WHATSNEW 3.7 KB

123456789101112131415161718192021222324252627282930313233343536373839404142434445464748495051525354555657585960616263646566676869707172737475767778798081828384858687888990919293949596979899100101102103104105106107108109110111112113114115116117118119120121122123124125126127128129130131132133134135136137138139140141142143144145146147148149150151152153
  1. 2019-09-24 Matias A. Fonzo <selk@dragora.org>
  2. Version 2019Sep24 released.
  3. * Binutils 2.32 + updates from 2_32-branch (version: 20190914).
  4. * Linux-libre 4.19.75 (headers).
  5. * GCC-9-20190921 (from snapshot, post gcc-9.2.0).
  6. - GMP 6.1.2.
  7. - MPFR 4.0.2 + cumulative patch (latest update: 2019-06-02).
  8. - MPC 1.1.0.
  9. - ISL 0.21.
  10. * Musl updated to the snapshot 20190913_f5eee489 (post musl-1.1.23).
  11. * Targets:
  12. armv7-a:
  13. - This soft float target has been renamed to 'armv7a_fp'.
  14. - The CPU name for the target triplet has been changed to be more
  15. generic (from armv7a to arm).
  16. armv7a_hf:
  17. - New target added. This is almost the same target as 'armv7a_fp'
  18. with the difference that this target is hard float, plus FPU
  19. 'vfpv3-d16'.
  20. armv7hl:
  21. - New target added. This is a hard float target for the armv7-a
  22. architecture (hard float) but with the FPU tuned for 'neon'.
  23. i586:
  24. - The option --enable-cld for the CLD instruction is now used by
  25. default (just in case).
  26. powerpc:
  27. - Option for decimal support has been disabled here.
  28. powerpc64:
  29. - The option --with-abi=elfv2 is now passed by default (just in case).
  30. - The target name has been renamed to 'powerpc64le' for clarification.
  31. riscv64:
  32. - New target added with multilib support. This is a hard float target
  33. for the architecture 'rv64gc' with the ABI 'lp64d'.
  34. s390x:
  35. - Directory to look into the corresponding (kernel) headers for the
  36. architecture has been fixed. (s390x -> s390).
  37. * Toolchain:
  38. - Option for binutils --disable-separate-code is now given by default.
  39. See: http://www.openwall.com/lists/musl/2019/08/12/4
  40. - Patches for GCC has been updated to work against Musl, GCC-9.
  41. Thanks to Szabolcs Nagy for provide it. :-)
  42. - config.sub and config.guess have been updated in order to detect more
  43. target triplets, for example those containing "armv7hl"...
  44. * Darkcrusade is now distributed using the tarlz format. For more
  45. information, visit http://lzip.nongnu.org/tarlz.html
  46. * Minor changes.
  47. 2018-05-26 Matias A. Fonzo <selk@dragora.org>
  48. Version 2018May26 released.
  49. * Binutils 2.30.
  50. * Linux-libre 4.14.41 (headers).
  51. * GCC-7-20180524 (from snapshot, post gcc-7.3.0).
  52. - GMP 6.1.2.
  53. - MPFR 4.0.1 + cumulative patch (latest update: 2018-05-01).
  54. - MPC 1.1.0.
  55. - ISL 0.19.
  56. * Musl 1db9a35 (from snapshot, post musl-1.1.19).
  57. * targets: powerpc64: added new target file.
  58. Thanks to David Edelsohn.
  59. * targets: s390x: added new target file. Thanks to David Edelsohn.
  60. * targets: i586, x32: fixed issue with multilib option.
  61. * targets: x86_64: enhanced multilib option for tuning GCC:
  62. "--with-arch=x86-64 --with-tune=generic".
  63. * The Apache license version 2 is now used by darkcrusade
  64. (scripts).
  65. * Minor changes.
  66. 2017-06-20 Matias A. Fonzo <selk@dragora.org>
  67. Version 2017Jun09 released.
  68. * The pass for the CC and CXX environment flags have been fixed for
  69. the final GCC -- previously resulting in a partial
  70. statically-linked cross-compiler.
  71. * Linux-libre 4.9.31 (headers).
  72. * Checksum files are now produced using sha256sum(1).
  73. 2017-01-06 Matias A. Fonzo <selk@dragora.org>
  74. Version 2017Jan05 released.
  75. * Binutils 2.26.1.
  76. * Linux-libre 4.4.39 (headers).
  77. * GCC 6.3.0.
  78. - GMP 6.1.2.
  79. - MPFR 3.1.5 (patch level 2).
  80. - MPC 1.0.3.
  81. - ISL 0.18.
  82. * Musl 1.1.16.
  83. 2016-06-19 Matias A. Fonzo <selk@dragora.org>
  84. * Initial release.