From: "A. Jans-Beken" Newsgroups: comp.os.msdos.djgpp Subject: Re: Array Addressing Bug in DJGPP 2.02 with GCC 2.81 ??? Date: Thu, 31 Dec 1998 13:25:28 +0100 Organization: World Access Lines: 45 Message-ID: <368B6D38.2C66A48@wxs.nl> References: <767k28$e46$1 AT latinum DOT dresearch DOT de> NNTP-Posting-Host: vl0210-0.dial.wxs.nl Mime-Version: 1.0 Content-Type: text/plain; charset=us-ascii Content-Transfer-Encoding: 7bit X-Trace: reader1.wxs.nl 915107525 24596 195.121.18.210 (31 Dec 1998 12:32:05 GMT) X-Complaints-To: abuse AT wxs DOT nl NNTP-Posting-Date: 31 Dec 1998 12:32:05 GMT X-Mailer: Mozilla 4.04 [en] (Win95; I) To: djgpp AT delorie DOT com DJ-Gateway: from newsgroup comp.os.msdos.djgpp Reply-To: djgpp AT delorie DOT com BJC wrote: > In article <767k28$e46$1 AT latinum DOT dresearch DOT de>, beck AT dresearch DOT de says... > > No, sorryly this is a bug in you :-) > > heh, it figures ;-) > > > >i compiled this program with DJGPP 2.01 and PGCC 1.01 and all was > > well...infact it left 2 > > >bytes space between str2 and str1. > > > > It's difficult to say why. Maybe the optimizers in this compilers think, > > that the strings should be allocated > > 8 bytes aligned... > > hmm... > Actually the compiler should optimize to 4 byte boundaries. Remember str1 - its 5 bytes. This will be rounded up to the nearest multiple of 4 (which is 8). Your str2 happens to be exactly 4 bytes so no optimization is needed. In the DJGPP faq it is stated that optimizing to 4 byte boundaries is needed to gain a big speed increase in program execution (it is a 32-bit compiler). If you really are low on memory you may instruct the compiler explicitly to not optimize to 4-byte boundaries with the __packed__ attribute. Read all about this in the docs that came with djgpp. > > So the problem is your for loop. Note, that in C the typical construct for > > traversing n elements is > > for (i = 0; i < n; ++i) > > JonnyCarsonMode(tm) ON > "oh, i did not know that." > JonnyCarsonMode(tm) OFF > > heehee, well like i said i'm still new to programming. > Thanks for the replys!! > > -- > brian james chapman > email: bchapman(at)feist(dot)com