[Date Prev][Date Next][Thread Prev][Thread Next][Thread Index]

Re: [XaraXtreme-dev] Mac build



Hi,

I got so far to know that it fails in main2.cpp:InitPostTools1 - but all the &&'ing makes it impossible to step through the initialisation with gdb - is there any way to do this nicely?

dom

On 19.03.2006, at 12:12, Alex Bligh wrote:



--On 19 March 2006 12:03 +0100 Dominik Wagner <dom@xxxxxxxxxxxxxxxx> wrote:

any suggestions?

Yes. This is the rather annoying "the Kernel doesn't initialize problem".
What's happening is it's running through all the initialization in
main1.cpp, main2.cpp & main3.cpp, and something (looks like something
quite far down) is failing. Error reporting isn't quite right under
LX so you aren't seeing the error. Annoyingly the list of initializers
is one big if() statement with && between the calls so setting breakpoints
is not easy.

You can often find these by putting a few breakpoints in wxOil/ errors.cpp in Error::XSetError (all the various flavours) which /should/ find what is producing the error. If not, I'm afraid it's a case of stepping in and out of each of the various init jobs (put a breakpoint in the obvious place in each of main1, main2, main3.cpp) and finding which class ::Init function is
returning FALSE (clearly finding which of the "main" files causes the
problem is the first step). It's a little time-consuming, but normally once you've done that debugging why the init fails is pretty simple. I guess is
something pretty recent as the Mac build used to init OK.

I expect you could insert something like
 && TRACE(_T("Got here"))
etc. into the init lists.

Alex

--
Dominik Wagner          Mail: dom@xxxxxxxxxxxxxxxx
TheCodingMonkeys        http://www.codingmonkeys.de/
Blog - DasGenie: !Scrap http://scrap.dasgenie.com/

Attachment: smime.p7s
Description: S/MIME cryptographic signature