By using this site, you agree to our updated Privacy Policy and our Terms of Use. Manage your Cookies Settings.
457,949 Members | 1,511 Online
Bytes IT Community
+ Ask a Question
Need help? Post your question and get tips & solutions from a community of 457,949 IT Pros & Developers. It's quick & easy.

Cross-compiler way to handle memory allocation failure ?

P: n/a
Hi!
What is the method you use to handle memory allocation failure in C++
using the "new" operator?
I used to only check if a NULL pointer is returned after the "new" call,
but that's probably only working in MSVC++.
I also read about the "bad_alloc" exception, but that doesn't seem to be
thrown by MSVC++.
Now, one could of course say "don't use MSVC++", but I'd rather like to
hear a good solution for being able to handle memory allocation failure
in a cross-compiler way (we're using MSVC++ on windows and gcc on
linux). Any ideas anyone?
Is installing a new handler the only way to be sure?
Koen
Jul 19 '05 #1
Share this Question
Share on Google+
1 Reply


P: n/a
"Koen" <no@ssppaamm.com> wrote in message news:<bd**********@gaudi2.UGent.be>...
Hi!
What is the method you use to handle memory allocation failure in C++
using the "new" operator?
I used to only check if a NULL pointer is returned after the "new" call,
but that's probably only working in MSVC++.
I also read about the "bad_alloc" exception, but that doesn't seem to be
thrown by MSVC++.
Not in the last two versions. You have to go back to the 6.0 series,
which predates the standard. Both VC7.0 and VC7.1 get it right.
Now, one could of course say "don't use MSVC++", but I'd rather like to
hear a good solution for being able to handle memory allocation failure
in a cross-compiler way (we're using MSVC++ on windows and gcc on
linux). Any ideas anyone?
Is installing a new handler the only way to be sure?


I believe installing a new handler is also buggy on VC6. Upgrading from
VC6 to VC7 is relatively painless; i've done it on more than one project.
As others mentioned, new(nothrow) is an option.

For windows, this often is a bit of a non-issue though. Windows will
swap to death before failing, the user will probably kill off the
program first. The only case I know that fails fast enough to matter is
running into the 2Gb limit, if your PC has more than 2Gb RAM. I don't
know if that's implemented properly in VC6, though.

Regards,
--
Michiel Salters
Jul 19 '05 #2

This discussion thread is closed

Replies have been disabled for this discussion.