# 00:17:59 |
cpackham |
quits : Ping timeout: 264 seconds |
# 00:20:06 |
cpackham |
joins #crosstool-ng |
# 01:23:37 |
cpackham |
quits : Ping timeout: 255 seconds |
# 01:37:57 |
cpackham |
joins #crosstool-ng |
# 04:05:29 |
hays |
quits : Remote host closed the connection |
# 04:26:02 |
hays |
joins #crosstool-ng |
# 04:27:08 |
hays |
quits : Client Quit |
# 04:29:00 |
hays |
joins #crosstool-ng |
# 04:38:16 |
cpackham |
quits : Ping timeout: 245 seconds |
# 05:09:58 |
Net147 |
quits : Ping timeout: 255 seconds |
# 05:19:55 |
Net147 |
joins #crosstool-ng |
# 05:19:57 |
Net147 |
quits : Changing host |
# 05:19:57 |
Net147 |
joins #crosstool-ng |
# 05:53:01 |
Net147 |
quits : Ping timeout: 276 seconds |
# 10:14:59 |
Net147 |
joins #crosstool-ng |
# 10:15:01 |
Net147 |
quits : Changing host |
# 10:15:02 |
Net147 |
joins #crosstool-ng |
# 10:23:35 |
Net147 |
quits : Read error: Connection reset by peer |
# 10:25:56 |
Net147 |
joins #crosstool-ng |
# 10:25:59 |
Net147 |
quits : Changing host |
# 10:25:59 |
Net147 |
joins #crosstool-ng |
# 10:33:55 |
Net147 |
quits : Quit: Quit |
# 10:34:11 |
Net147 |
joins #crosstool-ng |
# 10:34:13 |
Net147 |
quits : Changing host |
# 10:34:13 |
Net147 |
joins #crosstool-ng |
# 10:47:20 |
Net147 |
quits : Read error: Connection reset by peer |
# 10:50:17 |
Net147 |
joins #crosstool-ng |
# 10:50:20 |
Net147 |
quits : Changing host |
# 10:50:20 |
Net147 |
joins #crosstool-ng |
# 10:56:26 |
Net147 |
quits : Ping timeout: 252 seconds |
# 11:05:05 |
Net147 |
joins #crosstool-ng |
# 11:05:08 |
Net147 |
quits : Changing host |
# 11:05:08 |
Net147 |
joins #crosstool-ng |
# 19:25:44 |
zagura |
quits : Ping timeout: 252 seconds |
# 19:26:07 |
zagura |
joins #crosstool-ng |
# 19:35:32 |
cpackham |
joins #crosstool-ng |
# 19:42:18 |
Guest59 |
joins #crosstool-ng |
# 19:45:14 |
Guest59 |
Hi there, I wanted to compile crosstool-ng on my vm (Ubuntu Linux on x86_64) so that I can comple MAME for Raspberry Pi (arm). After quite some time compiling, the following warning is generated: |
# 19:45:14 |
Guest59 |
[ERROR]Â Â Â /opt/pi/mame_raspberrypi_cross_compile/build/ctng_rpi2_buster_armhf/.build/armv7-rpi2-linux-gnueabihf/src/binutils/gold/x86_64.cc:1590:5: error: narrowing conversion of 'elfcpp::GNU_PROPERTY_LOPROC' from 'unsigned int' to 'int' [-Wnarrowing] |
# 19:45:15 |
Guest59 |
In the meantime, I think I understand what that warning means, but I have no clue on how to resolve that. Any ideas? Is this a know issue? |
# 19:56:43 |
milkylainen |
It's a type of sign unsign warning. You're probably converting values the compiler knows won't fit in their unsigned representation in an int. |
# 20:31:21 |
Guest59 |
quits : Quit: Client closed |
# 20:45:59 |
milkylainen |
Would have been nice with some more information though... |
# 20:46:04 |
milkylainen |
Oh well. |
# 21:07:14 |
cpackham |
Based on that one line it looks like it was actually the ct-ng build that erroring out. |
# 21:07:57 |
cpackham |
I'd have suggested looking for a binutils fix or maybe using bfd instead of gold |
# 21:08:27 |
milkylainen |
cpackham: Yes the error came from a ct-ng printout, but the actual error came from picky compiler settings for gold? |
# 21:09:35 |
cpackham |
I didn't think we built binutils with -Werror but again that's something that could have been checked |
# 21:10:23 |
milkylainen |
Yeah, I should have been more responsive. |
# 21:10:25 |
milkylainen |
:\ |
# 21:10:44 |
cpackham |
I must dust of my gold patch series. I don't think it'd have helped with that issue but might get more coverage out of our builds |
# 21:12:49 |
milkylainen |
Your opinion, is it mature enough these days? I remember trying it a long time ago, but it wasn't trouble free enough for everything I used it for. Or did it suffer less attention? |
# 21:13:03 |
milkylainen |
I don't use gold at all really. |
# 21:14:49 |
cpackham |
I'm not really sure. gold has been the "new" linker for so long the I'm not sure what the distinction between it and bfd is |
# 21:15:07 |
cpackham |
I have seen various people say gold is faster |
# 21:15:25 |
cpackham |
but I've not seen any data to actually back that up |
# 21:16:10 |
cpackham |
from a ct-ng perspective I'd like to be able to offer both (which we do but not by default) |
# 21:16:48 |
milkylainen |
Yeah. I try to avoid the typical omg-c++bloblink so I guess I'm not reaping much of the benefits. |
# 21:17:39 |
milkylainen |
I think it was very large things that annoyed people? linking time in minutes or so? |
# 21:19:08 |
milkylainen |
I remember a client doing "we don't need dynamic loading binaries. Lets go static c++ 600M binaries!" |
# 21:19:11 |
milkylainen |
:D |
# 21:21:56 |
milkylainen |
btw. what happened to the latest binutils bump? |
# 21:27:21 |
Guest59 |
joins #crosstool-ng |
# 21:28:57 |
Guest59 |
quits : Client Quit |
# 21:46:47 |
cpackham |
Binutils bump went in |
# 21:47:29 |
cpackham |
Had to tweak how we build elf2flt because we don't properly install bintuils for the host |
# 21:49:16 |
milkylainen |
elf2flt. Haven't used it for ages. |
# 21:50:42 |
milkylainen |
last time was some old 50Mhz OKI cpu running nommu linux. |
# 21:52:53 |
milkylainen |
Would be interesting to hear userstories of various projects still using binflats. |
# 21:55:23 |
cpackham |
I've never used it personally |