Getting ready for release 2.0.0-beta.4

ANNOUNCE.rst

@@ -1,16 +1,17 @@
 ===============================================================
- Announcing C-Blosc2 2.0.0-beta.3
+ Announcing C-Blosc2 2.0.0-beta.4
  A simple, compressed and fast data store library for C
 ===============================================================
 
 What is new?
 ============
 
-This is the third beta version, a quick release to fix the fact that the previous one was made on a branch different than master.  Also, this includes minor fixes in the trailer section of the frame; hence, the trailer version has been bumped to 1.
+This is the fourth beta version, adding an important fix for making 64-bit containers to actually work.  Also, a new pluggable threading backend has been added, allowing better interaction with the threading mechanism that you are using in the rest of your program
+  (e.g. OpenMP or Intel TBB).  Finally, a cache has been added for on-disk offsets.  This accelerates the reading of slices from disk quite a lot (up to 50% with my benchmarks).
 
 In principle, C-Blosc2 should be backward compatible with C-Blosc, so you can start using it right away and slowly begin to use its new functionality, like the new filters, prefilters, super-chunks and frames.  See docs in: https://blosc-doc.readthedocs.io
 
-**IMPORTANT**: Please note that, even if the API has been declared frozen, that does not mean that Blosc2 is ready for production yet: internal structures can change, formats can change and most importantly, bugs can be normal at this stage.  So *do not assume* that your blosc2 data can be read with future versions.
+**IMPORTANT**: Please note that, even if the API has been declared frozen, that does *not* mean that Blosc2 is ready for production yet: internal structures can change, formats can change and most importantly, bugs can be normal at this stage.  So *do not assume* that your blosc2 data can be read with future versions.
 
 For more info, please see the release notes in:
 

DEVELOPING-GUIDE.rst

@@ -1,14 +1,14 @@
 Some conventions used in C-Blosc2
 =================================
 
-* Use C99 designated initialization whenever possible (specially in examples)
+* Use C99 designated initialization whenever possible (specially in examples).
 
 * Use _new and _free for memory allocating constructors and destructors and _init and _destroy for non-memory allocating constructors and destructors.
 
 
 Naming things
 -------------
 
-Naming is one of the most time-consuming tasks, but critical for communicating effectively.  Here it is a preliminary list of names that I am not comfortable with: 
+Naming is one of the most time-consuming tasks, but critical for communicating effectively.  Here it is a preliminary list of names that I am not comfortable with:
 
 * We are currently calling `filters` to a data transformation function that essentially produces the same amount of data, but with bytes shuffled or transformed in different ways.  Perhaps `transformers` would be a better name?

RELEASE_NOTES.md

@@ -6,6 +6,22 @@
 Changes from 2.0.0-beta.3 to 2.0.0-beta.4
 =========================================
 
+* New pluggable threading backend.  Instead of having Blosc use its own
+  thread pool, you can instead call
+  `blosc_set_threads_callback(threads_callback, callback_data)` to install
+  your own threading backend.  This gives Blosc the possibility to use the
+  same threading mechanism as one you are using in the rest of your program
+  (e.g. OpenMP or Intel TBB), sharing the same threads, rather than starting
+  its own threads that compete with yours for the CPU cores. See #PR 81.
+  Thanks to Steven G. Johnson.
+
+* The endianness of the platform that is writing the data in chunks is stored
+  now in the headers of the chunks.  This info is not used yet, but this
+  should allow a good hint for implementing format compatibility among
+  platforms with different endianness in other layers.  See PR #84. 
+
+* Fixed a nasty bug that prevented frames to go more than 2 GB in size.
+
 * Added a cache for on-disk offsets.  This accelerates the reading of slices
   from disk quite a lot (up to 50% with my benchmarks).
 

blosc/blosc2.h

@@ -33,9 +33,9 @@ extern "C" {
 #define BLOSC_VERSION_MINOR    0    /* for minor interface/format changes  */
 #define BLOSC_VERSION_RELEASE  0    /* for tweaks, bug-fixes, or development */
 
-#define BLOSC_VERSION_STRING   "2.0.0-beta.4.dev"  /* string version.  Sync with above! */
+#define BLOSC_VERSION_STRING   "2.0.0-beta.4"  /* string version.  Sync with above! */
 #define BLOSC_VERSION_REVISION "$Rev$"   /* revision version */
-#define BLOSC_VERSION_DATE     "$Date:: 2019-09-02 #$"    /* date version */
+#define BLOSC_VERSION_DATE     "$Date:: 2019-09-13 #$"    /* date version */
 
 
 /* The VERSION_FORMAT symbols below should be just 1-byte long */