# 02:54:32 |
mnt_real |
joins #crosstool-ng |
# 06:36:35 |
kos_tom |
joins #crosstool-ng |
# 07:02:05 |
y_morin |
joins #crosstool-ng |
# 07:45:29 |
meyrntc |
joins #crosstool-ng |
# 07:45:42 |
mnt_real |
quits : Remote host closed the connection |
# 08:01:26 |
meyrntc |
hi y_morin; is there a reason for gmp not to being compiled for CT_TARGET and installed into the CT_PREFIX_DIR directory? |
# 08:02:18 |
y_morin |
meyrntc: hey! Crosstool-NG only builds what is absolutely required to build a toolchain. |
# 08:02:42 |
y_morin |
meyrntc: GMP on the target is not required for the toolchain, so is not built. |
# 08:02:51 |
y_morin |
Ditto MPC, MPFR, PPL... |
# 08:03:37 |
meyrntc |
ok; what about makeing it optional? :) |
# 08:04:42 |
y_morin |
meyrntc: Crosstool-NG is only about building the toolchain, so it's not in the scope. |
# 08:17:03 |
y_morin |
meyrntc: Thinking back, if there is a hard-requirement that the _same_ version of those libs are needed on the target for compatibility reasons, then we can add them. |
# 08:17:29 |
y_morin |
meyrntc: For example, that's the reason why we build the binutils libs for the target, because they are needed by oprofile, and must be the exact same ones as present in the toolchain. |
# 08:17:54 |
y_morin |
meyrntc: but I doubt this is the case for the companion libraries... |
# 08:24:34 |
meyrntc |
Cannot say, if it's really needed. We just need the gmp libs for our different cross-compilations, so I extended the script to build and install for the target, too. Other thing: when restarting the build, the gcc.sh tries to make a symbolic link the the compiler (ln -sv ...), which fails. Perhaps, it should remove an existing link beforehand? |
# 08:28:48 |
meyrntc |
And a third thing: Perhaps crosstool-ng could silently ignore, if the CT_PREFIX_DIR itself cannot be removed (in case where it is the top level writable directory for the installing user), when the CT_RM_RF_PREFIX_DIR is set. |
# 08:31:23 |
kos_tom |
y_morin, oh, oprofile needs the same binutils on the target as the one in the cross-toolchain ? |
# 08:41:02 |
y_morin |
kos_tom: I don't recall exactly why, but that was the reason why it was added into crosstool-NG. |
# 08:41:40 |
y_morin |
kos_tom: hard to tell, it was a looong time ago, and the history at the time got lost... |
# 08:42:40 |
y_morin |
meyrntc: I see what you mean: CT_PREFIX_DIR is writable, but not the parent, so we can't remove CT_PREFIX_DIR, while we can remove its content. |
# 08:43:23 |
y_morin |
meyrntc: in this case, we should not even try to remoce CT_PREFIX_DIR, but only its content. That would solve both cases, right? |
# 08:44:13 |
y_morin |
kos_tom: OK I see why you're interested in that! ;-) |
# 08:45:47 |
y_morin |
is now known as: y_morin|away |
# 09:13:12 |
y_morin|away |
is now known as: y_morin |
# 09:52:16 |
y_morin |
is now known as: y_morin|lunch |
# 11:10:32 |
y_morin|lunch |
is now known as: y_morin |
# 11:30:22 |
kos_tom |
quits : Quit: Leaving |
# 11:33:23 |
kos_tom |
joins #crosstool-ng |
# 14:30:35 |
meyrntc |
quits : Quit: thank you |
# 15:16:59 |
Buglouse |
quits : Quit: Love IRC |
# 15:17:34 |
Buglouse |
joins #crosstool-ng |
# 15:46:49 |
y_morin |
quits : Quit: Back on-line in about a few hours... |
# 16:06:05 |
kos_tom |
quits : Ping timeout: 260 seconds |
# 16:43:49 |
y_morin |
joins #crosstool-ng |
# 16:44:37 |
y_morin |
quits : Remote host closed the connection |
# 16:51:25 |
y_morin |
joins #crosstool-ng |
# 17:57:46 |
y_morin |
quits : Remote host closed the connection |
# 17:57:57 |
linuxjacques |
joins #crosstool-ng |
# 17:58:26 |
y_morin |
joins #crosstool-ng |
# 20:20:54 |
y_morin |
quits : Quit: Nite all! |
# 21:22:52 |
bhundven |
joins #crosstool-ng |
# 21:23:12 |
bhundven |
ah... no yann. |
# 23:18:28 |
linuxjacques |
quits : Quit: Page closed |