Previously when using system gd there was no way to use the Zend allocator methods to limit its memory usage. This PR fixes this by using the new memory allocator setters in libgd/libgd#692. Also, this PR gives the possibility to unbundle ext/gd/libgd.

At least the number of failures is going down each time a fix is pushed, I can wait and we are clearly closer to succeed.

Implement application-supplied memory allocators

Document missing types

Clear methods and testing reworked

Struct instead of lone function pointers

Fix gdmem malloc test

@cmb69, @pierre, @remicollet, @vapier I'm deeply sorry for being late with this PR, was occupied with things IRL and I didn't know how to fix this PR up until now.

So if anyone of you can merge this I'd be glad for it. :pray_tone1:

#788 fix bug in HEIF usage, stride is require (#801)

fix bug #788 in HEIF usage, stride is require

#792, document gdImageEllipse (#803)

#797 possible leak on png error, returns from setjmp lost the row_poi… (#802)

• #797 possible leak on png error, returns from setjmp lost the row_pointers ref somehow

• skip that one using mingw until we have a reliable portable FP related tests suites (may never happpen)

Create SECURITY.md (#775)

A simple instruction for security researchers.

Fix #808, When RAQM is used, and it fails or no text processed, the glyph may be leaked (#809)

Fix #806, getPixelInterpolateWeight, getPixelOverflowTC, getPixelOver… (#807)

• Fix #806, getPixelInterpolateWeight, getPixelOverflowTC, getPixelOverflowPalette, gdImageRotate* take bgcolor as GD's truecolor, result is a TC image as well.Clean dead code.

• Not needed at this stage

Fix #810, Wrong image freed, src_cloned should be freed if set. (#811)

Fix #808, When RAQM is used, and it fails or no text processed, the glyph may be leaked

Fix #808, improve fix here, at this stage info is not null (deref earlier too)

Fix #808, free info. Not sure what happens, it keeps being removed on cherry-pick to 2.3....

Fix #808, free info. Not sure what happens, it keeps being removed on cherry-pick to 2.3....

Fix #812, ensure operands priorty works as expected (#813)

Fix #815, gd_topal, explicit null dereferenced (kind of false positiv… (#816)

• Fix #815, gd_topal, explicit null dereferenced (kind of false positive however it does hurt to add that check)

• Fix #815, != NULL

sync Changelog

Fix tests based on coverity reports (#819)

• Partial #818, unused arg

• Partial #818, init var

• partail #818, fix va_args usage

• partail #818, handle f* calls and avoid possible call to malloc with negative values

• partail #818, prevent double free

• partail #818, resource leak if test fail

• partail #818, null deref fix

• partail #818, avoid double free on fp failure

• Partial #818, fix error msg

• Partial #818, leak on error

• Partial #818, null deref

• Partial #818, avoid possible negative index on failure

• partial #818, does not free if we return if requested new size overflow

• partial #818, avoid double free, free where the alloc happened

Bug/818 (#820)

• Partial #818, unused arg

• Partial #818, init var

• partail #818, fix va_args usage

• partail #818, handle f* calls and avoid possible call to malloc with negative values

• partail #818, prevent double free

• partail #818, resource leak if test fail

• partail #818, null deref fix

• partail #818, avoid double free on fp failure

• Partial #818, fix error msg

• Partial #818, leak on error

• Partial #818, null deref

• Partial #818, avoid possible negative index on failure

• partial #818, does not free if we return if requested new size overflow

• partial #818, avoid double free, free where the alloc happened

• partial #818, fix assert logic and test exp_size for <=0

• partial #818, fix assert logic for color idx test

Bug/818 (#821)

• Partial #818, unused arg

• Partial #818, init var

• partail #818, fix va_args usage

• partail #818, handle f* calls and avoid possible call to malloc with negative values

• partail #818, prevent double free

• partail #818, resource leak if test fail

• partail #818, null deref fix

• partail #818, avoid double free on fp failure

• Partial #818, fix error msg

• Partial #818, leak on error

• Partial #818, null deref

• Partial #818, avoid possible negative index on failure

• partial #818, does not free if we return if requested new size overflow

• partial #818, avoid double free, free where the alloc happened

• partial #818, fix assert logic and test exp_size for <=0

• partial #818, fix assert logic for color idx test

• partial #818, new case for possible leak, improve logic

• partial #818, fix assert logic for color idx test

partial #818, fix again that logic. I need to find something more generic and convenient to handle errors, assert and free used resources on fail assert

enable interlace transform when reading png (#823)

update changelog

chore: update browser-repl's webpack to 5.75.0

chore: update boxednode to 2.0.1

chore: update browser-repl's webpack to 5.75.0

chore: update boxednode to 2.0.1

There's mongodb-js/boxednode#41 that addresses an issue in the e2e test as it doesn't display a clear stack trace.

I've also raised browser-repl's webpack version to ^5.75.0 as that version processes node:module much more correctly.

Though I theorize that if all of devDependencies are hoisted then many of these CI failures will disappear.

fix: update browser-repl's webpack to 5.75.0

This fixes a TypeError that occurred in an e2e test in mongodb-js/mongosh#1427.

[Timed out 10000ms] expected prompt
+ expected - actual
-Current Mongosh Log ID:	640b3942e163d12c46b210ce
-Using Mongosh:		0.0.0-dev.0
-
-For mongosh info see: https://docs.mongodb.com/mongodb-shell/
-
-> {
-  key: Binary(Buffer.from("921e06bf637caa8a5940c418304b70c3f0243ec22b5cd9843314ebb37b54a134196d9bb89ac75151a47c5397d70fb03b417607704beebb859cb492e84fe36fbd5dde788837303f8af07f5247bf789bfb1a7e04477487f2e8f0feac5a6f9856d8", "hex"), 0)
-}
-> ... ... ... ... mongodb://localhost:30270/?directConnection=true&serverSelectionTimeoutMS=2000
-> test-1678457154415
-> ... ... ... ... ... ... ... ... ... ... ... 
-> mongodb://localhost:30270/?directConnection=true&serverSelectionTimeoutMS=2000
-> [prompt search starts here]node:mongosh/mongosh:40
-          Object.defineProperty(entry, 'getFileName', {
-                 ^
-
-TypeError: Cannot redefine property: getFileName
-    at Function.defineProperty (<anonymous>)
-    at node:mongosh/mongosh:40:18
-    at eval (eval at innerEval (/private/tmp/m/boxednode/mongosh/node-v16.19.1/out/Release/node:100:360919), <anonymous>:55:67)
-    at Array._arv8.map (eval at innerEval (/private/tmp/m/boxednode/mongosh/node-v16.19.1/out/Release/node:100:360919), <anonymous>:72:182)
-    at Function.epst (node:mongosh/mongosh:37:23)
-    at Function.epst (node:mongosh/mongosh:47:18)
-    at maybeOverridePrepareStackTrace (node:internal/errors:141:29)
-    at prepareStackTrace (node:internal/errors:115:5)
-    at beforeInspector (node:internal/errors:754:20)
-    at processPromiseRejections (node:internal/process/promises:279:13)
-    at processTicksAndRejections (node:internal/process/task_queues:97:32)
+/^([^<>]*> ?)+$/m

To alter a property of an Object using defineProperty said function must have the descriptor parameter with its configurable field set to true.

Hopefully this change will display a much more accurate log in that PR.

@YakoYakoYokuYoku Thank you! I’ve updated/removed the package-lock files here, that’s required in order for this to pass CI. We’ll see how that goes :slightly_smiling_face:

This surprises me, I thought that only package.json and lerna.json needed to be modified, but it seems that everything was moved to the root package-lock.json. Regardless, I have faith that this is going to work. Also if lerna bootstrap doesn't work then npm install should do the trick.

Hi @YakoYakoYokuYoku! In order to accept external contributions, we need a signed Contributor Agreement. Could you do that?

Done, now this is ready for reviews. And thanks for reminding me to sign the CA.

Basically this adds a workspaces entry in package.json that reflects the one from lerna.json and in the latter the useWorkspaces entry is set to true.

This was tested with the following commands.

npm install
npm run compile-cli
node packages/cli-repl/bin/mongosh.js

With this PR there's now the possibility to hoist devDependencies, deprecate usages of lerna bootstrap and to update lerna to a much newer version.

rust: Use 'pie' relocation mode under UML

UML expects a position independent executable for some reason, so tell rustc to generate pie objects. Otherwise we get a bunch of relocations we can't deal with in libcore.

Signed-off-by: David Gow davidgow@google.com

arch: um: Disable FP/SIMD instruction to match x86

The kernel disables all SSE and similar FP/SIMD instructions on x86-based architectures (partly because we shouldn't be using floats in the kernel, and partly to avoid the need for stack alignment, see: https://gcc.gnu.org/bugzilla/show_bug.cgi?id=53383 )

UML does not do the same thing, which isn't in itself a problem, but does add to the list of differences between UML and "normal" x86 builds.

In addition, there was a crash bug with LLVM < 15 / rustc < 1.65 when building with SSE, so disabling it fixes rust builds with earlier compiler versions, see: https://github.com/Rust-for-Linux/linux/pull/881

Signed-off-by: David Gow davidgow@google.com

rust: rename PointerWrapper to ForeignOwnable

As discussed elsewhere, this is to keep the rust branch in sync with upstream to simplify porting code from rust to rust-next.

Signed-off-by: Wedson Almeida Filho walmeida@microsoft.com

gpio: pl061: fix new clippy error

Without this fix, we get the following new clippy error:

error: deref which would be done by auto-deref
   --> drivers/gpio/gpio_pl061_rust.rs:285:75
    |
285 |         let gen_inner = unsafe { data.as_mut().map_unchecked_mut(|d| &mut (**d).inner) };
    |                                                                           ^^^^^ help: try this: `d`
    |

Signed-off-by: Wedson Almeida Filho walmeida@microsoft.com

Merge pull request #961 from wedsonaf/foreign-ownable-rust

rust: rename PointerWrapper to ForeignOwnable

Merge pull request #962 from wedsonaf/pl061-clippy

gpio: pl061: fix new clippy error

Merge pull request #881 from sulix/rust_uml_fix

Fix building under UML

rust: doctests: Mark generated rustdoc kunit tests extern "C"

The generated rustdoc tests were marked no_mangle, but not extern "C". Since these are called from C by KUnit, they should use the C ABI.

Signed-off-by: David Gow davidgow@google.com

Merge pull request #968 from sulix/rust_doctest_extern

rust: doctests: Mark generated rustdoc kunit tests extern "C"

rust: forward gfp_t and its flags in kernel

Many places of the kernel make use of gfp_t and its flags to perform certain types of allocations. On the other hand the kernel crate does not expose them for consumers, in particular, they are needed for the USB abstractions. For now gfp_t and some common flags will be available through kernel while in the future more flags can be added.

Signed-off-by: Martin Rodriguez Reboredo yakoyoku@gmail.com

rust: device number abstraction

Signed-off-by: Martin Rodriguez Reboredo yakoyoku@gmail.com

rust: PM message abstractions

Signed-off-by: Martin Rodriguez Reboredo yakoyoku@gmail.com

rust: support for inodes

Signed-off-by: Martin Rodriguez Reboredo yakoyoku@gmail.com

rust: implement try_lock for Mutex

Signed-off-by: Martin Rodriguez Reboredo yakoyoku@gmail.com

rust: xarray: Add an abstraction for XArray

The XArray is an abstract data type which behaves like a very large array of pointers. Add a Rust abstraction for this data type.

The initial implementation uses explicit locking on get operations and returns a guard which blocks mutation, ensuring that the referenced object remains alive. To avoid excessive serialization, users are expected to use an inner type that can be efficiently cloned (such as Arc), and eagerly clone and drop the guard to unblock other users after a lookup.

Future variants may support using RCU instead to avoid mutex locking.

This abstraction also introduces a reservation mechanism, which can be used by alloc-capable XArrays to reserve a free slot without immediately filling it, and then do so at a later time. If the reservation is dropped without being filled, the slot is freed again for other users, which eliminates the need for explicit cleanup code.

Signed-off-by: Martin Rodriguez Reboredo yakoyoku@gmail.com Signed-off-by: Asahi Lina lina@asahilina.net

rust: add support for USB device drivers

Introduces APIs for USB devices and interfaces. Driver traits and implementations are aimed at interfaces only.

Signed-off-by: Martin Rodriguez Reboredo yakoyoku@gmail.com

rust: support for URBs

Adds an API for creating, filling and submitting USB Request Blocks. With support for URB contexts, specialized trait for transport buffers, which mimics ForeignWrapper and completion routines.

Signed-off-by: Martin Rodriguez Reboredo yakoyoku@gmail.com

rust: add USB descriptors

Adds a few useful descriptors for querying data upon USB abstract types.

Signed-off-by: Martin Rodriguez Reboredo yakoyoku@gmail.com

rust: USB simple sample

Adds a sample USB driver that catches incoming bulk requests and writes their transport data to a miscellaneous device.

Signed-off-by: Martin Rodriguez Reboredo yakoyoku@gmail.com