When we convert a Lua table to an Object, we consider the table a
"dictionary" if it contains only string keys, and an array if it
contains all numeric indices with no gaps. While rare, Lua tables can
have both strictly numeric indices and gaps (e.g. { [2] = 2 }). These
currently cannot be serialized because it is not considered an array.
However, we know the maximum index of the table and as long as all of
the keys in the table are numeric, it is still possible to serialize
this table as an array. The missing indices will have nil values.
Problem:
When `vim._foldupdate()` is invoked inside a scheduled callback, the
cursor may have moved to a line with a closed fold, e.g., after `dd` on
the line that is one line above a folded region. Then it opens the fold,
which is unnecessary and distracting. Legacy foldexprs do not have this
issue.
Solution:
Don't explicitly open folds on cursor.
Note:
This doesn't completely prevent spurious opening of folds. That is due
to bugs in treesitter foldexpr algorithm, which should be addressed
separately.
Problem: Confusing error for missing key.
Solution: Use the actualy key for the error. (closesvim/vim#9241)
5c1ec439f0
Co-authored-by: Bram Moolenaar <Bram@vim.org>
FUNC_ATTR_* should only be used in .c files with generated headers.
Defining FUNC_ATTR_* as empty in headers causes misuses of them to be
silently ignored. Instead don't define them by default, and only define
them as empty after a .c file has included its generated header.
Problem: buffer text with composing chars are converted from UTF-8
to an array of up to seven UTF-32 values and then converted back
to UTF-8 strings.
Solution: Convert buffer text directly to UTF-8 based schar_T values.
The limit of the text size is now in schar_T bytes, which is currently
31+1 but easily could be raised as it no longer multiplies the size
of the entire screen grid when not used, the full size is only required
for temporary scratch buffers.
Also does some general cleanup to win_line text handling, which was
unnecessarily complicated due to multibyte rendering being an "opt-in"
feature long ago. Nowadays, a char is just a char, regardless if it consists
of one ASCII byte or multiple bytes.
If Nvim is in the process of exiting then we do not want to allocate any
new refs onto the event loop, because they will not be freed and will
result in a memory leak.
We already have an extensive suite of static analysis tools we use,
which causes a fair bit of redundancy as we get duplicate warnings. PVS
is also prone to give false warnings which creates a lot of work to
identify and disable.
long is 32 bits on windows, while it is 64 bits on other architectures.
This makes the type suboptimal for a codebase meant to be
cross-platform. Replace it with more appropriate integer types.
Problem:
The next command after `silent !{cmd}` or `silent lua print('str')`
prints an empty line before printing a message, because these commands
set `msg_didout = true` despite not printing any messages.
Solution:
Set `msg_didout = true` only if `msg_silent == 0`
long is 32 bits on windows, while it is 64 bits on other architectures.
This makes the type suboptimal for a codebase meant to be
cross-platform. Replace it with more appropriate integer types.
- Move vimoption_T to option.h
- option_defs.h is for option-related types
- option_vars.h corresponds to Vim's option.h
- option_defs.h and option_vars.h don't include each other
This involves two redesigns of the map.c implementations:
1. Change of macro style and code organization
The old khash.h and map.c implementation used huge #define blocks with a
lot of backslash line continuations.
This instead uses the "implementation file" .c.h pattern. Such a file is
meant to be included multiple times, with different macros set prior to
inclusion as parameters. we already use this pattern e.g. for
eval/typval_encode.c.h to implement different typval encoders reusing a
similar structure.
We can structure this code into two parts. one that only depends on key
type and is enough to implement sets, and one which depends on both key
and value to implement maps (as a wrapper around sets, with an added
value[] array)
2. Separate the main hash buckets from the key / value arrays
Change the hack buckets to only contain an index into separate key /
value arrays
This is a common pattern in modern, state of the art hashmap
implementations. Even though this leads to one more allocated array, it
is this often is a net reduction of memory consumption. Consider
key+value consuming at least 12 bytes per pair. On average, we will have
twice as many buckets per item.
Thus old implementation:
2*12 = 24 bytes per item
New implementation
1*12 + 2*4 = 20 bytes per item
And the difference gets bigger with larger items.
One might think we have pulled a fast one here, as wouldn't the average size of
the new key/value arrays be 1.5 slots per items due to amortized grows?
But remember, these arrays are fully dense, and thus the accessed memory,
measured in _cache lines_, the unit which actually matters, will be the
fully used memory but just rounded up to the nearest cache line
boundary.
This has some other interesting properties, such as an insert-only
set/map will be fully ordered by insert only. Preserving this ordering
in face of deletions is more tricky tho. As we currently don't use
ordered maps, the "delete" operation maintains compactness of the item
arrays in the simplest way by breaking the ordering. It would be
possible to implement an order-preserving delete although at some cost,
like allowing the items array to become non-dense until the next rehash.
Finally, in face of these two major changes, all code used in khash.h
has been integrated into map.c and friends. Given the heavy edits it
makes no sense to "layer" the code into a vendored and a wrapper part.
Rather, the layered cake follows the specialization depth: code shared
for all maps, code specialized to a key type (and its equivalence
relation), and finally code specialized to value+key type.
Downloading the necessary files all at once instead of doing dependency
handling with luarocks speeds up installation immensely. We speed up the
process even more by using luv as a replacement for the C modules in the
busted dependencies, which allows us to skip costly compilation times.
Co-authored-by: bfredl <bjorn.linse@gmail.com>
Unfortunately the gc=false objects can refer to a dangling tree if the
gc=true tree was freed first. This reuses the same tree object as the
node itself is keeping alive via the uservalue of the node userdata.
(wrapped in a table due to lua 5.1 restrictions)
Problem: `push_tree`, every time its called for the same TSTree with
`do_copy=false` argument, creates a new userdata for it. Each userdata,
when garbage collected, frees the same TSTree C object.
Solution: Add flag to userdata, which indicates, should C object,
which userdata points to, be freed, when userdata is garbage collected.
