- 30 Oct, 2018 37 commits
-
-
Khaled Hosny authored
Co-Authored-By:
n8willis <nwillis@glyphography.com>
-
Khaled Hosny authored
Co-Authored-By:
n8willis <nwillis@glyphography.com>
-
Nathan Willis authored
-
Khaled Hosny authored
Co-Authored-By:
n8willis <nwillis@glyphography.com>
-
Nathan Willis authored
-
Nathan Willis authored
-
Nathan Willis authored
-
Nathan Willis authored
-
Nathan Willis authored
-
Nathan Willis authored
-
Nathan Willis authored
-
Nathan Willis authored
-
Nathan Willis authored
-
Nathan Willis authored
-
Nathan Willis authored
-
Nathan Willis authored
-
Nathan Willis authored
-
-
Behdad Esfahbod authored
-
Behdad Esfahbod authored
-
Behdad Esfahbod authored
-
Behdad Esfahbod authored
-
Behdad Esfahbod authored
-
Ebrahim Byagowi authored
-
Behdad Esfahbod authored
Speeds up runnings in-house tests from over 20s to 2s.
-
Behdad Esfahbod authored
-
Behdad Esfahbod authored
-
Behdad Esfahbod authored
-
Ebrahim Byagowi authored
-
Behdad Esfahbod authored
I keep forgetting that primitive types are NOT initialized during construction. :|
-
Behdad Esfahbod authored
-
Behdad Esfahbod authored
-
Behdad Esfahbod authored
-
Behdad Esfahbod authored
-
Behdad Esfahbod authored
-
Behdad Esfahbod authored
See a85641446c30247c4e948263f0f8c1147ed4efb9
-
Behdad Esfahbod authored
Disable copy/assign on them, as they shouldn't. Make constructor / destructor call init_shallow/fini_shallow, and make those idempotent. So, these three can be constructed on stack now and no init/fini call is needed. As such, hb_auto_t<> is not needed anymore. I'll remove that separately.
-
- 29 Oct, 2018 3 commits
-
-
Behdad Esfahbod authored
-
Behdad Esfahbod authored
It was arbitrary that this struct had it and not dozens of others.
-
Behdad Esfahbod authored
-