g_assertions

[John Emmas]

Thanks Dave, you put me on the right track.

This turned out to be a glitch with my development IDE (Code::Blocks).
g_assert() ends up calling g_log() which requires a pre-processor definition
called G_LOG_DOMAIN.  In my case, G_LOG_DOMAIN needed to be a string -
namely, "glibmm" (including the quote marks).

To set this up for my gcc compiler, the IDE needs to send this:-

g++ [various options] -DG_LOG_DOMAIN=\"glibmm\"

However, that doesn't seem to work for cygwin.  I need to set it up slightly
differently - namely:-

g++ [various options] -DG_LOG_DOMAIN="\"glibmm\""

So, probably not a problem with Cygwin.  More likely a problem with
Code::Blocks.  I probably need to flag it up to them.

Thanks again,

John



----- Original Message ----- 
From: "Dave Korn"
Sent: 22 September 2008 13:39
Subject: RE: g_assertions


> John Emmas wrote on 22 September 2008 13:55:
>
>> Thanks Phil,
>>
>> That sounds very plausible but I'm not sure it's right in this case.
>
>  I'm pretty sure of it.  The only thing that can cause that compiler error
> is
> characters between the backslash and the newline.
>
>> I
>> only have 2 copies of glib.h on my system and I'm pretty sure they came
>> with the Cygwin installation.  I looked at both of them in a hex editor
>> and they seem to have Unix style line-endings.
>
>  I can't make your testcase compile here.  You didn't show the command
> line
> but without some kind of -I option it's not going to even find glib.h, and
> without you showing the command line we don't know which one or from
> where.
>
>  The next step in investigating would be to compile it again, adding the
> "--save-temps" flag, so you get a copy of the preprocessed output in a .i
> file.  Then check /that/ for EOLs of the wrong type, maybe using something
> like
>
> tail -20 *.i | od -c
>
> and looking out for "\r"s.  (I chose "-20" as a guess, just to get the
> pre-processed version of my_func and skip out most of the #include headers
> and
> other irrelevancies that end up in the preprocessed output).
>
>    cheers,
>      DaveK
> -- 
> Can't think of a witty .sigline today....
>
> 


--
Unsubscribe info:      http://cygwin.com/ml/#unsubscribe-simple
Problem reports:       http://cygwin.com/problems.html
Documentation:         http://cygwin.com/docs.html
FAQ:                   http://cygwin.com/faq/