Commit graph

19 commits

Author SHA1 Message Date
Aleksey Kladov
da65cff18b Add one more test 2020-07-26 11:05:28 +02:00
Aleksey Kladov
39a2bc5e3c Expose package roots more directly 2020-07-21 12:52:51 +02:00
Aleksey Kladov
3e688d2a93 Simplify 2020-07-20 18:01:42 +02:00
Aleksey Kladov
46ac9ff5e3 Simplify exclusion logic 2020-07-18 16:40:10 +02:00
Aleksey Kladov
dc2094cfa5 Minor, push allocations down 2020-07-14 15:57:10 +02:00
Yuki Okushi
6f423466d1
Add a license field to all the crates 2020-07-14 10:57:26 +09:00
Aleksey Kladov
847135495f no doctests 2020-07-11 13:31:02 +02:00
Aleksey Kladov
a1ef6cc553 Optimize VFS processing 2020-07-10 22:30:24 +02:00
Aleksey Kladov
86bc4d20b3 Also reload when adding new examples, tests, etc 2020-07-10 22:30:24 +02:00
Aleksey Kladov
69b79e3a73 Replace ad hocery with science 2020-07-07 23:28:48 +02:00
Aleksey Kladov
5d2225f4bc Fix symbol search in salsa
Previous solution for binning paths into disjoint directories was
simple and fast -- just a single binary search.

Unfortunatelly, it wasn't coorrect: if the ditr are

  /d
  /d/a
  /d/c

then partitioning the file /d/b/lib.rs won't pick /d as a correct
directory.

The correct solution here is a trie, but it requires exposing path
components.

So, we use a poor man's substitution -- a *vector* of sorted paths,
such that each bucket is prefix-free

closes #5246
2020-07-07 17:38:02 +02:00
Aleksey Kladov
e336a96998 FileSetConfig works with empty set of roots
Closes #5139
2020-07-01 09:06:51 +02:00
Jonas Schievink
72fe70f2f8 Make VFS join methods fallible 2020-06-26 16:30:50 +02:00
Aleksey Kladov
6937bcb67d Less error-prone naming 2020-06-24 16:59:15 +02:00
Aleksey Kladov
e6c61d5072 Cleanup project.json deserialization 2020-06-24 16:16:52 +02:00
Laurențiu Nicola
c15c22139f Make Debug less verbose for VfsPath and use Display in analysis-stats 2020-06-24 14:07:03 +03:00
Aleksey Kladov
dad1333b48 New VFS 2020-06-23 17:51:06 +02:00
Aleksey Kladov
b9f3c5d585 Speedup VFS::partition
The task of `partition` function is to bin the flat list of paths into
disjoint filesets. Ideally, it should be incremental -- each new file
should be added to a specific fileset.

However, preliminary measurnments show that it is actually fast enough
if we just optimize this to use a binary search instead of a linear
scan.
2020-06-19 15:07:32 +02:00
Aleksey Kladov
c002322bde New VFS API 2020-06-16 13:42:29 +02:00