Re: [Liberty-eiffel] Liberty-eiffel Digest, Vol 4, Issue 5

[Raphael Mack]

;-)

Hans, this sounds somehow similar to my situation. With one difference:
in my case I try to find the point in OUR compiler. I search to place
where to adjust the loading of the root class, to make the compiler take
the current working directory as start point to use the class loading
algorithm for the root class...

Cheers,
Rapha

Am Dienstag, den 17.09.2013, 18:50 +0200 schrieb H. Zwakenberg | Ocean
Consulting GmbH: 
>  ok, this is getting odd:
> 
>  in three tcc source files there are six occurences of a "memory full" 
>  message being emitted. Each of these strings I extended with 
>  corresponding filename and line number. After recompiling, I hoped to 
>  learn where it quits when compiling compile_to_c22.c.
> 
>  Much to my amazement, the same "memory full" message was generated, 
>  naked - without any of the filename and line number info I added.  I 
>  then verified that I indeed had only a single tcc.exe in my path and 
>  current dir.  There's only a single one, so it must be the newly 
>  compiled one...  To be certain of this, I deleted the tcc.exe file and 
>  verified there is none on my harddisk anymore.  Then I recompiled and 
>  repeated the tests, same results.
> 
>  Another assumption that I verified is that perhaps the preprocessor has 
>  issues.  So I started 'tcc -E', to let it only pre-process.  The output 
>  was piped to a file.  The preprocessor seems to work, no error messages 
>  there.  Compiling the preprocessed file then produces the same error 
>  message as before...
> 
>  I then started to grep the sources for just the word 'full', hoping 
>  that somewhere a string variable was given this value and that it got 
>  concatenated to another string containing 'memory' somewhere.  Alas, 
>  this brought no solution. Also, I grepped for all 'printf(' occurences, 
>  hoping that this might give me some insights or perhaps some further 
>  ideas.  Unfortunately, this didn't make me any smarter...
> 
>  To make a long short:  I haven't the foggiest where the offending error 
>  message is emitted.  As I'm only compiling and not linking, it can only 
>  be from the parser and compiler.  As I don't know where it gets emitted, 
>  I cannot set breakpoints at smart locations and try to find out what is 
>  going on here.  There is of course another method:  grabbing the 
>  preprocessed file and single step through it.  Anyone of you up to this 
>  task?  After all, it's just a 1.7 Mb file containing only about 46000 
>  lines... ;)  Really, I've got better things to do, like integrating 
>  PellesC....
> 
>  Speaking of which: I sent Pelle a message asking him for his formal 
>  o.k. to use his compiler as one of our backends.  I'll keep you posted.
> 
>  stumped,
>  Hans
> 
> 
>  Am Dienstag, den 17.09.2013, 12:00 +0200 schrieb 
>  address@hidden:
> > Send Liberty-eiffel mailing list submissions to
> >     address@hidden
> >
> > To subscribe or unsubscribe via the World Wide Web, visit
> >     https://lists.gnu.org/mailman/listinfo/liberty-eiffel
> > or, via email, send a message with subject or body 'help' to
> >     address@hidden
> >
> > You can reach the person managing the list at
> >     address@hidden
> >
> > When replying, please edit your Subject line so it is more specific
> > than "Re: Contents of Liberty-eiffel digest..."
> >
> >
> > Today's Topics:
> >
> >    1. TCC self-compiling on Windows
> >       (H. Zwakenberg | Ocean Consulting GmbH)
> >
> >
> > 
> > ----------------------------------------------------------------------
> >
> > Message: 1
> > Date: Tue, 17 Sep 2013 16:37:36 +0200
> > From: "H. Zwakenberg | Ocean Consulting GmbH" 
> > <address@hidden>
> > To: Liberty eiffel <address@hidden>
> > Subject: [Liberty-eiffel] TCC self-compiling on Windows
> > Message-ID:
> > 
> >     <address@hidden>
> >
> > Content-Type: text/plain; charset=UTF-8; format=flowed
> >
> >  Hi group,
> >
> >  a quick update: this afternoon I found a bit of time to adapt the
> >  scripts to allow TCC to compile itself on Windows. Better that than
> >  having to install yet another toolchain... That's done now, so the 
> > basis
> >  is there to try and find out why one of the compile_to_c sources is 
> > a
> >  problem for it...
> >
> >  cheers
> >  Hans
> >
> >
> >
> >
> > ------------------------------
> >
> > _______________________________________________
> > Liberty-eiffel mailing list
> > address@hidden
> > https://lists.gnu.org/mailman/listinfo/liberty-eiffel
> >
> >
> > End of Liberty-eiffel Digest, Vol 4, Issue 5
> > ********************************************
> 
>