t/41screen_unicode test failures

[James McCoy]

** t/41screen_unicode.test **
> Single width UTF-8
> Wide char
> Combining char
> 10 combining accents should not crash
# Assert screen_cell 0,0 failed:
# Expected: {0x65,0x301,0x302,0x303,0x304,0x305} width=1 attrs={} fg=rgb(240,240,240) bg=rgb(0,0,0)
# Actual: {0x65,0x301,0x302,0x303,0x304,0x305,0xf0f001} width=1 attrs={} fg=rgb(240,240,240) bg=rgb(0,0,0)
> 40 combining accents in two split writes of 20 should not crash
# Assert screen_cell 0,0 failed:
# Expected: {0x65,0x301,0x301,0x301,0x301,0x301} width=1 attrs={} fg=rgb(240,240,240) bg=rgb(0,0,0)
# Actual: {0x65,0x301,0x301,0x301,0x301,0x301,0xf0f001} width=1 attrs={} fg=rgb(240,240,240) bg=rgb(0,0,0)

-- System Information:
Debian Release: jessie/sid
  APT prefers unstable
  APT policy: (500, 'unstable'), (1, 'experimental')
Architecture: amd64 (x86_64)
Foreign Architectures: i386

Kernel: Linux 3.13-1-amd64 (SMP w/8 CPU cores)
Locale: LANG=en_US.UTF-8, LC_CTYPE=en_US.UTF-8 (charmap=UTF-8)
Shell: /bin/sh linked to /bin/dash

[James McCoy (jamessan)]

Finally got around to looking at this more, and it might be a compiler issue. I only hit this with "gcc -O2". Using -O0/-O1 or clang instead hits https://bugs.launchpad.net/libvterm/+bug/1314448.

[James McCoy (jamessan)]

Fixed in r618