hashes for refs should be an opt-in feature (#18098)

This commit is contained in:
Miran 2021-05-30 23:55:51 +02:00 • committed by GitHub
commit 50e98e6efa
No known key found for this signature in database
GPG key ID: 4AEE18F83AFDEB23
4 changed files with 16 additions and 5 deletions

View file

@ -50,8 +50,9 @@
- On POSIX systems, we now ignore `SIGPIPE` signals, use `-d:nimLegacySigpipeHandler` - On POSIX systems, we now ignore `SIGPIPE` signals, use `-d:nimLegacySigpipeHandler`
for previous behavior. for previous behavior.
- `hashes.hash` now supports `object` and `ref` (can be overloaded in user code). - `hashes.hash` can now support `object` and `ref` (can be overloaded in user code),
For a transition period, use `-d:nimLegacyNoHashRef` to avoid defining `hash(ref)`. if `-d:nimEnableHashRef` is used.
- `hashes.hash(proc|ptr|ref|pointer)` now calls `hash(int)` and honors `-d:nimIntHash1`, - `hashes.hash(proc|ptr|ref|pointer)` now calls `hash(int)` and honors `-d:nimIntHash1`,
`hashes.hash(closure)` has also been improved. `hashes.hash(closure)` has also been improved.
@ -82,7 +83,9 @@
other defines. other defines.
## Standard library additions and changes ## Standard library additions and changes
- Added support for parenthesized expressions in `strformat` - Added support for parenthesized expressions in `strformat`
- Fixed buffer overflow bugs in `net` - Fixed buffer overflow bugs in `net`
@ -99,7 +102,6 @@
the OpenSSL DLLs (e.g. libssl-1_1-x64.dll, libcrypto-1_1-x64.dll) you the OpenSSL DLLs (e.g. libssl-1_1-x64.dll, libcrypto-1_1-x64.dll) you
now also need to ship `cacert.pem` with your `.exe` file. now also need to ship `cacert.pem` with your `.exe` file.
- Make `{.requiresInit.}` pragma to work for `distinct` types. - Make `{.requiresInit.}` pragma to work for `distinct` types.
- Added a macros `enumLen` for returning the number of items in an enum to the - Added a macros `enumLen` for returning the number of items in an enum to the
@ -324,6 +326,8 @@
- Fixed premature garbage collection in asyncdispatch, when a stack trace override is in place. - Fixed premature garbage collection in asyncdispatch, when a stack trace override is in place.
## Language changes ## Language changes
- `nimscript` now handles `except Exception as e`. - `nimscript` now handles `except Exception as e`.
@ -368,6 +372,8 @@
- `typeof(voidStmt)` now works and returns `void`. - `typeof(voidStmt)` now works and returns `void`.
## Compiler changes ## Compiler changes
- Added `--declaredlocs` to show symbol declaration location in messages. - Added `--declaredlocs` to show symbol declaration location in messages.
@ -394,7 +400,8 @@
- TLS: OSX now uses native TLS (`--tlsEmulation:off`), TLS now works with importcpp non-POD types, - TLS: OSX now uses native TLS (`--tlsEmulation:off`), TLS now works with importcpp non-POD types,
such types must use `.cppNonPod` and `--tlsEmulation:off`should be used. such types must use `.cppNonPod` and `--tlsEmulation:off`should be used.
- Now array literals(JS backend) uses JS typed arrays when the corresponding js typed array exists, for example `[byte(1), 2, 3]` generates `new Uint8Array([1, 2, 3])`. - Now array literals(JS backend) uses JS typed arrays when the corresponding js typed array exists,
for example `[byte(1), 2, 3]` generates `new Uint8Array([1, 2, 3])`.
- docgen: rst files can now use single backticks instead of double backticks and correctly render - docgen: rst files can now use single backticks instead of double backticks and correctly render
in both rst2html (as before) as well as common tools rendering rst directly (e.g. github), by in both rst2html (as before) as well as common tools rendering rst directly (e.g. github), by
@ -426,6 +433,8 @@
the official Nim style guide. To be enabled, this has to be combined either the official Nim style guide. To be enabled, this has to be combined either
with `--styleCheck:error` or `--styleCheck:hint`. with `--styleCheck:error` or `--styleCheck:hint`.
## Tool changes ## Tool changes
- The rst parser now supports markdown table syntax. - The rst parser now supports markdown table syntax.

View file

@ -236,7 +236,7 @@ proc hash*[T](x: ptr[T]): Hash {.inline.} =
assert cast[pointer](a[0].addr).hash == a[0].addr.hash assert cast[pointer](a[0].addr).hash == a[0].addr.hash
hash(cast[pointer](x)) hash(cast[pointer](x))
when not defined(nimLegacyNoHashRef): when defined(nimEnableHashRef):
proc hash*[T](x: ref[T]): Hash {.inline.} = proc hash*[T](x: ref[T]): Hash {.inline.} =
## Efficient `hash` overload. ## Efficient `hash` overload.
runnableExamples: runnableExamples:

View file

@ -8,6 +8,7 @@ And we get here
''' '''
joinable: false joinable: false
targets: "c cpp js" targets: "c cpp js"
matrix: "-d:nimEnableHashRef"
""" """
# xxx wrap in a template to test in VM, see https://github.com/timotheecour/Nim/issues/534#issuecomment-769565033 # xxx wrap in a template to test in VM, see https://github.com/timotheecour/Nim/issues/534#issuecomment-769565033

View file

@ -1,5 +1,6 @@
discard """ discard """
targets: "c cpp js" targets: "c cpp js"
matrix: "-d:nimEnableHashRef"
""" """
import std/hashes import std/hashes