cubature 2.0.4

• cubintegrate now matches method via match.arg (Issue 25)

• Address gcc version 10.0 changes due to -fno-common default setting.

cubature 2.0.3

• Fixed up stack overrun in Cuba-4.2/src/divonne/Split.c (lines 119–128 utilizing a flag for first time through loop
• More cleanup of Makevars

cubature 2.0.2

• Fixed up uninitialized count for hcubature (Thanks, Ehsan Masoudi)
• Tentative fix of infinite value for minfluct in Cuba-4.2/src/suave/Integrate.c in lines 197–205. (Thanks, Prof. Brian Ripley)
• Miscellaneous fixes for compilation on various platforms.
  • Cleaned up Makevars and Makevars.win to remove some unused flags. Tentative fix for solaris in Makevars and Cuba-4.2/src/common/stddecl.c.
  • Removed printing of statefile in all the src/*/Integrate.c routines to avoid segfault on solaris.
  • Added ret_code for return value in call to getloadavg in Cuba-4.2/src/common/Fork.c. Also added an #ifdef SOLARIS for including appropriate header under solaris.
  • Fixed up call to MASTER variadic macro by including dummy argument 0 in Cuba-4.2/src/common/Fork.c, line 146.
  • Replaced multi-statement macro with a do { } while(0) hack.
  • Reworked code around Cuba-4.2/src/divonne/Rule.c checks for key values (lines 593 to 600).
  • Fixed up call to WORKER variadic macro by including dummy argument 0 in Cuba-4.2/src/common/Parallel.c, line 393.
  • Moved #ifdef FRAMECOPY outside in Cuba-4.2/src/common/Parallel.c to avoid embedding directives in macro args, lines 70-80.

cubature 2.0

• Major update. Integrates Cuba 4.2 library
• Allows for finite or infinite limits of integration (courtesy of Simon Gaure)
• Deprecates doChecking argument. Now it does nothing and will be removed in future versions.

cubature 1.4-1

• Fixed up scale argument to tests to conform to pass checks.
• Removed any reference to orphaned package R2Cuba and updated vignette with information on imminent 2.0 release.

cubature 1.4

• Fixed up the C call so that it is re-entrant (brought to my attention by Pierre de Villemereuil). This should be considered a bug fix!

• Corrected private notes

cubature 1.3-13

• Generated package docs using pkgdown
• Synced up to cubature-1.0.3
• Added LICENSE file (request of Nick Youngblut)

cubature 1.3-12

• Minor typographical and documentation fixes

cubature 1.3-11

• Merged Manuel Koller’s registration for C code and vignette fix for NA.
• Moved cubature header and exp_cubature headers to inst/include for linking to other packages

cubature 1.3-10

• Renamed Readme.md to README.md
• Removed references to ab-initio website that caused some hassle due to misconfigured site.

cubature 1.3-9

cubature 1.3-8

cubature 1.3-7

• Registered .Call stuff and removed microbenchmark suggestion in favor of benchr.

cubature 1.3-4

• Generated smaller pcubature header (clencurt.h) using M = 16 and put back C cubature source in tree.

cubature 1.3-3

• Moved the cubature-1.0.2 C library to github to avoid hitting the CRAN size limit.

cubature 1.3-2

• Added vignette showing huge gains due to vectorization

cubature 1.3-1

• Added pcubature
• Added vector versions

cubature 1.3-0

• Moved to SGJ cubature-1.0.2 version.
• adaptIntegrate and hcubature are aliases
• hcubature function gains a norm argument that is set to a sensible default so as not to affect depending packages

cubature 1.2-0

• Moved to Rcpp framework
• Added tests in preparation for SGJ cubature-1.0.2 version

cubature 1.1-3

• Roxygenized in preparation for upgrade to newer version of cubature library on abinitio website

cubature 1.1-2

• Registered native cubature functions adapt_integrate and adapt_integrate_v so that they are directly callable from C (courtesy of Simen Gaure)

cubature 1.1-2

• Fixed typo in doc for function adaptIntegrate; default value for doChecking was incorrectly stated as TRUE

cubature 1.1-1

• Added doChecking argument (default FALSE) to save some computation time in evaluating integrand (9% speedup).

cubature 1.1

• Synced up to SGJ cubature routines dated 2010-10-18 on his website
• Bugfix: potential memory leak fixed up in heap routine (my oversight!)
• Routine adaptIntegrate gains … argument (request of Baptiste Auguie)
• Corrected radius constant in testFn2 to match cubature output exactly

cubature 1.0

• Original version of package based on Steven G. Johnson’s cubature routines at http://ab-initio.mit.edu/wiki/index.php/Cubature