- Wine-GitLab - list.winehq.org

[PATCH 0/1] MR10532: crypt32: Implement PFXExportCertStoreEx support
by Benoît Legat (＠blegat) April 1, 2026

April 1, 2026

Part 2 of my attempt to run the [Niko Home Control programming software](https://appdb.winehq.org/objectManager.php?sClass=application&iId… under Wine. This .NET application ends up calling \``PFXExportCertStoreEx` (or `PFXExportCertStore` ) since it redirects to it which is a stub at the moment. According to the [documentation](https://learn.microsoft.com/en-us/windows/win32/api/wincrypt…, if the `EXPORT_PRIVATE_KEYS` flag is set, we should export the private key as well. For this, we use `NCryptExportKey` which is part of `ncrypt` so we have to add `ncrypt` to the `Makefile.in` as well. For exporting, I used GnuTLS library. As it seems to be only available in the unix side, I added a helper in `unixlib.c`. -- https://gitlab.winehq.org/wine/wine/-/merge_requests/10532

2 1

Re: [PATCH v9 0/2] MR10460: msvcrt: Add support for %Z printf specifier.
by Trung Nguyen (＠trungnt2910) April 1, 2026

April 1, 2026

On Wed Apr 1 06:21:04 2026 +0000, Piotr Caban wrote: > I think that it's not supported by MSVC. Thanks for the suggestions. For the first point about portability, did you mean the syntax of compound literals (a C99 [standard feature](https://en.cppreference.com/w/c/language/compound_literal.html)) or the `UNICODE_STRING` struct itself? For the syntax, it seems that MSVC does support it: https://godbolt.org/z/KxPGahc7z I chose the earliest MSVC version available there. -- https://gitlab.winehq.org/wine/wine/-/merge_requests/10460#note_134633

1 0

[PATCH v6 0/7] MR10515: vbscript: Bind local variables, class properties, and loop variables at compile time.
by Francis De Brabandere (＠francisdb) April 1, 2026

April 1, 2026

**New opcodes:** - `OP_local` / `OP_assign_local` / `OP_set_local` read/write local vars and args by index - `OP_local_prop` / `OP_assign_local_prop` / `OP_set_local_prop` read/write class properties by index - `OP_step_local` / `OP_incc_local` For-loop step/increment with indexed counter - `OP_enumnext_local` For-Each iteration with indexed loop variable This mirrors jscript's `OP_local` / `bind_local()` optimization (`dlls/jscript/compile.c`) and addresses the FIXME in `lookup_identifier()` noting that class property access should be bound at compile time. Names that cannot be resolved at compile time (globals in Execute/ExecuteGlobal code, dynamic vars, host objects) continue to use the existing string-based path. ### Local variable / argument / class property access (1M iterations, 10 reads per iteration) | Benchmark | Windows (VM) | Wine master | Wine (6 commits) | **Speedup** | vs Windows | |-----------|-------------|-------------|-------------------|---------|------------| | Local vars | 156 ms | 10,134 ms | 964 ms | **10.5×** | 6.2× slower | | Arguments | 171 ms | 6,656 ms | 968 ms | **6.9×** | 5.7× slower | | Class props | 265 ms | 6,673 ms | 957 ms | **7.0×** | 3.6× slower | ### For-loop variants (10M iterations) | Benchmark | Windows (VM) | Wine master | Wine (6 commits) | **Speedup** | vs Windows | |-----------|-------------|-------------|-------------------|---------|------------| | Empty For | 117 ms | 6,222 ms | 1,535 ms | **4.1×** | 13× slower | | For + assign | 203 ms | 10,226 ms | 1,734 ms | **5.9×** | 8.5× slower | | Do-While manual | 437 ms | 8,773 ms | 1,828 ms | **4.8×** | 4.2× slower | | For Step 2 | 62 ms | 6,125 ms | 1,542 ms | **4.0×** | 25× slower | | Nested For | 125 ms | 7,742 ms | 1,554 ms | **5.0×** | 12× slower | | For local Sub | 109 ms | 3,062 ms | 1,542 ms | **2.0×** | 14× slower | | For R8 counter | 203 ms | 10,882 ms | 1,484 ms | **7.3×** | 7.3× slower | | For-Each array | 7 ms | 6,339 ms | **160 ms** | **40×** | 23× slower | The remaining ~4–13× gap vs Windows in For-loops is from `VarAdd`/`VarCmp` going through full `VariantChangeTypeEx` machinery (including locale alloc/free per call). A separate MR ([!10529](https://gitlab.winehq.org/wine/wine/-/merge_requests/10529)) adds I2/I4 fast paths in oleaut32 that bring integer loops within 1.2–1.9× of Windows when combined with this branch. -- v6: vbscript: Fast-path For-Each iteration over SAFEARRAY. https://gitlab.winehq.org/wine/wine/-/merge_requests/10515

2 7

[PATCH v5 0/7] MR10515: vbscript: Bind local variables, class properties, and loop variables at compile time.
by Francis De Brabandere (＠francisdb) April 1, 2026

April 1, 2026

**New opcodes:** - `OP_local` / `OP_assign_local` / `OP_set_local` read/write local vars and args by index - `OP_local_prop` / `OP_assign_local_prop` / `OP_set_local_prop` read/write class properties by index - `OP_step_local` / `OP_incc_local` For-loop step/increment with indexed counter - `OP_enumnext_local` For-Each iteration with indexed loop variable This mirrors jscript's `OP_local` / `bind_local()` optimization (`dlls/jscript/compile.c`) and addresses the FIXME in `lookup_identifier()` noting that class property access should be bound at compile time. Names that cannot be resolved at compile time (globals in Execute/ExecuteGlobal code, dynamic vars, host objects) continue to use the existing string-based path. ### Local variable / argument / class property access (1M iterations, 10 reads per iteration) | Benchmark | Windows (VM) | Wine master | Wine (6 commits) | **Speedup** | vs Windows | |-----------|-------------|-------------|-------------------|---------|------------| | Local vars | 156 ms | 10,134 ms | 964 ms | **10.5×** | 6.2× slower | | Arguments | 171 ms | 6,656 ms | 968 ms | **6.9×** | 5.7× slower | | Class props | 265 ms | 6,673 ms | 957 ms | **7.0×** | 3.6× slower | ### For-loop variants (10M iterations) | Benchmark | Windows (VM) | Wine master | Wine (6 commits) | **Speedup** | vs Windows | |-----------|-------------|-------------|-------------------|---------|------------| | Empty For | 117 ms | 6,222 ms | 1,535 ms | **4.1×** | 13× slower | | For + assign | 203 ms | 10,226 ms | 1,734 ms | **5.9×** | 8.5× slower | | Do-While manual | 437 ms | 8,773 ms | 1,828 ms | **4.8×** | 4.2× slower | | For Step 2 | 62 ms | 6,125 ms | 1,542 ms | **4.0×** | 25× slower | | Nested For | 125 ms | 7,742 ms | 1,554 ms | **5.0×** | 12× slower | | For local Sub | 109 ms | 3,062 ms | 1,542 ms | **2.0×** | 14× slower | | For R8 counter | 203 ms | 10,882 ms | 1,484 ms | **7.3×** | 7.3× slower | | For-Each array | 7 ms | 6,339 ms | **160 ms** | **40×** | 23× slower | The remaining ~4–13× gap vs Windows in For-loops is from `VarAdd`/`VarCmp` going through full `VariantChangeTypeEx` machinery (including locale alloc/free per call). A separate MR ([!10529](https://gitlab.winehq.org/wine/wine/-/merge_requests/10529)) adds I2/I4 fast paths in oleaut32 that bring integer loops within 1.2–1.9× of Windows when combined with this branch. -- v5: vbscript: Fast-path For-Each iteration over SAFEARRAY. https://gitlab.winehq.org/wine/wine/-/merge_requests/10515

2 7

[PATCH 0/1] MR10530: crypt32: Add CERT_STORE_NO_CRYPT_RELEASE_FLAG to CertOpenStore() flags when called from CertOpenSystemStore().
by Dmitry Timoshkov (＠dmitry) April 1, 2026

April 1, 2026

-- https://gitlab.winehq.org/wine/wine/-/merge_requests/10530

3 2

[PATCH 0/1] MR10529: crypt32: Also accept CERT_CHAIN_ENGINE_CONFIG without dwExclusiveFlags.
by Hans Leidekker (＠hans) April 1, 2026

April 1, 2026

-- https://gitlab.winehq.org/wine/wine/-/merge_requests/10529

2 1

[PATCH v2 0/1] MR10528: oleaut32: Add fast paths for VarAdd and VarCmp with integer operands.
by Francis De Brabandere (＠francisdb) April 1, 2026

April 1, 2026

Add fast paths for `VarAdd` and `VarCmp` when both operands are simple integer types (VT_I2, VT_I4, VT_EMPTY), skipping the expensive `VariantChangeType`/`VariantCopyInd` machinery. Profiling VBScript For-loops showed that `VarAdd`/`VarCmp` are called every iteration for the loop counter, and the full VARIANT conversion pipeline (including locale allocation in ucrtbase) was the dominant cost. With this patch, a VBScript For-loop inside a Sub goes from 3,041ms to 1,712ms for 10M iterations (**1.8× faster**). Benchmark results (VBScript `cscript`, 10M iterations unless noted): | Benchmark | Before | After | Speedup | |-----------|--------|-------|---------| | For-loop in Sub | 3,041 ms | 1,712 ms | **1.8×** | | Empty For (global) | 6,125 ms | 4,792 ms | **1.3×** | | For Step 2 (5M) | 3,058 ms | 2,404 ms | **1.3×** | | For R8 counter | 10,824 ms | 10,878 ms | ~same (R8 not fast-pathed) | Global-scope loops show smaller gains because VBScript's string-based identifier lookup still dominates there (see !10515). Combined with !10515, integer For-loops reach 1.2–1.5× of native Windows performance (up to 51× faster than current master). -- v2: oleaut32: Add fast paths for VarAdd and VarCmp with integer operands. https://gitlab.winehq.org/wine/wine/-/merge_requests/10528

2 1

[PATCH 0/1] MR10528: oleaut32: Add fast paths for VarAdd and VarCmp with integer operands.
by Francis De Brabandere (＠francisdb) April 1, 2026

April 1, 2026

Skip the expensive VariantChangeType/VariantCopyInd machinery when both operands are simple integer types (VT_I2, VT_I4, VT_EMPTY). For VarAdd, handle overflow promotion (I2+I2 that overflows to I4, I4+I4 that overflows to R8) matching native Windows behavior. For VarCmp, compare via LONGLONG to handle mixed I2/I4 correctly. This eliminates the dominant cost in VBScript For-loop step/compare operations, where VarAdd and VarCmp are called every iteration with trivial I2 operands but previously went through the full VARIANT conversion pipeline including locale allocation in ucrtbase. -- https://gitlab.winehq.org/wine/wine/-/merge_requests/10528

2 1

Re: [PATCH v9 0/2] MR10460: msvcrt: Add support for %Z printf specifier.
by Piotr Caban (＠piotr) April 1, 2026

April 1, 2026

On Tue Mar 31 18:55:56 2026 +0000, Elizabeth Figura wrote: > I believe we have some of those structures elsewhere already. Is there a > known compiler that doesn't support them? I think that it's not supported by MSVC. -- https://gitlab.winehq.org/wine/wine/-/merge_requests/10460#note_134611

1 0

[PATCH v3 0/5] MR10515: vbscript: Bind local variables, arguments, and class properties at compile time.
by Francis De Brabandere (＠francisdb) April 1, 2026

April 1, 2026

Resolve identifiers to direct indices during compilation instead of doing runtime wcsicmp string scans in lookup_identifier(). This adds new opcodes (OP_local, OP_assign_local, OP_set_local, OP_step_local, OP_incc_local, OP_local_prop, OP_assign_local_prop, OP_set_local_prop) that access local variables, function arguments, and class properties by index, eliminating the linear string comparison overhead for the most common cases. This mirrors jscript's OP_local / bind_local() optimization (dlls/jscript/compile.c) and addresses the FIXME in lookup_identifier() noting that class property access should be bound at compile time. Names that cannot be resolved at compile time (globals, dynamic vars, host objects) continue to use the existing string-based path. It's still 10x slower compared to windows but I have another related MR coming for that and it's related to the loop Windows ``` Local vars: 156.3 ms Arguments: 156.3 ms Class props: 390.6 ms Iterations: 1,000,000 ``` | Phase | Local vars | Arguments | Class props | |-------|-----------|-----------|-------------| | Before (baseline) | 10,200 ms | 6,700 ms | 6,710 ms | | `OP_local` | 5,130 ms | 2,035 ms | 5,920 ms | 2.0× / 3.3× | | `OP_assign_local` | 1,880 ms | 1,125 ms | 5,000 ms | 5.4× / 6.0× | | `OP_step_local`, `OP_incc_local` | 960 ms | 968 ms | 4,830 ms | **10.6× / 6.9× | | `OP_local_prop`, `OP_assign_local_prop` | 968 ms | 972 ms | 968 ms | **10.5× / 6.9× | -- v3: vbscript: Bind global-scope Dim variables at compile time. https://gitlab.winehq.org/wine/wine/-/merge_requests/10515

2 5