Running emerge -u dev-db/mysql -up shows: # emerge -up dev-db/mysql These are the packages that I would merge, in order: Calculating dependencies ...done! [ebuild U ] dev-db/mysql-3.23.56 [3.23.54a] ..... checking if conversion of longlong to float works... no configure: error: Your compiler cannot convert a longlong value to a float! If you are using gcc 2.8.# you should upgrade to egcs 1.0.3 or newer and try again !!! ERROR: dev-db/mysql-3.23.56 failed. !!! Function econf, Line 273, Exitcode 1 !!! econf failed Seems that the configure file in /var/tmp/portage/mysql-3.23.56/work/mysql-3.23.56 is assuming at line 16914 that I need to do a check for Sparc and gcc 2.8.# gcc on this system is at version 3.2.2 I don't know if this is a stupid noobe error or what? Reproducible: Always Steps to Reproduce: 1. 2. 3. Actual Results: checking if conversion of longlong to float works... no configure: error: Your compiler cannot convert a longlong value to a float! If you are using gcc 2.8.# you should upgrade to egcs 1.0.3 or newer and try again !!! ERROR: dev-db/mysql-3.23.56 failed. !!! Function econf, Line 273, Exitcode 1 !!! econf failed Expected Results: mysql should be upgraded I tried this a variety of times. failed each. some from console... nothing else running... some from kde... same results every time.
please post /var/tmp/portage/mysql-3.23.56/work/mysql-3.23.56/config.log as an attachment
from the user: configure:16912: checking if conversion of longlong to float works configure:16939: gcc -o conftest -O3 -DDBUG_OFF -rdynamic conftest.c -lpthread -lz -lcrypt -lnsl -lm -lpthread >&5 /usr/lib/gcc-lib/i686-pc-linux-gnu/3.2.2/../../../libpthread.a(pthread.o)(.text+ 0x136): In function `__pthread_initialize_minimal': : undefined reference to `_dl_cpuclock_offset' /usr/lib/gcc-lib/i686-pc-linux-gnu/3.2.2/../../../libpthread.a(pthread.o)(.text+ 0x13b): In function `__pthread_initialize_minimal': : undefined reference to `_dl_cpuclock_offset' collect2: ld returned 1 exit status configure:16942: $? = 1 configure: program exited with status 1 configure: failed program was: #line 16922 "configure" #include "confdefs.h" #include <stdio.h> typedef long long longlong; main() { longlong ll=1; float f; FILE *file=fopen("conftestval", "w"); f = (float) ll; fprintf(file,"%g\n",f); close(file); exit (0); } configure:16966: result: no configure:16973: error: Your compiler cannot convert a longlong value to a float! If you are using gcc 2.8.# you should upgrade to egcs 1.0.3 or newer and try again
could you please post 'emerge info' ?
Created attachment 11338 [details] emerge info output This is the requested output from "emerge info" on the failing system.
Created attachment 11339 [details] emerge info output (dup of 11338 - ignore!) This is the requested output from "emerge info" on the failing system.
Comment on attachment 11339 [details] emerge info output (dup of 11338 - ignore!) I would delete this if there was a "delete" buttom somewhere.
could you please do: 'emerge binutils glibc' then try to emerge mysql again.
Kurt: any updates on this?
Robin I tried the 'emerge binutils glibc'. It worked just fine... then tried both 'emerge -u mysql', as well as 'emerge mysql'.... Same failure. Any idea what I am suffering from? Am I the only one??? Thanks Kurt
Could you please try some of the steps in bug #16707. It seems that is the same thing. Namely: USE='-java' emerge gcc
I went and read the 16707 bug. I don't see any similarity. but... I ran the: USE='-java' emerge gcc This worked just fine. then i ran emerge -u mysql same failure then I ran USE='-java' emerge -u mysql same failure The 16707 bug was in reference to '_dl_cpuclock_offset', mine is in reference to longlong. I hate to be dense, but I don't see the corelation.
Kurt: the config.log that you posted for vapier showed that it was the '_dl_cpuclock_offset' that was causing the configure to give the weird error about longlong.
Robin Thanks for the swinging a 2x4 my way <grin> I see that now... I was way to focused on the longlong anyway.... A bit more investigation now that I have been reminded to READ! The section that is failing has a comment that preceeds it that indicates it is a check for a sparc using gcc 2.8.# So, I bit reformatting so I can tell where the "if"s are closed and I comment out the section and re-run ./configure from the vartmp/portage/mysql-3.23.56/work/mysql-3.23.56 directory. This seems to make it all the way through the process. Problem is, if I re-run emerge, I suspect that all my modifications to configure are going to disappear. Also of note from reading the code a bit. The error (_dl_cpuclock_offset) seems to be coming from "-lpthread" which is imbedded in the gcc command twice. Removing this and re-running the code fragment for longlong works just fine. So, I seem to be suffering from 1. possession of a pthread library 2. having a sparc sibling that I need to check on 3. insufficent knowledge of the internals for the emerge process of building, compiling, sandboxing, upgrading, etc... I'm somewhat tempted to go find the pthread library and rename it... but I suspect that wouldn't really help. Oh yeah... in case you are curious what the gcc line looks like by the time it gets around to the longlong to float test.... here it is. gcc -o conftest -O3 -DDBUG_OFF -march=pentiumpro -pipe -DHAVE_ERRNO_AS_DEFINE=1 -O3 -rdynamic conftest.c -lpthread -lz -lcrypt -lnsl -lm -lpthread Curious and learning.... Thanks as always Kurt L Vanderwater
I really can't figure out why -lpthread twice is causing problems. Could you please try: 'emerge -D binutils gcc glibc binutils gcc', then 'emerge mysql'. I know I have gcc/binutils in the first line twice, I want to see something about cross-linking between them. When the mysql emerge fails, please attach the entire config.log file here again.
Kurt: any progess on this?
On bugs.gentoo.org I've seen problems with dl_cpuclock_offset for php,mysql and now I've got problems with avifile....
Kurt: I'm strongly inclined to suspect deeper problems with your system somewhere, and I'm not sure where. Please try upgrading to glibc-2.3.2-r1 and recompiling, other people have seen this fix similar problems.
is this still a problem ?
no response from user.