[avr-libc-dev] RE: [avr-gcc-list] Find problem with gcc patching libc

[Weddington, Eric]

I thought that fixincludes in gcc was completely disabled for the AVR
target. If not, then it should be. Fixincludes does not need to change
anything in avr-libc as it will probably just mess things up. So I would
suggest looking into that area and make sure it is disabled for the AVR.

Eric
 

> -----Original Message-----
> From: Andy H [mailto:address@hidden 
> Sent: Saturday, June 07, 2008 10:21 AM
> To: Weddington, Eric
> Cc: Mike Stein; address@hidden; address@hidden
> Subject: Re: [avr-gcc-list] Find problem with gcc patching libc
> 
> We seem to be both using 4.4 head and same configuration 
> switches. Though different directories perhaps.
> 
> I have just noted that stdio.h in 1.6.2 is version 1.29 - 
> where Mike is using 1.30. So I assume he is using HEAD 
> version of libc??
> 
> I will thus check this version to see if it triggers problem.
> 
> I started at almost same time I added AVR target test. Which 
> failed because of this issue. However, I was able to see same 
> problem just before that on some other test.
> Any tests with stdio.h will thus fail (on Mike's setup)
> 
> I think changes fixincludes are making are somewhat logical. 
> Maybe Cygwin and Debian are wrong?
> 
> I am not sure what determines if file is fixed or not. The 
> gcc code looks for missing functions and adds missing 
> definitions. But what causes that to happen (or not) is a 
> mystery to me.
> Debian shows fixing messages for files prior to test. Cygwin 
> does not. Just to confuse the issue!
> 
> 
> Andy
> 
> 
> Weddington, Eric wrote: 
> 
>        
>       
>         
> 
>               -----Original Message-----
>               From: 
>               
> address@hidden 
>               
> [mailto:address@hidden
>               org] On Behalf Of Andy H
>               Sent: Saturday, June 07, 2008 9:26 AM
>               To: Mike Stein; address@hidden; 
> address@hidden
>               Subject: [avr-gcc-list] Find problem with gcc 
> patching libc 
>               
>               I was looking for cause of problem with Mike 
> Steins tests 
>               setup, where 
>               stdio.h gets patched by gcc.
>               This defines unused file functions but assume fpos_t is 
>               defined. Thus we 
>               get compilation failures.
>               
>               My cygwin and Debian linux setups dont have 
> this problem.
>               
>               I have rebuilt my Debian setup with clean 
> binutils, gcc and 
>               libc today 
>               from scratch so I now have 1.6.2 libc (same as Mike)
>               
>               Despite this  I cannot replicate problem at 
> all.  Any ideas?
>               
>               I am re-running gcc testsuite and hopefully the 
> result will 
>               be the same 
>               as 1.6.1 (or better)!
>               
>               Mike, you may want to try clean build. This 
> problem only occurred 
>               recently, and was not connected to any patches.
>                   
> 
>       
>       Since this is GCC's fixincludes that is doing the 
> patching (I assume),
>       then what version of GCC are you using, Mike, that is 
> causing the
>       problem? Andy, what version of GCC are you using that 
> is NOT doing the
>       patching?
>         
> 
>