Mailing-List: contact cygwin-developers-help AT sourceware DOT cygnus DOT com; run by ezmlm List-Subscribe: List-Archive: List-Post: List-Help: , Sender: cygwin-developers-owner AT sources DOT redhat DOT com Delivered-To: mailing list cygwin-developers AT sources DOT redhat DOT com From: Matt X-Sender: matt AT cesium DOT clock DOT org To: cygwin-developers AT cygwin DOT com Subject: Re: failed malloc()? In-Reply-To: <20001125200651.B5139@redhat.com> Message-ID: MIME-Version: 1.0 Content-Type: TEXT/PLAIN; charset=US-ASCII Date: Sat, 25 Nov 2000 17:24:31 -0800 On Sat, 25 Nov 2000, Christopher Faylor wrote: > On Sat, Nov 25, 2000 at 03:41:59PM -0800, Matt wrote: > >I've started reviewing some of the winsup code and had a question as to > >how to most properly correct unchecked pointers. That is, I know to add a > >statement to check the freshly created/passed pointer, but not what to do > >in the instance that pointer is NULL. > > > >Some places in the cygwin code do a system_printf() which is what I will > >assume is desired, but wanted to check to make sure what the > >standard/preferred way of handling such things is (if there is one). > > > >Here one of the sections of code I am looking at (cygcheck.c:91,92): > > paths[num_paths] = (char *) malloc (maxlen + 1); > > memcpy (paths[num_paths], s, maxlen); > > Um. That's not cygwin, that's cygcheck. Proper programming practice would > be to check every single malloc and issue an error if there is a problem. > > You can't use system_printf in this case since it isn't cygwin that you're > looking at. > > It is usually a sign of some severe memory corruption if a program like > cygcheck isn't able to allocate its own buffers so I am not too worried > that there is no protection in this code. I wanted to start with something simple first, and once I got comfortable/proficient, moving onto cygwin itself. What would be the preferred method to output these error messages? I am using the format "file: function: message". For example, "cygcheck.cc: get_word(): Readfile returned error (GetLastError)". Thanks for the help! -- it's better to burn out than to fade away