Google Git
Sign in
apache / apr / f4d474d9874d69a06e9dfbab1e33ed6e412d41b9
commitf4d474d9874d69a06e9dfbab1e33ed6e412d41b9[log] [tgz]
authorWilliam A. Rowe Jr <wrowe@apache.org>Wed Jun 08 22:39:58 2016 +0000
committerWilliam A. Rowe Jr <wrowe@apache.org>Wed Jun 08 22:39:58 2016 +0000
treebcbf4fbe941d405f0782ca93c73aed963cb1ede9
parentb65726d0823c1e4c64a434635556c1a1fa2acb58 [diff]
Upon closer scrutiny, the optimization of using short variables
(as opposed to the array of shorts) really didn't hold up.
Revert back to processing in 'int' amounts, persist in using
short values in the character case folding table, which are
more easily retrieved based on word alignment.  Aligning and
representing the table values as 'int' doesn't exhibit any
additional benefit and wastes some DATA page space.



git-svn-id: https://svn.apache.org/repos/asf/apr/apr/trunk@1747461 13f79535-47bb-0310-9956-ffa450edef68
1 file changed
tree: bcbf4fbe941d405f0782ca93c73aed963cb1ede9
  1. atomic/
  2. buckets/
  3. build/
  4. crypto/
  5. dbd/
  6. dbm/
  7. docs/
  8. dso/
  9. encoding/
  10. file_io/
  11. helpers/
  12. hooks/
  13. include/
  14. locks/
  15. memcache/
  16. memory/
  17. misc/
  18. mmap/
  19. network_io/
  20. passwd/
  21. poll/
  22. random/
  23. shmem/
  24. strings/
  25. strmatch/
  26. support/
  27. tables/
  28. test/
  29. threadproc/
  30. time/
  31. tools/
  32. uri/
  33. user/
  34. util-misc/
  35. xlate/
  36. xml/
  37. apr-config.in
  38. apr.dsp
  39. apr.dsw
  40. apr.pc.in
  41. build.conf
  42. buildconf
  43. CHANGES
  44. CMakeLists.txt
  45. config.layout
  46. configure.in
  47. emacs-mode
  48. libapr.dsp
  49. libapr.rc
  50. LICENSE
  51. Makefile.in
  52. Makefile.win
  53. NOTICE
  54. NWGNUmakefile
  55. README
  56. README.cmake
  57. README.FREETDS
  58. SConscript
  59. SConstruct
  60. STATUS