Need to find a 5 year old or older version of gcc compiler that will run on windows.

mpvmud · 06-25-2013, 12:00 AM

Like my title says. I am trying to compile a code that is about 5 years out of date for current compilers. Last time I compiled it was in about 2009. Since then gcc has been updated to the point that a once compilable code no longer compiles now.

Any help in trying to find a compiler I can use for this would be great help. I have spent a good 3 days on bing and google trying to find anything and has ended in failure.

Thanks in advance for your help!

Cheers.

NevemTeve · 06-25-2013, 03:17 AM

Off: is the problem with the code so serious? Would you mind quoting the error message?

theNbomr · 06-25-2013, 10:12 AM

The 'releases' directory of the GNU gcc site mirrors has all old versions dating back to the nineties, and version 1.1.2 It would always be the preferred location for downloading GCC. Of course you'll need some working compiler to build the version you choose.

When compilers get smarter, they usually help us find problems with the code we compile with them. Wouldn't it be a good idea to use that to fix the problems identified by the newer compilers?

--- rod.

gnashley · 06-25-2013, 11:08 AM

I'd look in the old releases at mingw:
http://sourceforge.net/projects/ming...inGW/Base/gcc/

Version 3 should be the right one for you.

John VV · 06-25-2013, 05:10 PM

Mingw dose have the Ancient gcc 3.2 in the archives

Why not just hack the code to bring it up to the 20'th century specs

as i recall the current stable in Mingw is gcc 4.1 with gcc 4.6 in the testing section

mpvmud · 06-25-2013, 06:04 PM

Quote:

Originally Posted by NevemTeve

Off: is the problem with the code so serious? Would you mind quoting the error message?

pal.c:116: error: expected â)â before â*â token
make: *** [pal.o] Error 1

not sure why it is making the strange ascii letters but I have checked the code and it should be fine. The line is:

Quote:

shield_rush (CHAR_DATA * ch, char *argy)
{
CHAR_DATA *victim;
SINGLE_OBJECT *obj;
SPELL_DATA *spl;
int chan;
bool notft = FALSE;
int tmpi;
char buf[STD_LENGTH];
int dam;
int damm;
DEFINE_COMMAND("rush", shield_rush, POSITION_FIGHTING, 0, LOG_ALWAYS, "Shield Rush")
check_fgt (ch);
chan = 0;

So I do not know why this line of code out of all the code is acting up. Also in my header file (if I remove pal.c and run make) I get warnings and errors on code that has worked fine for almost 20 years.

Quote:

gcc -c -g -mtune=i486 -Wno-unused -DLinux -DNEED_EXTERN -DOLD_FIGHT autoeq.c
In file included from autoeq.c:9:
emlen.h:3192: warning: built-in function âpowâ declared as non-function
emlen.h:4398: error: array type has incomplete element type
In file included from autoeq.c:9:
emlen.h:4677: warning: conflicting types for built-in function âexeclpâ
emlen.h:4678: warning: conflicting types for built-in function âexeclâ
make: *** [autoeq.o] Error 1

So I am by far confused by why the code is just not working and I am by far not an expert in code but this code has not been touched since at least 1994. I have worked on things I have added but it has nothing to do with any of the errors or warnings coming up.

sundialsvcs · 06-25-2013, 07:34 PM

Quote:

Originally Posted by theNbomr

When compilers get smarter, they usually help us find problems with the code we compile with them. Wouldn't it be a good idea to use that to fix the problems identified by the newer compilers?

That's definitely the way that I would suggest "dealing with" this issue. Perhaps there are compiler options or directives (which maybe didn't exist at that time) which would resolve any issues. Or, definitely, it might be the case that there is a latent problem with this software which would warrant more careful examination.

I would strongly suggest that you use the git version-control system at this point: set up the directory as a "git repository" and "check in" the software exactly as it is right now. Then, make a "branch" and now start making changes that are necessary to bring the code up to snuff. If you're using "git" in this way, you can always at any time (a) revert the code back to any previous state, and (b) identify the precise line-by-line changes that you have actually made. ("git" runs equally well on both systems and does not require a server.)

Although it does demand more work and investigation of a piece of software that you might not be too keen to get deeply involved in, I do think that you will be more satisfied with the final results.

frieza · 06-25-2013, 08:09 PM

Quote:

Originally Posted by sundialsvcs

Although it does demand more work and investigation of a piece of software that you might not be too keen to get deeply involved in, I do think that you will be more satisfied with the final results.

I have to second that, as the errors you are seeing could be just a newer compiler flagging something incorrect that the older version missed, or it could be that something went awry in storage, either way bringing the code up to modern standard is probably the best option.

John VV · 06-25-2013, 08:32 PM

Quote:

pal.c:116

in the section ending on line 116 is a issue

however there is a second error in a header

Quote:

In file included from autoeq.c:9:
emlen.h:3192: warning: built-in function âpowâ declared as non-function
emlen.h:4398: error: array type has incomplete element type
In file included from autoeq.c:9:
emlen.h:4677: warning: conflicting types for built-in function âexeclpâ
emlen.h:4678: warning: conflicting types for built-in function âexeclâ
make: *** [autoeq.o] Error 1

" -- emlen.h:4398: error: array type has incomplete element type -- "
that array error would also cause the unexpected token on line 116

a question on this
" -mtune=i486"
are you using a computer from 1990 - 1994 ? and is it really an almost antique 486

if so then for 15+ year old code for a 15 to 20 year old box
gcc 2 might be needed

NevemTeve · 06-26-2013, 03:54 AM

Well if it is non-secret, you might upload it somewhere...