diff options
author | Kaz Kylheku <kaz@kylheku.com> | 2021-04-20 04:36:05 -0700 |
---|---|---|
committer | Kaz Kylheku <kaz@kylheku.com> | 2021-04-20 04:36:05 -0700 |
commit | d467598989bcd815adf30e42f9a67a73929f9d71 (patch) | |
tree | 2c4f39c4ab225bda9c409e2db90a319aedccc3bc /configure | |
parent | 1b78b60d156e3c8593efdc04133131bfcaffaeab (diff) | |
download | txr-d467598989bcd815adf30e42f9a67a73929f9d71.tar.gz txr-d467598989bcd815adf30e42f9a67a73929f9d71.tar.bz2 txr-d467598989bcd815adf30e42f9a67a73929f9d71.zip |
Revert bogus LIT_ALIGN commit from 2015.
This reverts commit 0343c6f32c5bd8335e88595cb9d23506625b2586.
I don't see evidence that the claim in the commit is true:
that wide literals are not four-byte-aligned on Darwin in
spite of sizeof(wchar_t) being 4. Not even with the old clang
in my old VM where I first thought I discovered this.
* configure: do not set up LIT_ALIGN == 2 for Darwin.
* lib.h (litptr): Remove LIT_ALIGN < 4 && SIZEOF_WCHAR_T == 4
case.
* HACKING: Undocument bogus claim.
Diffstat (limited to 'configure')
-rwxr-xr-x | configure | 6 |
1 files changed, 1 insertions, 5 deletions
@@ -1503,11 +1503,7 @@ if [ -z "$lit_align" ] ; then exit 1 fi - if [ -n "$darwin_target" ] && [ "$(arch)" = "i386" ] ; then - lit_align=2 - else - lit_align=$SIZEOF_WCHAR_T - fi + lit_align=$SIZEOF_WCHAR_T fi printf "%d\n" "$lit_align" |