jumplist: avoid extra jumplist tail entry #9805

abdelhakeem · 2019-03-28T16:10:14Z

setpcmark() should not add pcmark to jumplist if jumplist tail is the same.

src/nvim/mark.c

justinmk · 2019-03-28T20:12:46Z

Thank you @abdelhakeem , I'm quite interested in this feature myself. Can you add a basic test? I would create a new test file test/functional/normal/jump_spec.lua.

See test/functional/example_spec.lua to get started.

justinmk · 2019-03-30T11:16:48Z

If we change it to always avoid adding a duplicate consecutive jumplist item, does that avoid the setpcmark_ext extra function?

abdelhakeem · 2019-03-30T12:05:07Z

If we change it to always avoid adding a duplicate consecutive jumplist item, does that avoid the setpcmark_ext extra function?

Actually I misunderstood the issue at first, hence my wording led to misunderstanding...
The issue we're having here is not about duplicate consecutive jumplist items (that might be another issue though), it's the call to setpcmark that is not followed by a jump. setpcmarkadds the current position to the jumplist by default (not duplicated), with the assumption that anyone would call it before an actual jump, so it pushes the current position. do_argfile just wants to update the current position mark when it calls setmark('\''), it shouldn't update the jumplist in this case then.

Remember the jumplist in the :drop issue looks like this:

 jump line  col file/text
   2     1    0 file1
   1     1    0 file2 contents
>

No duplicate entries, just that after adding 1 1 0 file2 contents we don't jump as assumed by setpcmark.

abdelhakeem · 2019-03-30T17:54:34Z

Related: #9775 (comment)

test/functional/normal/jump_spec.lua

justinmk · 2019-03-30T18:52:28Z

@abdelhakeem Thanks, this is clear now!

Do you want to also fix main.c in this PR, as mentioned in #9775 (comment) ? Testing it is easy:

local funcs = helpers.funcs
local eq = helpers.eq
it('screen test', function()
  clear()
  eq('\n jump line  col file/text\n>', funcs.execute('jumps'))
end)

abdelhakeem · 2019-03-30T20:50:39Z

The fix in main "breaks" two tests which expect the buggy behavior:

[  FAILED  ] 2 tests, listed below:
[  FAILED  ] .../abdelhakeem/neovim/test/functional/shada/marks_spec.lua @ 143: ShaDa support code is able to dump and restore jump list
.../abdelhakeem/neovim/test/functional/shada/marks_spec.lua:155: Expected objects to be the same.
Passed in:
(string) '
 jump line  col file/text
   4     2    0 Xtestfile-functional-shada-marks-2
   3     1    0 Xtestfile-functional-shada-marks-2
   2     2    0 Xtestfile-functional-shada-marks
   1     1    0 Xtestfile-functional-shada-marks
>'
Expected:
(string) '
 jump line  col file/text
   5     2    0 Xtestfile-functional-shada-marks-2
   4     1    0 Xtestfile-functional-shada-marks-2
   3     2    0 Xtestfile-functional-shada-marks
   2     1    0 Xtestfile-functional-shada-marks
   1     1    0 
>'

stack traceback:
	.../abdelhakeem/neovim/test/functional/shada/marks_spec.lua:155: in function <.../abdelhakeem/neovim/test/functional/shada/marks_spec.lua:143>

[  FAILED  ] ...bdelhakeem/neovim/test/functional/shada/merging_spec.lua @ 903: ShaDa jumps support code merges jumps when reading
...bdelhakeem/neovim/test/functional/shada/merging_spec.lua:913: Expected objects to be the same.
Passed in:
(string) '
 jump line  col file/text
   5     2    0 /a/b/c
   4     2    0 /a/b/d
   3     3    0 /a/b/d
   2     2    0 /a/b/e
   1     2    0 /a/b/f
>'
Expected:
(string) '
 jump line  col file/text
   6     2    0 /a/b/c
   5     2    0 /a/b/d
   4     3    0 /a/b/d
   3     2    0 /a/b/e
   2     2    0 /a/b/f
   1     1    0 
>'

justinmk · 2019-03-30T21:22:02Z

The fix in main "breaks" two tests which expect the buggy behavior:

@abdelhakeem that's ok, those tests merely assert that we match Vim behavior. We are intentionally deviating now, so the tests can just be updated (just delete the lines in the test).

abdelhakeem · 2019-03-31T22:19:21Z

@justinmk
It keeps popping up in different places as expected, this setpcmark_ext approach isn't reliable.
I found out that jumplist navigation commands call cleanup_jumplist (in movemark) before changing position, which already removes duplicates but it does not remove this redundant entry at the tail, so I added the following to cleanup_jumplist instead of the original fix:

--- a/src/nvim/mark.c
+++ b/src/nvim/mark.c
@@ -1191,6 +1191,18 @@ void cleanup_jumplist(void)
   if (curwin->w_jumplistidx == curwin->w_jumplistlen)
     curwin->w_jumplistidx = to;
   curwin->w_jumplistlen = to;
+
+  if (curwin->w_jumplistlen
+      && curwin->w_jumplistidx == curwin->w_jumplistlen) {
+    // When pointer is below last jump, remove it if it
+    // matches current position.
+    const xfmark_T *fm_last = &curwin->w_jumplist[curwin->w_jumplistlen - 1];
+    if (fm_last->fmark.fnum == curbuf->b_fnum
+        && equalpos(fm_last->fmark.mark, curwin->w_cursor)) {
+      curwin->w_jumplistlen--;
+      curwin->w_jumplistidx--;
+    }
+  }
 }

It (hopefully) solved all instances of the problem, but the tests are weirdly always hanging here with this change though:

[----------] Running tests from /home/abdelhakeem/neovim/test/functional/core/fileio_spec.lua
[ RUN      ] fileio fsync() codepaths #8304: 767.11 ms OK
█

Any ideas?

EDIT: Got it. ShaDa calls cleanup_jumplist just after setpcmark on exit, so with no jumplist entries besides the one matching current position this cleans this entry and ShaDa writing expects at least one entry. Just reversed their order.

test/functional/normal/jump_spec.lua

fixes neovim#9775

justinmk · 2019-04-01T22:51:11Z

@abdelhakeem so far as I've tested it, this is a really nice change that fixes a very old Vim issue. Nice work!

This release represents ~2700 commits since v0.3.4, the previous non-maintenance release. Besides the highlights listed below, this release features vast improvements to documentation, internal subsystems and test/CI infrastructure, and 700+ patches merged from Vim. FEATURES: New API functions: nvim_create_buf: create various kinds of buffers nvim_get_context, nvim_load_context 8e6b0a7 #10619 API: Context: save/restore/inspect editor state nvim_input_mouse: perform mouse actions nvim_open_win: create floating windows (and external, for supporting UIs) nvim_paste: paste text at cursor nvim_put: put text at cursor nvim_select_popupmenu_item: perform popupmenu actions nvim_set_keymap: create/delete mappings nvim_set_vvar: set v: variables nvim_ui_pum_set_height nvim_ui_try_resize_grid nvim_win_close: close windows nvim_win_get_config: get window configuration nvim_win_set_config: reconfigure windows New UI events: redraw.grid_destroy redraw.hl_group_set 8a3f858 #10504 UI/highlight: expose builtin highlight groups using hl_group_set event redraw.msg_clear redraw.msg_history_show redraw.msg_ruler redraw.msg_set_pos redraw.msg_show redraw.msg_showcmd redraw.msg_showmode redraw.win_close redraw.win_external_pos redraw.win_float_pos redraw.win_hide redraw.win_pos API f5c56f0 #9170 API/Lua: nvim_buf_attach: support Lua callback 82d48c0 #9896 API: emit nvim_error_event on failed async request b9ad12e #9992 UI/nvim_ui_attach(): add `override` option 3d1ed7c #9993 UI/ext_messages: learn more message kinds 8ed54bb #9547 proper multiline error message for rpcrequest, API wrappers Lua This release introduces "Nvim-Lua standard library". See ":help lua-stdlib". 89d7e24 #9463 Lua stdlib: vim.inspect, string functions 8e941c5 #9740 Lua: generate documentation from docstrings 1cbe014 #9301 lua/stdlib: Introduce vim.shared c83926c #10123 Lua: introduce vim.loop (expose libuv event-loop) 81e1dbc #10120 Lua: vim.schedule(cb) 1f54f68 #10688 Lua: minimal UTF-16 support needed for LSP 6fb0020 #10513 Lua encoding support c0993ed Lua: support getting UTF-32 and UTF-16 sizes of replaced text b0e2619 Lua: add {old_byte_size} to on_lines buffer change event UI: - The Nvim 0.3.4 UI protocol introduced line-based updates instead of legacy char-based updates. Nvim 0.4 continues to evolve the UI protocol. See ":help ui". Legacy UI clients are supported. See ":help api-contract". 9a1675b #6619 Floating windows - Can be (re)positioned, anchored, external. - Are real windows showing real buffers. No shortcuts, hacks, or compromises. - Support all features and API of normal windows, plus more. 6427894 #8455 Multigrid: "windows drawn on separate grids" - Windows are logically isolated internally. - Windows are sent to UIs as distinct objects, so that UIs can control layout instead of being stuck with the classic TUI layout. - Per-window font-size, dimenions, line-spacing. - Compositor: Internal subsystem for composing grids. 3855204 #6917 UIEnter, UILeave 788bcbb #9923 ui: ":syn blend=", 'winblend' 7cf7c0a #9575 ui: 'redrawdebug' option for flexible debugging of redrawing 5c836d2 #9607 wildoptions=pum (enabled by default) 37f8df8 #9571 UI: 'pumblend' option for semi-transparent popupmenu c403a95 #9446 Visual: highlight char-at-cursor - Traditionally Vim's visual selection does "reverse mode", which perhaps conflicts with the non-blinking block cursor. But 'guicursor' defaults to a vertical bar for selection=exclusive, and this confuses users who expect to see the text highlighted. :terminal fc27dc9 #8550 autocmds: TermEnter, TermLeave d13803f #9810 keymap, terminal: more keycodes 3b56f59 #9535 :terminal : Fix F1-F4 key codes 2d4a37e #10370 :ls : show "R", "F" for terminal-jobs fd0fd75 #9966 terminal: swap priority of terminal, editor highlights 7bb858c #9494 libvterm 0.1 TUI 3afb397 syntax, TUI: support "strikethrough" ccbcd39 #9408 TUI: "title stacking" unconditionally 298608f #9509 TUI: detect background color, set bg=dark/light 42f492a #9097 TUI: handle Smulx extension capability (extended underline) 424ddd0 #10205 TUI: support rgba background detection 9b43832 #9601 TUI: italics in tmux, Terminal.app f6fb370 #9793 keymap: support more (keypad) keycodes 3340e08 #9423 TUI: Konsole DECSCUSR fixup :checkhealth d0fd66b health/provider.vim: check curl HTTPS support c38862a #10490 checkhealth: try yarn if npm is missing 43356a4 #9929 health: check if tmux enabled true colors ec5a4d8 #9548 checkhealth: validate locale providers (clipboard, python, etc.): 96be8a2 #10161 Allow reloading providers (useful for UIs/clients) db3c797 #9487 provider: improve error message if provider is missing Various: 36762a0 #9295 signs: support multiple columns 801fe79 #10382 eval: wait() (wait for any condition) 9df3a67 #10400 MsgArea highlight; message grid a9bea8c #10790 keymap: allow modifiers to multibyte chars, like <m-ä> 25e0a44 #10878 #4448 paste: redesign (10x+ faster pasting; extensible vim.paste Lua hook) ef5037e #9706 autocmd: introduce "++once" feature 175398f #9616 add CompleteChanged autocmd 7fcf2f9 #9717 TextYankPost: add v:event["inclusive"] 3a699a7 #8364 termdebug.vim plugin ca1ce59 #9709 performance: use os_copy to create backups ed0e96c man.vim: set 'linebreak' 70f6939 #9564 events: add "Signal" event f89d0d8 #9568 inccommand: auto-disable if folding is slow FIXES: 41bb68b #10584 process_stop: uv: do not close stdin first/explicitly e50aa2a #10117 normal: Don't exit CTRL-O mode after processing K_EVENT 95fa71c #9504 :recover : Fix crash on non-existent *.swp 5a836d4 #9507 screen: don't unconditionally clear messages on window scroll 149dcbf #10021 channel: refactor events, prevent recursive invocation of events d19ff73 #10107 Fix multiple c_CTRL-D showing statusline b65a7b7 #10103 Fix wildmode=list,full and display+=msgsep interaction 0be6d3c #9634 fsync: Ignore ENOTSUP. Fix writing to SMB. b247c6f #10025 kbtree: pointer UB and unitialized value fixes 018e0d5 #9643 API/buffer-updates: always detach on buf-reload 400ee59 #9961 API: fix cursor position when lines are added 769f44e #9911 win/defaults: Use "…/nvim-data/site" in 'runtimepath' 83d5716 #9911 spellfile.vim: store files in stdpath('data') 8dbf231 #9887 RPC: conform message-id type to msgpack-RPC spec 5f996e3 #9894 options: properly reset directories on 'autochdir' 4c4a570 #9807 various CursorMoved fixes 943bedf #9853 event-loop: do not set CA_COMMAND_BUSY 9d207fd #9693 dictwatcheradd(): support b:changedtick 2d50bf3 #9789 mac: fix locale detection c563133 #9754 :mksession : restore tab-local working directories 092e7e6 #9703 #9703 executable(): return false if user is not owner 11a481f #9686 env var fixes/improvements 8e54847 #9666 #7920 os/env: Fix completion of multibyte env var names 5193826 #10468 Fix is_executable_in_path() on Windows 8eaa452 #9516 win: exepath(), executable() fixes f55c1e4 #10544 reltimefloat(): allow negative result b08dc3e #10561 win: jobstart(), system(): $PATHEXT-resolve exe 7cc2b72 #10392 TextYankPost: spurious/too-early dispatch during delete 6e01ed6 OpenBSD: stop jobs/processes properly 58dd5fc #10522 jobstop(): close channel before process_stop() 8363202 #10959 improved resize behavior (all UIs) c6eb1f4 #10830 API: fix nvim_command_output buffer overflow cbfd18c #10763 startup: handle 'guicursor' after user config b8f2436 #10915 jobwait(): fix race if job exits quickly 2fafed6 #10765 clipboard: handle/avoid SIGTERM with previous owner 8aca932 #9954 clipboard: setreg("*") with clipboard=unnamed 3f10c5b #9480 performance: clipboard/macOS: assume that pbcopy works 48efafc #10398 screen: disable redrawing inside VimResized 5e4b93a #10389 API/Lua: make nvim_execute_lua use native lua floats, not special tables 8c6f5b7 #9934 Spurious quote mark in command line when typing <C-R> a8a38f3 Lua 5.2/5.3 compat :terminal 47b7b47 #10700 :terminal : update buffer when switching tabpage 5225c1e #9605 terminal: Fix potential invalid local 'scrollback' 894f6be #8325 :terminal : set topline based on window height 8171e96 #9551 Improve :terminal resize d928b03 #9856 :stopinsert should leave terminal-mode 3f71218 #9926 :terminal : fix: Using `:stopinsert` while in normal mode 5020daa #9883 ui/terminal: make terminal state redraw like any other state TUI: 9f19e8d #9443 TUI: Do not disable BCE for builtin terminfos a4076e5 #9474 win/TUI: fix text overrides line numbers 533d4a3 #9645 TUI: do not resize host-terminal on startup b51e5d8 #9688 tui_tk_ti_getstr: handle weird value 1f5eac1 #10785 TUI: fix data-race during resize CHANGES: 9697c7f #8194 fix menu_get() 7f2e43c #9520 improve Lua error messages c234318 #9526 Remove jemalloc baf93d9 #9581 UI: always use concrete colors for default_colors_set 91688b4 #9563 defaults: set 'scrollback' to -1 by default bb24fec #10136 defaults: exclude "S" from 'shortmess' ddd0eb6 #8540 startup: -es/-Es (silent/batch mode): skip swapfile 3536249 #9805 jumplist: avoid extra tail entry 939d905 #10573 channels: reflect exit due to signals in exit status code 45c34bd #10689 :doautocmd : Never show "No matching autocommands" fb19aee #9110 API: make nvim_win_set_option() set window-global, not buffer-local abfc8b3 #10778 emsg_multiline: log Vim errors 06d9cc7 #10657 exists("$FOO"): return false for empty env var 6616d1d #10743 win/env: Vim-compat: Empty string deletes env var 7d66483 #10662 win: expand nested env var in $HOME 2816bc8 #8349 edit.c: Disable indent during completion 58f505d #9829 startup: remove TUI init special-case Historically Vim/Nvim does backflips to handle input and show messages before a UI is available. This logical contradiction was already fixed for remote UIs (#9024 c236e80). Fixing it also for the TUI avoids problems on Windows, simplifies the logic, and avoids races like #9959.

Commit 3536249 ("jumplist: avoid extra tail entry neovim#9805", 2019-04-02) introduced a regression where, when the * command leaves the cursor in the same line, the jumplist entry is erroneously erased. This breaks the mapping nnoremap * *<C-O> The cursor remains in the same line when 1. the cursor is outside of a word boundary; 2. the searched-for word occurs next time in the same line; or 3. the searched-for word is unique. Only erase the jumplist entry when both the cursor line and column position are unchanged. This still leaves case neovim#3 confusingly broken, probably somewhere between neovim#2 and neovim#1 in severity.

Commit 3536249 ("jumplist: avoid extra tail entry neovim#9805", 2019-04-02) introduced a regression where, when the * or # command leaves the cursor in the same line, the jumplist entry is erroneously erased. In addition to interactive use, this breaks mappings like nnoremap * *<C-O> The cursor remains in the same line when 1. the cursor is outside of a word boundary; 2. the searched-for word occurs next time in the same line; or 3. the searched-for word is unique. Only erase the jumplist entry when both the cursor line and column position are unchanged. This still leaves case neovim#3 confusingly broken but hopefully it's enough to salvage interactive use -- why "jump back" when the cursor doesn't move? -- even if it turns out to be insufficient for aforementioned mapping. The slightly more clumsy alternative mapping of nnoremap * *`` doesn't help: `` updates the jumplist, too, so unique keywords move the cursor to the position before the * command instead of staying in place. If that command is used on a unique keyword after being used on a non-unique keyword, that position is the "next match of the searched-for word" from the previous usage. ref neovim#9874

groves · 2022-03-29T22:11:52Z

This change means if getjumplist is called during BufHidden or BufLeave autocmds, the buffer being left isn't added to the jumplist.

I'm writing a plugin that uses the jumplist to show recently edited buffers in windows and added autocommands that were checking the jumplist as the user went between buffers. Was really confused as to why my jumplist stopped updating when I was editing new files.

I can't see a straightforward way to fix that bug while keeping this behavior. It seems like it'd be nice to only cleanup the jumplist on updating it rather than when reading it, but that seems like a challenging thing to do. Thought I'd toss this in here in case anyone else could think of a more straightforward fix.

abdelhakeem · 2022-04-04T02:14:52Z

This change means if getjumplist is called during BufHidden or BufLeave autocmds, the buffer being left isn't added to the jumplist.

I'm writing a plugin that uses the jumplist to show recently edited buffers in windows and added autocommands that were checking the jumplist as the user went between buffers. Was really confused as to why my jumplist stopped updating when I was editing new files.

I can't see a straightforward way to fix that bug while keeping this behavior. It seems like it'd be nice to only cleanup the jumplist on updating it rather than when reading it, but that seems like a challenging thing to do. Thought I'd toss this in here in case anyone else could think of a more straightforward fix.

A hacky workaround could be to prevent getjumplist() from cleaning up the jumplist tail in autocmd context as it would be cleaned up later when read or used anyway:

neovim/src/nvim/eval/funcs.c

Line 3514 in a973fa5

cleanup_jumplist(wp, true);

 cleanup_jumplist(wp, !autocmd_busy);

I'm not sure if this would introduce another unexpected behavior though.

mhinz reviewed Mar 28, 2019

View reviewed changes

src/nvim/mark.c Outdated Show resolved Hide resolved

mhinz reviewed Mar 28, 2019

View reviewed changes

src/nvim/mark.c Outdated Show resolved Hide resolved

abdelhakeem force-pushed the jumplist-fix branch from 9567daa to 8142bb7 Compare March 28, 2019 17:19

marvim added the RFC label Mar 28, 2019

justinmk added the enhancement label Mar 28, 2019

abdelhakeem changed the title ~~[RFC] fix: avoid duplicate adjacent jumplist marks~~ [RDY] fix: avoid duplicate jumps in do_argfile Mar 29, 2019

marvim added RDY and removed RFC labels Mar 29, 2019

abdelhakeem changed the title ~~[RDY] fix: avoid duplicate jumps in do_argfile~~ [WIP] fix: avoid duplicate jumps in do_argfile Mar 29, 2019

marvim added WIP and removed RDY labels Mar 29, 2019

abdelhakeem force-pushed the jumplist-fix branch from 34da210 to f831634 Compare March 30, 2019 14:27

abdelhakeem changed the title ~~[WIP] fix: avoid duplicate jumps in do_argfile~~ [RFC] fix: avoid duplicate jumps in do_argfile Mar 30, 2019

marvim added RFC and removed WIP labels Mar 30, 2019

justinmk reviewed Mar 30, 2019

View reviewed changes

test/functional/normal/jump_spec.lua Outdated Show resolved Hide resolved

justinmk reviewed Mar 30, 2019

View reviewed changes

test/functional/normal/jump_spec.lua Outdated Show resolved Hide resolved

justinmk reviewed Mar 30, 2019

View reviewed changes

test/functional/normal/jump_spec.lua Outdated Show resolved Hide resolved

abdelhakeem changed the title ~~[RFC] fix: avoid duplicate jumps in do_argfile~~ [RFC] fix: avoid extra jump in main and do_argfile Mar 30, 2019

abdelhakeem changed the title ~~[RFC] fix: avoid extra jump in main and do_argfile~~ [WIP] fix: avoid extra jump in main and do_argfile Mar 31, 2019

marvim added WIP and removed RFC labels Mar 31, 2019

abdelhakeem changed the title ~~[WIP] fix: avoid extra jump in main and do_argfile~~ [WIP] fix: avoid extra jumplist tail entry Apr 1, 2019

abdelhakeem force-pushed the jumplist-fix branch from f831634 to 7aa9c89 Compare April 1, 2019 15:37

abdelhakeem changed the title ~~[WIP] fix: avoid extra jumplist tail entry~~ [RFC] fix: avoid extra jumplist tail entry Apr 1, 2019

marvim added RFC and removed WIP labels Apr 1, 2019

justinmk reviewed Apr 1, 2019

View reviewed changes

test/functional/normal/jump_spec.lua Outdated Show resolved Hide resolved

jumplist: avoid extra tail entry

2a3581e

fixes neovim#9775

justinmk force-pushed the jumplist-fix branch from 7aa9c89 to 2a3581e Compare April 1, 2019 22:48

justinmk changed the title ~~[RFC] fix: avoid extra jumplist tail entry~~ jumplist: avoid extra jumplist tail entry Apr 1, 2019

justinmk merged commit 3536249 into neovim:master Apr 1, 2019

justinmk removed the RFC label Apr 1, 2019

blueyed mentioned this pull request Sep 18, 2019

m' doesn't save position to jump list #11054

Closed

commonquail mentioned this pull request Jul 8, 2020

jumplist: alleviate *, # command regression #12607

Closed

justinmk mentioned this pull request Oct 4, 2023

fix(jumplist): Ctrl+o, Ctrl+i behave unintuitively when deleting buffers #25461

Merged

Provide feedback

Saved searches

Use saved searches to filter your results more quickly

Uh oh!

Uh oh!

jumplist: avoid extra jumplist tail entry #9805

jumplist: avoid extra jumplist tail entry #9805

Uh oh!

abdelhakeem commented Mar 28, 2019

Uh oh!

Uh oh!

Uh oh!

justinmk commented Mar 28, 2019

Uh oh!

justinmk commented Mar 30, 2019

Uh oh!

abdelhakeem commented Mar 30, 2019 •

edited

Loading

Uh oh!

abdelhakeem commented Mar 30, 2019

Uh oh!

Uh oh!

Uh oh!

Uh oh!

justinmk commented Mar 30, 2019 •

edited

Loading

Uh oh!

abdelhakeem commented Mar 30, 2019

Uh oh!

justinmk commented Mar 30, 2019 •

edited

Loading

Uh oh!

abdelhakeem commented Mar 31, 2019 •

edited

Loading

Uh oh!

Uh oh!

justinmk commented Apr 1, 2019

Uh oh!

groves commented Mar 29, 2022

Uh oh!

abdelhakeem commented Apr 4, 2022 •

edited

Loading

Uh oh!

Uh oh!

Uh oh!

jumplist: avoid extra jumplist tail entry #9805

jumplist: avoid extra jumplist tail entry #9805

Uh oh!

Conversation

abdelhakeem commented Mar 28, 2019

Uh oh!

Uh oh!

Uh oh!

justinmk commented Mar 28, 2019

Uh oh!

justinmk commented Mar 30, 2019

Uh oh!

abdelhakeem commented Mar 30, 2019 • edited Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

Uh oh!

abdelhakeem commented Mar 30, 2019

Uh oh!

Uh oh!

Uh oh!

Uh oh!

justinmk commented Mar 30, 2019 • edited Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

Uh oh!

abdelhakeem commented Mar 30, 2019

Uh oh!

justinmk commented Mar 30, 2019 • edited Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

Uh oh!

abdelhakeem commented Mar 31, 2019 • edited Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

Uh oh!

Uh oh!

justinmk commented Apr 1, 2019

Uh oh!

groves commented Mar 29, 2022

Uh oh!

abdelhakeem commented Apr 4, 2022 • edited Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

Uh oh!

Uh oh!

abdelhakeem commented Mar 30, 2019 •

edited

Loading

justinmk commented Mar 30, 2019 •

edited

Loading

justinmk commented Mar 30, 2019 •

edited

Loading

abdelhakeem commented Mar 31, 2019 •

edited

Loading

abdelhakeem commented Apr 4, 2022 •

edited

Loading