home *** CD-ROM | disk | FTP | other *** search
Text File | 1990-03-16 | 1.3 KB | 38 lines | [TEXT/GEOL] |
- Item 2727711 15-March-90 09:33PST
-
- From: VALOIS1 Valois, John E.
-
- To: CPLUS.APPLE$ C++ Interest List--Apple Employees
- CPLUS.DEV$ C++ Interest List--Developers
- NAUTIL France - Dev, Nautil Info Lyon,IDV
-
- Sub: RE>Dead Code Stripping wi
-
- Attn: C++ Apple
- Attn: C++ Developers
- Attn: Etienne Vautherin
- SentBy: Jack Valois
- Subject RE>Dead Code Stripping with C++
- Etienne, let me take a stab at answering this one for you. I am new to C++
- but, I had this one explained to me once and I think that I have it right now.
-
- >Is it necessary to keep in memory one instance of each "dead strip candidate"
- class?
- >Could the sample in "Release Notes, Preliminary C++ Interfaces For MacApp
- 2.0ß9" p.3, be written:
-
- >if (gDeadStripSuppression) {
- > TFoo* aFoo = new TFoo; aFoo.Free;
- > TFoo2* aFoo2 = new TFoo2; aFoo2.Free;
- >}
-
- The code fragment that you add to prevent dead code stripping is there to fake
- out the linker. If you examine the variable gDeadStripSuppression you will
- find out that it is never true. So, that code fragement never gets executed.
- Therefore you won't have an instance of each "dead strip candidate class" in
- memory.
-
- Jack Valois
-
-
-