Prep 5.5.0 release candidate

CMakeLists.txt

@@ -25,7 +25,7 @@ if(MSVC)
     add_compile_options("/wd4324")
 endif()
 
-project(astcencoder VERSION 5.4.0)
+project(astcencoder VERSION 5.5.0)
 
 set(CMAKE_CXX_STANDARD 14)
 set(CMAKE_CXX_STANDARD_REQUIRED ON)

Docs/ChangeLog-5x.md

@@ -9,39 +9,43 @@ clocked at 4.2 GHz, running `astcenc` using AVX2 and 6 threads.
 <!-- ---------------------------------------------------------------------- -->
 ## 5.5.0
 
-**Status:** In development.
-
-The 5.5.0 release is a minor maintenance release.
-
-* **General:**
-  * **Update:** Update stb_image to v1.30.
-  * **Update:** Update Wuffs to v0.3.4.
-  * **Update:** Update TinyEXR to v1.0.13.
-  * **Improvement:** Decompressing a block using an unsupported partition index
-    when using `ASTCENC_FLG_SELF_DECOMPRESS_ONLY` now returns the standard
-    block error color.
-  * **Bug fix:** Core library now correctly detects LDR error blocks when
-    compiled for AVX2 and returns max magenta rather than black.
-  * **Bug fix:** Front-end wrapper now uses C++ RAII to manage lifetime of
-    memory and file handles, fixing a number of resource leaks when on an error
-    handling path.
-  * **Bug fix:** Throw errors in the command line wrapper for cases where
-    `.astc`, `.dds`, and `.ktx` input image sizes would result in integer
-    overflow in calculations of total block count or data size.
-  * **Bug fix:** Throw errors in the library API for cases where input image
-    sizes would result in integer overflow in calculations of total texel
-    count, total block count, or compressed image byte size.
-  * **Bug fix:** Fix incorrect offset calculation in `compute_averages()` when
+**Status:** June 2026
+
+The 5.5.0 release is a minor maintenance release, fixing minor functional
+issues and improving robustness when processing invalid images.
+
+* **Codec library updates:**
+  * **Bug fix:** Correct `vfloat8` NaN inequality comparisons for AVX2 builds,
+    so that decompressing an invalid encoding returns the correct max magenta
+    error color for UNORM output formats instead of black.
+  * **Bug fix:** Add overflow detection so that the API returns errors for
+    cases when input image size would result in integer overflow in
+    calculations of total texel count, total block count, or compressed image
+    byte size.
+  * **Bug fix:** Fix bad offset calculation in `compute_averages()` when
     using 3D block sizes.
   * **Bug fix:** Add missing low-clamp when storing decompressed HDR data into
     a UNORM8 image. Prior to this fix, output colors would be incorrect if a
-    HDR void-extent block contained a negative FP16 color value.
-  * **Bug fix:** Avoid undefined behavior on NaN to int conversion, that is
-    triggered by NaNs generated by `atan2()` calls where both inputs are zero.
+    HDR void-extent block contained a negative FP16 constant color component.
+  * **Improvement:** Return standard error color when decompression requires an
+    unavailable partitioning when using an `ASTCENC_FLG_SELF_DECOMPRESS_ONLY`
+    context. Note that this is still not allowed API usage and may trigger
+    asserts in a debug build.
+* **Command line tool updates:**
+  * **Update:** Update stb_image to v2.30.
+  * **Update:** Update Wuffs to v0.3.4.
+  * **Update:** Update TinyEXR to v1.0.13.
+  * **Bug fix:** Use C++ RAII to manage resource lifetime, removing resource
+    leaks when an error condition is encountered.
+  * **Bug fix:** Report errors for `.astc`, `.dds`, and `.ktx` input image
+    sizes that would result in integer overflow in calculations of total block
+    count or data size.
+  * **Bug fix:** Avoid undefined behavior on NaN to int conversion triggered by
+    `atan2()` calls when both inputs are zero.
   * **Bug fix:** Avoid undefined behavior on left shift of negative integer,
     that is triggered for some HDR encodings in `hdr_alpha_unpack()`.
-  * **Bug fix:** Report error messages for failed `.dds` and `.ktx` input image
-    loading to stderr, to match other formats.
+  * **Bug fix:** Redirect error logging for failed `.dds` and `.ktx` input
+    image loading to stderr, and improve message consistency.
 
 <!-- ---------------------------------------------------------------------- -->
 ## 5.4.0