[Tinycc-devel] Fwd: RE :Re: Recent change breaks test3 on RPi

[Thomas Preud'homme]

Forwarding here as the mail being in english indicates it was meant for a wide 
audience.

----------  Message transmis  ----------

Objet : RE :Re: [Tinycc-devel] Recent change breaks test3 on RPi
Date : mercredi 5 février 2014, 08:56:49
De : Christian JULLIEN <address@hidden>
 À : address@hidden <address@hidden>

Hi Thomas,

About signess, C standard permits an implementation to choose what it prefers.
To me, tcc should used the same sign as the default sign used by boostrap 
compiler (and of course tests should test the same)
For example, an ARM board A, for some good reason, can choose signed char 
while ARM board B can choose unsigned char by default.
In this case, the associated gcc compiler on those boards are certainly be 
configured with A or B default sign.

When boostraping tcc with gcc, the resulting tcc should have the same char 
sign as default for gcc options used for boostrap.
Hence, tcc configuration boostrap should ask the sign of char to bootstraping 
compiler (e.g. gcc but may also be clang or whatever).

tcc should also honnor -fsigned-char or -funsigned-char option (gcc 
compatible) to change char default sign.

Finally, tcc ./configure should include --char-sign=signed/unsigned to force a 
specific mode.

M2c

Christian 


----- Message d'origine -----
De : "Thomas Preud'homme" &lt;address@hidden&gt;
Date mer. 05/02/2014 07:15 (GMT +01:00)
À : "address@hidden" &lt;address@hidden&gt;, "Christian 
Jullien" &lt;address@hidden&gt;
Objet : Re: [Tinycc-devel] Recent change breaks test3 on RPi

On February 5, 2014 1:18:39 PM GMT+08:00, Christian Jullien 
&lt;address@hidden&gt; wrote:
&gt; Hi All,
&gt; 
&gt;  
&gt; 
&gt; This failure is new on RPI

[SNIP falling char sign test]

Actually it's not. What's new is that it's the only failure for this test now. 
The reason you notice it only now is because recently grischka made the diff 
between the output of tcctest when compiled with tcc and the output when 
compiled with gcc a failure. Before it just displayed the diff but without 
editing in error.

I would like to fix this but I need to see first why tcc and gcc differ. 
Daniel, 
if you read this, can you explain me why gcc seems to use a signed char while 
tcc uses an unsigned char?


Best regards,

Thomas
-----------------------------------------