file-{rename,untrack,delete}: AutoloadJonas Bernoulli
2025-03-31magit-rebase-autosquash: With a prefix allow selecting beginningJonas Bernoulli
2025-03-31magit--git-insert: Do not return empty string as error messageJonas Bernoulli
When RETURN-ERROR is `full' this function returns all stderr output. If there was nothing on stderr, then it returned the empty string, now it returns the exit code in that case too. If a caller of `magit--git-wash' uses `wash-anyway' for KEEP-ERROR, it likely does so because a non-zero exit code does not necessarily mean that an error occurred. If `magit--git-insert' cannot find any error string in such a case, that is because there wasn't actually an error, and it must return the exit code. This does not affect the case when there is an actual error because that would be accompanied by non-empty stdout. Callers can then still interpret the non-zero, non-string return value as an error, but they don't have to. (magit--git-wash ... 'wash-anyway ...) does not and because it no longer ends up inserting the empty string and some separator whitespace before the output from stdout, stdout washing does no longer break. The command `magit-diff-paths' uses `magit--git-wash' as described, and it broke when support for `full' as a special RETURN-ERROR was added in [1: 1f17e515cf]. (Like "diff", "git diff --no-index" exits with 1 if there are differences.) Fixes second coming of #5093. 1: 2024-11-15 1f17e515cfbef3acfd2bdf990cd206d6ae02010b magit--git-{insert,wash}: Potentially use complete error text
2025-03-31magit-git-wash: Always keep error outputJonas Bernoulli
Remove the variable `magit--git-wash-keep-error' which previously could be used to disable that. The variable was added in [1: 78a979fde5] and its value was changed to t in [2: 76253bda0d]. This has generally held up well enough to be permanently enabled. Function `magit--git-wash', which was added in [1: 78a979fde5], can be used in cases where the section should be canceled in case of an error. Also extend the docstrings of these two functions. 1: 2023-03-18 78a979fde52815242b165f083d171259db28e0b4 Support displaying errors that occur when washing a section 2: 2024-11-15 76253bda0d05d252e118e56955447b6e4d6c0535 magit--git-wash-keep-error: Enable by default
2025-03-31magit-git-debug: Add docstringJonas Bernoulli
2025-03-31magit-git-debug: Rename value all to include-successJonas Bernoulli
"All" is misleading as this does not include each and every git invocation. There are low-level functions which by pass even this most extreme setting.
2025-03-31magit-toggle-git-debug: Fix key binding in docstringJonas Bernoulli
2025-03-31magit-git-executable: Change name of output bufferJonas Bernoulli
Naming it *magit-git-debug* because there is a variable by that name, which enables debugging that does *not* use this buffer.