X-Authentication-Warning: delorie.com: mail set sender to djgpp-bounces using -f
X-Recipient: djgpp AT delorie DOT com
Subject: Re: DJGPP v2.05 and Windows 10
To: djgpp AT delorie DOT com
References: <562BCDB5 DOT 8000402 AT iki DOT fi>
 <CAB9Rao_2QazAWzGWU5nxVHzCQvXi7n4Ni2A008VGvHq9UuKXVg AT mail DOT gmail DOT com>
From: "Andris Pavenis (andris DOT pavenis AT iki DOT fi) [via djgpp AT delorie DOT com]" <djgpp AT delorie DOT com>
Message-ID: <562DBA23.9040506@iki.fi>
Date: Mon, 26 Oct 2015 07:29:07 +0200
User-Agent: Mozilla/5.0 (X11; Linux x86_64; rv:38.0) Gecko/20100101
 Thunderbird/38.3.0
MIME-Version: 1.0
In-Reply-To: <CAB9Rao_2QazAWzGWU5nxVHzCQvXi7n4Ni2A008VGvHq9UuKXVg@mail.gmail.com>
Content-Type: text/plain; charset=utf-8; format=flowed
Content-Transfer-Encoding: 7bit
Reply-To: djgpp AT delorie DOT com
Errors-To: nobody AT delorie DOT com
X-Mailing-List: djgpp AT delorie DOT com
X-Unsubscribes-To: listserv AT delorie DOT com
Precedence: bulk

On 10/24/2015 09:46 PM, Louis Santillan (lpsantil AT gmail DOT com) [via djgpp AT delorie DOT com] wrote:
> Andris, is it possible that you have a bad memory chip?  Maybe
> something in 4.x, 5.x, 6.x gcc grows the memory image into the bad
> pages while 3.x does not and/or does utilize (read/write) those bad
> pages.

I do not think that bed memory could be a reason. I would have noticed it log time ago
in Linux (all mentioned tests were done in Win 10 VM under Fedora 22 x86_64). I have seen
random compilation errors in Linux on a different  system which was caused by RAM problems
some  years ago. Never seen anything like that on this system.

Also
- initial impression is that gcc-4.4.7 is OK and works under Win10 (bootstrapped it under Win10
using gcc-3.4.6 without problems)
- some changes in environment (update VirtualBox 4.3 to VirtualBox-5.0, change of DJGPP installation
to have more than 1 gcc version installed at the same time with additional bin directories with 
different
names) causes symptoms to change. Now I have one reproducible NTVDM crash I can workaround by 
increasing
gcc garbage collection threshold (could be caused by memory corruption, so avoiding GC could avoid 
crash)

Andris

>
> On Sat, Oct 24, 2015 at 11:28 AM, Andris Pavenis
> (andris DOT pavenis AT iki DOT fi) [via djgpp AT delorie DOT com] <djgpp AT delorie DOT com>
> wrote:
>> Did some additional tests on WIndows 10.
>>
>> I'm getting random compile failures with gcc-4.9.3, gcc4.5.2 and gcc 6.0.0
>> 20151018 experimental (last development version I built for DJGPP).
>> Running
>>
>> echo '#include <iostream>" | gcc -E -dD -x c++ -
>>
>> in loop and comparing outputs is sufficient for detecting random corruption.
>> I have not succeeded to find why it happens and how to workaround that.
>> Additionally I'm also getting random failures when trying to repeatedly
>> compile same preprocessed C++ file.
>>
>> The situation is however not completely hopeless:
>> Installed gcc-3.4.6 and bootstrapped the same gcc version (3.4.6) without
>> problems under Windows 10. No random gcc failures observed.
>>
>> So some difference (or differences) between gcc-3.4.6 and recent version
>> cause problems.
>> Some suspects:
>> - change to C++ as implementation language
>> - noticeable increase of size of executables and resources required by them
>>
>> Andris
>>
>>
>>
>>
>>