Problem: After ":cd" fails ":cd -" is incorrect.
Solution: Set the previous directory only after successfully changing
directory. (Richard Doty, closesvim/vim#9419, closesvim/vim#8983)
3d0abad5bf
Adjust the test's error message check due to missing patch
vim-patch:8.2.3973: tiny build fails
Problem: Tiny build fails.
Solution: Adjust #ifdefs
0f7a5e758c
vim-patch:8.2.3978: build error when using dynamycally loaded Python 3
Problem: Build error when using dynamycally loaded Python 3.
Solution: Adjust #ifdef.
6b1a99dfe3
vim-patch:8.2.4013: build failure without the spell feature
Problem: Build failure without the spell feature.
Solution: Adjust #ifdefs.
e60b3c47d7
vim-patch:8.2.4032: ATTRIBUTE_NORETURN is not needed
Problem: ATTRIBUTE_NORETURN is not needed.
Solution: Use NORETURN(). (Ozaki Kiichi, closesvim/vim#9487)
e12406526a
vim-patch:8.2.4048: gcc complains about use of "%p" in printf
Problem: gcc complains about use of "%p" in printf.
Solution: Add (void *) typecast. (Dominique Pellé, closesvim/vim#9494)
c14f667626
Problem: Command completion in cmdline window uses global user commands,
not local commands for the window where it was opened from.
Solution: Use local commands. (closesvim/vim#9168)
a119812437
These versions of python has reached End-of-life. getting rid
of python2 support removes a lot of logic to support two
incompatible python versions in the same version.
Problem: After restoring a session buffer order can be quite different.
Solution: Create buffers first. (Evgeni Chasnovski, closesvim/vim#9520)
26ebf1f036
---------------
As in Vim, this basically reverts 8.1.0829 providing different solution
(see vim/vim#9520).
Regarding Neovim, this basically reverts changes from #15062. Test about
restoring same terminals was a bit too restrictive with using actual
buffer ids, which changed with this patch (now they should be in the
same order as at `mksession` call), so I tweaked it.
Problem: No proper test for moving the window separator.
Solution: Add a test. Add comment in code. (closesvim/vim#9656)
a0c4e2f2d7
Remove the assertion as it is now possible for `fr` to be `NULL`.
The test fails without clearing messages. Not sure if this is a bug.
Problem: ":retab 0" may cause illegal memory access.
Solution: Limit the value of 'tabstop' to 10000.
652dee4486
ex_retab change is N/A (+vartabs always available).
Nvim's set_num_option validation logic was refactored, hence why it looks
different from Vim's.
Also use XFREE_CLEAR in other places.
Problem: Record buffer wrong if character in Select mode was not typed.
Solution: Only delete the tail from the record buffer if the character was
typed. (closesvim/vim#9650)
fbf4f1ca15
Problem: Crash when recording and using Select mode.
Solution: When deleting the last recorded character check there is something
to delete.
a4bc2dd7cc
Problem: With 'virtualedit' set to "block" block selection is wrong after
using "$". (Marco Trosi)
Solution: Compute the longest selected line. (closesvim/vim#8495)
b17ab86e7b
Problem: No proper test for maintaining change mark in diff mode.
Solution: Run the test with internal and external diff. (Sean Dewar,
closesvim/vim#9424)
ccc1644f95
Problem: Not all register related code is covered by tests.
Solution: Add more test cases. (Yegappan Lakshmanan, closesvim/vim#5301)
54c8d229f5
Can't be fully ported until "set clipboard=autoselect,autoselectplus" is
re-implemented for Test_clipboard_regs (and last visual selection to PRIMARY
selection works).
Problem: :lockmarks does not work for '[ and '].
Solution: save and restore '[ and '] marks. (James McCoy, closesvim/vim#5222)
f4a1d1c054
Test_diff_maintains_change_mark doesn't actually fail without these changes.
This is fixed in v8.2.3936.
Problem: output of ":scriptnames" goes into the message history, while this
des not happen for other commands, such as ":ls".
Solution: Use msg_outtrans() instead of smsg(). (closesvim/vim#9551)
840f16202e
