| 1 |
Hi, Patrick and list |
| 2 |
|
| 3 |
This is a weekly report of a SoC project "Porting Gentoo to DragonFly |
| 4 |
BSD". |
| 5 |
|
| 6 |
From the last week to now, I've done these things. |
| 7 |
|
| 8 |
- install sandbox binary, portage |
| 9 |
- sync to portage tree |
| 10 |
- create "profiles/default/bsd/dfbsd", "profiles/arch/x86-dfbsd", |
| 11 |
"profiles/releases/dfbsd-2.6" like freebsd's ones. |
| 12 |
- create copy of the base package ebuilds and KEYWORD them as |
| 13 |
"~x86-dfbsd" |
| 14 |
- fix some of the ebuilds |
| 15 |
- install "eseleclt-python" by hand using ebuild. |
| 16 |
- USE='-*' emerge -1 portage |
| 17 |
|
| 18 |
Now I'm trying "emrge -uDN world" on DragonFly BSD. |
| 19 |
|
| 20 |
Here are some problems I'm struggling with. |
| 21 |
|
| 22 |
- "sort -z" |
| 23 |
DragonFly BSD's "sort" have no "-z" option, so that "sandbox" seems |
| 24 |
not work. (ebuild.sh uses the option to deal something "sandbox" |
| 25 |
related works) |
| 26 |
|
| 27 |
- sandbox porting |
| 28 |
Linux getcwd() set errno = ENAMETOOLONG when The path was longer |
| 29 |
than PATH_MAX. But this behavior is not compatible with POSIX. |
| 30 |
"sandbox" is written to follow this Linux manner, so sandbox not |
| 31 |
working so well on *BSD. |
| 32 |
|
| 33 |
Here is my work around patch. But I think it would be better to |
| 34 |
convert ERANGE to ENAMETOOLONG or reverse at somewhere (maybe after |
| 35 |
calling vanilla getcwd() is adequate) |
| 36 |
|
| 37 |
http://github.com/naota/gentoo-dragonflybsd/commit/810b0be2b897579327b14fc599ba3024983237c7 |
| 38 |
|
| 39 |
Regards, |
Archives