From: "Lewis" Received: from [192.168.100.201] (account lgrosenthal@2rosenthals.com HELO [192.168.100.23]) by 2rosenthals.com (CommuniGate Pro SMTP 5.4.10) with ESMTPSA id 939728 for cwmm-dev@2rosenthals.com; Sat, 01 Aug 2020 23:14:33 -0400 Subject: Re: [cwmm-dev] Compiling cwwm To: CWMM Developers Mailing List References: Message-ID: <5F262F97.9090901@2rosenthals.com> Date: Sat, 1 Aug 2020 23:14:31 -0400 User-Agent: Mozilla/5.0 (OS/2; Warp 4.5; rv:38.0) Gecko/20100101 Firefox/38.0 SeaMonkey/2.35 MIME-Version: 1.0 In-Reply-To: Content-Type: text/plain; charset=UTF-8; format=flowed Content-Transfer-Encoding: 7bit Hi, Dave... On 08/01/20 10:47 pm, Dave Yeo wrote: > On 07/31/20 11:45 PM, Steven Levine wrote: >> In , on 07/31/20 >> at 10:28 PM, "Dave Yeo" said: >> >> Hi Dave, >> >>>> WPTYPES.IDL:23 >>>> typedef unsigned long ULONG; >> >>> OK, I've installed the VACPP 3.08 including all fixes that Lewis kindly >>> donated, I still see in G:\IBMCPP\INCLUDE\OS2\WPTYPES.IDL having at line >>> 23, typedef int BOOL; >> >> My advice is make the G:\IBMCPP\INCLUDE\OS2\ directory disappear and >> forget it existed. The same is true for any other SOM components in >> \IBMCPP that have been superceded by newer components in the Toolkit. > > OK. My original env script put the 4.52 toolkit ahead of VACPP, which I'd > think did the same thing. > (Guessing: Unless something INCLUDEd VACPP ahead of it.) >> >> I apologize for taking so long to recall why this directory can exist. It >> took some review of the VAC VACPP.PKG Software Installer Package file to >> revive my memory. It's been over 20 years since I last had to think about >> this directory. In software installer terms, the content of >> IBMCPP\INCLUDE\OS2\ is part of >> >> COMPONENT >> NAME = 'Warp Toolkit Headers & Libraries', >> ID = 'WARPHLIB', >> >> This component existed because back in 1995 there was not yet a separately >> delivered toolkit. When the toolkit appeared, I uninstalled the VAC >> WARPHLIB component because to my form of common sense, keeping it around >> could only cause problems. Also back in the 90's disk space was a bit >> more of a scarce commodity. >> >> The same applies to the VAC SOM Component >> >> COMPONENT >> NAME = 'SOM DLL', >> ID = 'SOMDLL', >> >> It too got uninstalled when the separately installed toolkit appeared. >> >> The same thing applies to any copies of SOM .IR files that duplicate >> copies in %_BOOT:\OS2\ETC. They sould be gone. Nothing good can come >> from trying to use them. > > OK. It did take a bit for me to remember to edit the config.sys before > rebooting to make sure the VACPP stuff wasn't at the head of LIBPATH, PATH > etc. First try resulted in a pure blue screen where only CTRL-ALT-DEL > worked to reboot :) > >> >> This reminds me of an interesting feature of the VAC installer. Left to >> its own devices, it would backlevel the SOM DLLs on the boot drive with: >> >> FILE >> EXITWHEN = '"%TARGET%" == REMOTE', >> EXITIGNOREERR = 'YES', >> EXIT = 'REPLMOD %EPFIBOOTDRIVE%:\IBMCPP\DLL\SOM.DLL' >> >> and so on. Like I said before, nothing good could come of this. > > This didn't happen here luckily. All \os2\dll files are dated newer. > >> >> Review SOM\LIB in your tooklit and make sure you are using the >> backleveled: >> >> Toolkit\SOM\LIB\EMITC.DLL >> 4-27-98 12:35 49,425 0 EMITC.DLL >> >> 280973a9abe88dd8dbfdd57343ebff64 *EMITC.DLL >> >> Lars discovered some defects in the EMITC.DLL that shipped with the 4.5.2 >> toolkit. The fix is to use the EMITC.DLL from the 4.0 tooklit. It has >> less defects. > > Which leads to, where to get the 4.0 toolkit? > Stand by. >> >> Now that we have gone over all this, it's now obvious to me why you have >> >> G:\IBMCPP\INCLUDE\OS2\WPTYPES.IDL having at line 23, >> typedef int BOOL; >> >> It's a defect in the version of WPTYPES.IDL that shipped with VAC. I >> assume that the version of SC.EXE that shipped with VAC accepted accepted >> this usage without complaining because it did not understand that this was >> a defect. Someone must have noticed/reported this and the defect was >> fixed in the toolkit versions of SC.EXE and WPTYPES.IDL and part of the >> fix was to report the defective usage of int. >> > > So which toolkit to use? As noted earlier, I was running into problems > with a Warp V4 define. > I was hoping to reproduce Chris's environment to start with. Unluckily he > didn't really follow the rules and there's those hand written hh files > Yes, it seems that Chris had a setup which worked for him but which he did not document for others. So it goes. As we know, he didn't start out making a collaborative project at Netlabs, so we're sanding down the square corners to get his square peg to fit in a round hole. PM coming soon. -- Lewis