EQOffline:BotCount issue I think

Angelox · #1 01-14-2009, 10:16 AM

I tried 270 out and it works fine too - what are you using for compile? windows or Linix? and if windows , what version VC?

cubber · #2 01-14-2009, 10:36 AM

I don't own a windows box, all linux here. I will try pulling down clean sources and recompiling. I did a make clean before recompiling this version. Then when I found I should recompile world with -DEQBOTS in the makefile I just did a make clean in the world directory and recompiled that. Then copied world to my server directory.

cubber · #3 01-14-2009, 11:11 AM

I just re-ran make clean on my svn sources then did a make and rebuilt everything. This did the trick. Seems you can't just change something in the world makefile and remake just world, need to do it all.

So problem was between the seat and keyboard sorry guys!

**Derision** · #4 01-14-2009, 11:39 AM

It's not actually your fault, it is a problem with the makefiles.

Both the makefile.common for world and zone specify ../common/rulesys.o in the SF= line.

../common/rulesys.o gets compiled for world first, then, when zone gets compiled, it sees that ../common/rulesys.o is newer than ../common/rulesys.cpp and so it knows it doesn't need to recompile it.

Normally, this is not a problem, however in the case of code bracketed in #ifdef/#endif, where different defines are used between world and zone, it causes problems.

Instead of ../common/rulesys.o, the world and zone makefile.common should specify .obj/rulesys.o

This tells the makefile to compile it's own private version using it's own set of -D flags.

I'll update the makefiles.

The EQOffline:BotCount rule should really be a RULE_INT (integer) rather than a RULE_REAL (floating point).

cubber · #5 01-14-2009, 12:19 PM

Thanks for the clarification Derision.

Angelox · #6 01-14-2009, 01:13 PM

Quote:

Originally Posted by Derision

The EQOffline:BotCount rule should really be a RULE_INT (integer) rather than a RULE_REAL (floating point).

I'll get this fixed - I just at first didn't have any makefiles for with the WORLD makefile, and was going crazy trying everything, tell I realized what it was.

I've been trying to post replies on these forums, but for seems it goes dead at night for hours and now seems to drop a lot during the day too. Last night I couldn't get on here to reply, finally I just went to bed.

Angelox · #7 01-14-2009, 01:26 PM

Derision; I also noticed, it seemed at one time, when I did a 'make clean' from the source root, it would go through all the subs and include a 'make depend' , it doesn't seem to do that anymore, just includes 'make clean' only.
I wonder if something got removed?

Quote:

Originally Posted by Derision

It's not actually your fault, it is a problem with the makefiles.

Both the makefile.common for world and zone specify ../common/rulesys.o in the SF= line.

../common/rulesys.o gets compiled for world first, then, when zone gets compiled, it sees that ../common/rulesys.o is newer than ../common/rulesys.cpp and so it knows it doesn't need to recompile it.

Normally, this is not a problem, however in the case of code bracketed in #ifdef/#endif, where different defines are used between world and zone, it causes problems.

Instead of ../common/rulesys.o, the world and zone makefile.common should specify .obj/rulesys.o

This tells the makefile to compile it's own private version using it's own set of -D flags.

I'll update the makefiles.

**Derision** · #8 01-14-2009, 05:31 PM

Quote:

Originally Posted by Angelox

Derision; I also noticed, it seemed at one time, when I did a 'make clean' from the source root, it would go through all the subs and include a 'make depend' , it doesn't seem to do that anymore, just includes 'make clean' only.
I wonder if something got removed?

I don't know. The oldest source I have readily to hand on my Linux server is 0.7.0-1035 and it seems to exhibit the same behaviour as the current makefiles.

Angelox · #9 01-14-2009, 09:54 PM

I just started to make this post pointing out it was working in EQEmu-0.7.0-1129, but not now, and I saw why it doesn't work;
when someone uploaded the source to SVN, they also uploaded their .depend files , the makefile checks to see if there is .depend files, if not it will auto make them. So, you can quickly run a 'make clean' from the root, all that will happen and new custom .depend files are created. If there already are .depend files, then you have to force a 'make depend' in each directory.

Quote:

Originally Posted by Derision

I don't know. The oldest source I have readily to hand on my Linux server is 0.7.0-1035 and it seems to exhibit the same behaviour as the current makefiles.

EDIT:
That problem was me, I must have at one time run 'make clean' in my local SVN , and it created the .depend files - since it's on another PC with another version of my distro, it made the problems for me. I deleted the .depend files from SVN, it works again.