| Commit message (Collapse) | Author | Age | ||
|---|---|---|---|---|
| ... | ||||
| * | refactor(extmarks): use a more efficient representation | Björn Linse | 2022-01-15 | |
| | | | | | | | | | | | | | | | | | | | | | | | | | | | | marktree.c was originally constructed as a "generic" datatype, to make the prototyping of its internal logic as simple as possible and also as the usecases for various kinds of extmarks/decorations was not yet decided. As a consequence of this, various extra indirections and allocations was needed to use marktree to implement extmarks (ns/id pairs) and decorations of different kinds (some which is just a single highlight id, other an allocated list of virtual text/lines) This change removes a lot of indirection, by making Marktree specialized for the usecase. In particular, the namespace id and mark id is stored directly, instead of the 64-bit global id particular to the Marktree struct. This removes the two maps needed to convert between global and per-ns ids. Also, "small" decorations are stored inline, i.e. those who doesn't refer to external heap memory anyway. That is highlights (with priority+flags) are stored inline, while virtual text, which anyway occurs a lot of heap allocations, do not. (previously a hack was used to elide heap allocations for highlights with standard prio+flags) TODO(bfredl): the functionaltest-lua CI version of gcc is having severe issues with uint16_t bitfields, so splitting up compound assignments and redundant casts are needed. Clean this up once we switch to a working compiler version. | |||
| * | fix(extmark): fix missing virt_lines when using id param of set_extmark | Björn Linse | 2021-11-01 | |
| | | ||||
| * | refactor(decorations): mark decorations directly on the marktree | Björn Linse | 2021-10-23 | |
| | | | | | | | | | | | | | | This allows to more quickly skip though regions which has non-decorative marks when redrawing. This might seem like a gratuitous micro-optimization in isolation. But! Soon decorations are gonna crop into other hot inner-loop paths, including the plines.c code for calculating the horizontal and vertical space of text. Then we want to quickly skip over regions with "only" overlaying decorations (which do not affect text size) | |||
| * | fix(PVS/V1028): prevent possible overflow #16023 | dundargoc | 2021-10-17 | |
| | | | | | | | Full warning: "Possible overflow. Consider casting operands, not the result." https://pvs-studio.com/en/docs/warnings/v1028/ | |||
| * | refactor: reduce number of unique char casts (#15995) | dundargoc | 2021-10-12 | |
| | | ||||
| * | refactor: remove redundant casts | Dundar Göc | 2021-10-07 | |
| | | ||||
| * | refactor: format with uncrustify #15741 | dundargoc | 2021-09-22 | |
| | | ||||
| * | refactor(marktree): embed the keymap in the MarkTree struct | Björn Linse | 2021-08-22 | |
| | | ||||
| * | clang/'Dead assignment': ignore warning #15000 | dundargoc | 2021-07-05 | |
| | | | | | This assignment is informative/documenting though technically unnecessary. https://github.com/neovim/neovim/pull/11900#discussion_r381860165 . | |||
| * | extmark: fix deletable nodes in MarkTree sometimes getting skipped | snezhniylis | 2021-06-22 | |
| | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | As per #14236, performing extmark cleanup in a certain namespace does not guarantee removing all the extmarks inside given namespace. The issue resides within the tree node removal method and results in a couple of rare edge cases. To demonstrate what causes this bug, I'll give an example covering one of the edge cases. === AN EXAMPLE === (A) (B) (C) (D) (E) --------- --------- --------- --------- --------- <0, 1> <0, 1> <0, 1> <0, 1> <0, 1> <0, 2> <0, 2> <0, 2> <0, 2> <0, 2> <0, 3> <0, 3> <0, 3> <0, 3> <0, 3> <0, 4> <0, 4> <0, 4> <0, 4> <0, 4> <0, 5> <0, 5> <0, 5> <0, 5> <0, 5> <0, 6> <0, 6> <0, 6> <0, 6> <0, 6> <0, 7> <0, 7> <0, 7> <0, 7> <0, 7> <0, 8> <0, 8> <0, 8> <0, 8> <0, 8> <0, 9> <0, 9> * * <0, 9> * <0, 9> [0, 10] * [0, 10] <0, 9> [0, 11] [0, 11] [0, 11] [0, 11] [0, 11] [0, 12] [0, 12] * [0, 12] [0, 12] [0, 12] [0, 13] [0, 13] [0, 13] [0, 13] [0, 13] [0, 14] [0, 14] [0, 14] [0, 14] [0, 14] [0, 15] [0, 15] [0, 15] [0, 15] [0, 15] [0, 16] [0, 16] [0, 16] [0, 16] [0, 16] [0, 17] [0, 17] [0, 17] [0, 17] [0, 17] [0, 18] [0, 18] [0, 18] [0, 18] [0, 18] [0, 19] [0, 19] [0, 19] [0, 19] [0, 19] [0, 20] [0, 20] [0, 20] [0, 20] [0, 20] DIAGRAM EXPLANATION * Every column is a state of the marktree at a certain stage. * To make it simple, I don't draw the whole tree. What you see are 2 leftmost parent nodes ([0, 10], [0, 20]) and their children placed in order `MarkTreeIter` would iterate through. From top to bottom. * Numbers on this diagram represent extmark coordinates. Relative positioning and actual mark IDs used by the marktree are avoided for simplicity. * 2 types of brackets around coordinates represent 2 different extmark namespaces (`ns_id`s). * '*' shows iterator position. ACTUAL EXPLANATION Let's assume, we have two sets of extmarks from 2 different plugins: * Plugin1: <0, 1-9> * Plugin2: [0, 10-20] 1. Plugin2 calls `vim.api.nvim_buf_clear_namespace(buf_handle, ns_id, 0, -1)` to clear all its extmarks which results in `extmark_clear` call. 2. The iteration process goes on ignoring extmarks with irrelevant `ns_id` from Plugin1, until it reaches [0, 10], entering state (A). 3. At the end of cleaning up process, `marktree_del_itr` gets called. This function is supposed to remove given node and, if necessary, restructure the tree. Also, move the iterator to the next node. The bug occurs in this function. 4. The iterator goes backwards to the node's last child, to put it in the place of its deleted parent later. (B) 5. The parent node is deleted and replaced with its child node. (C) 6. Since now this node has 8 children, which is less than `MT_BRANCH_FACTOR - 1`, it get's merged with the next node. (D) 7. Finally, since at (B) the iterator went backward, it goes forward twice, skipping [0, 11] node, causing this extmark to persist, causing the bug. (E) ANALYSIS AND SOLUTION The algorithm works perfectly when the parent node gets replaced by its child, but no merging occurs. I.e. the exact same diagram, but without the (D) stage. If not for (D), it would iterate to <0, 9> and then to [0, 11]. So, iterating twice makes sense. The actual problem is in (C) stage, because the iterator index isn't adjusted and still pointing to no longer existent node. So my solution is to adjust iterator index after removing the child node. More info: https://github.com/neovim/neovim/pull/14719 | |||
| * | clang/'Logic error': zero-init oldbase array | Jan Edmund Lazo | 2021-06-02 | |
| | | | | | unrelative() must not have garbage structs. | |||
| * | unused macros: Remove a five unused macros. (#13517) | Edwin Pujols | 2020-12-11 | |
| | | | | | | | | | | | Removes: - GET_CHARTAB - PRT_PS_DEFAULT_BUFFER_SIZE - key_t - PROP_MASK - SCTX_INIT Referring #13505. | |||
| * | pvs/v547: comment out code for 'adjustment == 1' | Jan Edmund Lazo | 2020-11-08 | |
| | | | | | | | | That code never runs. Comment it out in case that it is required for debugging marktree.c. Use 'const' to mark constant variables. | |||
| * | 'clang/Logic error': zero-init MarkTreeIter vars | Jan Edmund Lazo | 2020-03-01 | |
| | | ||||
| * | clang/scan-build: restore required code | Hirokazu Hata | 2020-02-23 | |
| | | | | | ref: https://github.com/neovim/neovim/pull/11900#discussion_r381860165 | |||
| * | clang/scan-build: fix dead stores #11900 | Hirokazu Hata | 2020-02-18 | |
| | | ||||
| * | extmarks/bufhl: reimplement using new marktree data structure | Björn Linse | 2020-01-16 | |
| | | | | | | | | | | | | | | | | | | | | | Add new "splice" interface for tracking buffer changes at the byte level. This will later be reused for byte-resolution buffer updates. (Implementation has been started, but using undocumented "_on_bytes" option now as interface hasn't been finalized). Use this interface to improve many edge cases of extmark adjustment. Changed tests indicate previously incorrect behavior. Adding tests for more edge cases will be follow-up work (overlaps on_bytes tests) Don't consider creation/deletion of marks an undoable event by itself. This behavior was never documented, and imposes complexity for little gain. Add nvim__buf_add_decoration temporary API for direct access to the new implementation. This should be refactored into a proper API for decorations, probably involving a huge dict. fixes #11598 | |||
| * | Add new marktree data structure for storing marks | Björn Linse | 2020-01-14 | |
| This is inspired by Atom's "marker index" data structure to efficiently adjust marks to text insertions deletions, but uses a wide B-tree (derived from kbtree) to keep the nesting level down. | ||||
 |
index : rneovim.git | |
| Neovim fork with Rahm's personal hacks. |
| aboutsummaryrefslogtreecommitdiff |