# 04:25:12 |
cpackham |
quits : Ping timeout: 252 seconds |
# 19:22:57 |
cpackham |
joins #crosstool-ng |
# 20:29:56 |
Trel |
zagura: Should just exporting CT_TARGET_VENDOR set to something work when I build is there more to it? |
# 20:38:00 |
Trel |
I see it in menuconfig as an option which is how I did it currently |
# 21:10:48 |
zagura |
What do you mean with your question? |
# 21:11:09 |
zagura |
menuconfig/.config file is the place for the setting |
# 21:17:33 |
Trel |
I was asking if having that as an ENVAR when you did a build would affect it. |
# 21:18:03 |
Trel |
But I'm back to a different question because I just got the whole buildroot thing setup only to find, it's not viable for my case either. |
# 21:19:38 |
Trel |
I'm at a complete and total loss as to where to even look. |
# 21:28:35 |
bhundven |
cpackham: ð |
# 21:30:16 |
cpackham |
hey bhuvden: happy new year |
# 21:30:51 |
cpackham |
* bhundven: (new IRC client doesn't auto complete mentions) |
# 21:31:54 |
bhundven |
You too! It's been a while. Changed jobs. Moved (stuff is still in storage). It's been a wild ride. |
# 21:32:59 |
cpackham |
I've still got stuff in the garage from when I move 14 years ago |
# 21:33:49 |
bhundven |
I keep seeing the egrep warning... we need to move to `grep -E` to make configure happy |
# 21:34:24 |
cpackham |
Is `grep -E` in posix? (not that egrep is) |
# 21:35:05 |
bhundven |
I just checked on my mac (which uses bsd grep), and it is also extended grep |
# 21:36:15 |
cpackham |
We might want to drop ncurses 6.0 as that appears to have a patch that does a whole lot of egreping |
# 21:36:29 |
bhundven |
`AC_PROG_EGREP`: Check whether $GREP -E works, or else look for the best available egrep or gegrep that accepts the longest input lines possible. Set the output variable EGREP to whatever is chosen. The result can be overridden by setting the EGREP variable and is cached in the ac_cv_path_EGREP variable. |
# 21:37:29 |
bhundven |
homebrew (macos): $ which gegrep |
# 21:37:30 |
bhundven |
/opt/homebrew/bin/gegrep |
# 21:37:35 |
cpackham |
Is that not working for you? Or is it just errors in the packages that use egrep directly |
# 21:38:02 |
bhundven |
Hrm, might be upstream... in tools |
# 21:39:03 |
bhundven |
I always set `CT_LOG_SEE_TOOLS_WARN=y` |
# 21:40:18 |
cpackham |
We do seem to error out if `"$EGREP" != "$GREP -E"` |
# 21:40:37 |
bhundven |
Yea, it's probably an issue with upstream tools |
# 21:40:45 |
bhundven |
I just notice gcc complaining |
# 21:42:10 |
Trel |
zagura: I've been playing with buildroot and asked there, they're saying it doesn't do what you're saying you use it for, and from what I've tried that seems to be the case. I used my crosstools toolchain to build it, but absolutely nothing from it is included in the root it builds. |
# 21:42:29 |
Trel |
Did you do something custom to accomplish it? |
# 21:46:10 |
cpackham |
gcc seems inconsistent. some places use grep/egrep directly in configure.ac others have tests to check which grep flavors to use |
# 21:47:51 |
bhundven |
Got a new PR for ya |
# 21:48:21 |
bhundven |
update config.{guess,sub} |
# 21:48:33 |
bhundven |
it's been... 2 years |
# 22:26:12 |
bhundven |
cpackham: well, it's still reproducible with binutils 2.42 |
# 22:27:55 |
bhundven |
Added new pr for binutils-2.42 |
# 22:40:29 |
cpackham |
Snap https://github.com/crosstool-ng/crosstool-ng/pull/2095 |
# 22:41:39 |
cpackham |
There was a build failure on macos that might be something you can shed some light on. |
# 22:41:49 |
cpackham |
I got as far as reporting in upstream https://sourceware.org/bugzilla/show_bug.cgi?id=31320 |
# 22:43:44 |
cpackham |
GNU C Library version 2.39 was just released |
# 23:04:47 |
bhundven |
ah, hrm |
# 23:05:25 |
bhundven |
Bunitils 𤣠|
# 23:06:00 |
bhundven |
Oh, gold |
# 23:36:51 |
bhundven |
cpackham: Check out the comment in #2095 |
# 23:37:02 |
bhundven |
I also updated the bug on sourceware |
# 23:37:19 |
bhundven |
https://cplusplus.com/reference/cuchar/ |
# 23:54:48 |
bhundven |
I don't think binutils depends on gnulib currently |
# 23:56:17 |
bhundven |
Also, binutils already does a bunch of checks for `__APPLE__`, so I think my proposed change is "proper". |