- Update to version 1.0+git.20240530.96f8fc5:
  * ld: Global constructor/destructor support
  * Build in C++11 mode
  * Import 'config/ax_cxx_compile_stdcxx.m4' from GCC
  * Replace 'nvptx-none-run-single' with a locking mechanism in 'nvptx-none-run.exp'
  * Change github.com/MentorEmbedded to github.com/SourceryTools

• Files Changed
• Browse Source

- Update to version 1.0+git.20240419.9962793:
  * as: Document current handling of '//:IND_FUNC_MAP' [#45]
  * Handle ":IND_FUNC_MAP" in nvptx-as to ensure map-directive order is kept
  * Handle --hash-style argument in nvptx-ld

• Files Changed
• Browse Source

- Update to version 1.0+git.20230913.c321f1a:
  * as: Allow for arbitrary number of tokens in a 'stmt'
  * Ensure :VAR_MAP and :FUNC_MAP are output in order, part II [#29]
  * Ensure :VAR_MAP and :FUNC_MAP are output in order: add more test cases [#29]
  * as: Pass 'Vis' into 'parse_line_nosemi'
  * as: Inline 'alloc_comment'
  * as: When tokenizing a block comment, don't lose a preceding end of line marker
  * as: Add 'test/as/block_comment-1.test'
  * Bump architecture baseline from sm_35 to sm_50 [#30]
  * as: Remove unused 'FUNC_MAP', 'VAR_MAP' scanning
  * as: Clean up 'symbol's via C++ destructor
  * Enable 'test/as/ptxas/missing-1.test' also if 'ptxas' is available
  * Adjust for 'pex_run' error message change in "Update imported files: GCC, 2018-10-29 [#1, #42]"
  * Update imported files: GCC, 2023-06-16 [#1, #42]
  * Update imported files: GCC, 2022-05-18, II [#1, #42]
  * Update imported files: GCC, 2022-05-18 [#1, #42]
  * Update imported files: GCC, 2022-04-06 [#1, #42]
  * Update imported files: GCC, 2022-03-20 [#1, #42]
  * Switch to Autoconf 2.69, Automake 1.15.1 [#42]
  * Update imported files: GCC, 2018-10-31 [#1, #42]
  * Update imported files: GCC, 2018-10-29 [#1, #42]
  * Update imported files: GCC, 2016-07-29 [#1, #42]
  * Update imported files: GCC, 2014-05-01 [#1, #42]
  * Note "Imported Files" in 'README-development' [#1]
  * Add 'move-if-change' [#1]
  * Remove 'acinclude.m4'
  * Don't use 'gcc_AC_PROG_INSTALL'; use standard 'AC_PROG_INSTALL'
  * Don't use 'gcc_AC_PROG_LN_S'; assume that 'ln -s' does work
  * Have 'aclocal.m4' be "generated automatically by aclocal"
  * Move 'ACX_PKGVERSION', 'ACX_BUGURL' into 'config/acx.m4'

• Files Changed
• Browse Source

- Update to version 1.0+git.20230718.bbb6fe6:
  * Document "Program Instrumentation: Sanitizers" in 'README-development'
  * run: Move 'cuModuleGetFunction' for '__main' out of 'compile_file'
  * run: Plug 'compile_file' memory leaks
  * run: Localize 'cuda_lib'
  * as: Localize 'collect_wait'
  * as: Don't 'unlink (NULL)'
  * as: Don't use an obstack for 'ptxas' invocation [#1]
  * ld: Remove unused '#include "obstack.h"' [#1]
  * ld, nm: Stop 'using namespace std;'
  * run: For auto '--stack-size', consider free instead of total device memory [#8]
  * ld: Support archives appearing as input files
  * ld: Don't reject empty archives
  * ld: Make resolving of '-l'ibraries a separate step
  * ld: Verify no NUL-separated parts for 'process_refs_defs'
  * ld: Fix undefined behavior reading objects from archives
  * ld: Document padding of objects in archives
  * as: Plug 'alloc_stmt' memory leak
  * as: Plug 'read_file' memory leak
  * as: Plug 'tokenize' memory leak
  * as: Conceptually simplify capture of preamble '.target' directive's argument
  * 'test/GLOBAL_FUNCTION_DECL_f.s': 'f' is '.extern', not '.visible' [#38]

• Files Changed
• Browse Source

- Start using tar_scm OBS server for a tarball.
- Update to version 1.0+git.20230122.93e0090:
  * nm: '#define __STDC_FORMAT_MACROS' for '#include <inttypes.h>' [#40]
  * 'nm'
  * ld: Clean up the 'symbol_table' when we're done
  * as: Clean up the 'symbol_table' when we're done
  * ld: Clean up 'file_hash_entry' instances when we're done with the 'symbol_table'
  * as: Always 'free' the 'tok_preamble_target_arg'-deduced 'target_arg' string after 'ptxas' verification
  * as: 'obstack_free' after 'ptxas' verification
  * run: Close files after use (forwarded request 1073654 from marxin)

• Files Changed
• Browse Source

Fix Wolfang Engel complaints, rejecting the SLE update because of
"missing nvptx-tools.patch removal" notes.

- Remove nvptx-tools.patch as it is included in the above tarball update

• Files Changed
• Browse Source

- Update nvptx-tools tarball from git
  * includes changes in nvptx-tools.patch
  * deals with CUDA 11 dropping support for NVIDIA Kepler sm_30 and sm_32
- Package nvptx-none-run and nvptx-none-run-single tools

• Files Changed
• Browse Source

• Files Changed
• Browse Source

- Update nvptx-tools.patch to check for getopt to avoid providing
  an incompatible declaration.

• Files Changed
• Browse Source

• Browse Source