Date: Thu, 26 Mar 1998 16:22:07 +0300 (IDT) From: Eli Zaretskii To: Eugene Ageenko cc: djgpp AT delorie DOT com, Robert Hoehne Subject: Re: Strip and Coff2exe. File size increasing! In-Reply-To: <351A42B3.756912F2@cs.joensuu.fi> Message-ID: MIME-Version: 1.0 Content-Type: TEXT/PLAIN; charset=US-ASCII Precedence: bulk On Thu, 26 Mar 1998, Eugene Ageenko wrote: > > 1) The versions of your GCC and Binutils; > > Hey Man!, its me again, I have the LATEST binary distribution og GCC and related tools, > just taken 2 days ago from SimTel. Look there for versions. Please bear with me. I reply to too many problem reports on any given day, and cannot afford keeping track of everybody and their versions. I prefer to ask explicitly instead of relying on my failing memory and veering away of the real cause. > $strip test.exe > $dir > > TEST EXE 38 552 26.03.98 13:40 test.exe > > # now let's use different tactic > > $strip test > $coff2exe test > > TEST 34 456 26.03.98 13:41 test > TEST EXE 36 504 26.03.98 13:41 test.exe Robert, it seems like `strip' from the latest Binutils prepends the stub twice (the added size iz 4K instead of 2K). Can you please explain what happens here? > Now the question - Is this RIGHT, that file size increaes? What happens if you say "strip test.exe" 3 or 4 times? Does it add 2KB every time or only the first time? > Is it right that file size of stripped executable is larger that file size of > executable produced from stripped coff file? All the different ways to get a stripped binary (gcc -s, strip coff, and strip .exe) should all yield the same size. Any other result means a bug somewhere. In this case, the bug seems to be in the case when `strip' is invoked on stubified .exe program. > And also: why file size increases when we try to strip executable produced > from already stripped coff file??? Because it seems to add another 2KB-long stub. Can you look at the larger .exe and see if the stub is indeed prepended twice?