Thanks Bernd. Yes i understand this load issue and how the resolved pointers will not work from save to save. I figured that was the reason but was not sure. No reason to rush on the resolving on this issue as i have many other methods to get around what i was doing. Mostly i was just lazy using some c code that was already there. I will just rewrite it in forth and be done with it. PKS
On Tue, Sep 3, 2013 at 7:20 AM, Bernd Paysan <bernd.pay...@gmx.de> wrote: > Am Montag, 2. September 2013, 17:16:57 schrieb Philip Smith: > > Hi Bernd > > Thanks for the information! I now wonder what i should do? I know i > > can do this from a script like method but i want the convenience of the > > script method and the convenience of forth but i also want the speed of > > compiled binarys or close to it. I also would like the benefit of code > > that can work on other platforms. I guess for now i will just try to > > figure out a best compromise method that will give me the most > flexibility > > with the most speed. > > I will just have to think about this for sometime before i come up > with > > a solution. > > Is the main reason this does not work yet because of some dependency > > Gforth has with some other library or is this an internal Gforth issue? > > The main reason is that this is not trivial to do. We can't just save the > image with the resolved pointers into the library, because next time we > load > the library, it might have a different address (it will be the same > library, > though, as we load our wrapper, but ASLR is pretty common). > > > Maybe i could try and solve this or is there some solution that you Bernd > > and others are already working on? > > I'm not working on it right now, but it is on my things to do list. Maybe > I'll > move that up in the priority list. It will only go into the development > branch (git respository). > > > I have done some testing to see the speed benefit of precompiled > images > > and i find that there is up to 200 times speed benefit for some of my > > testing code. I know that seems like an amazing amount as 200 times is > > very noticeable. My original idea of having small parts of code run > from > > the command line and piped to other parts of code would improve if it can > > all be precompiled images. But another way of structuring code i will > need > > to devise to work around this issue then i guess! > > Thanks again for the answer! PKS > > -- > Bernd Paysan > "If you want it done right, you have to do it yourself" > http://bernd-paysan.de/ > -- \|||/ -----0oo---( o o )---oo0------ (_) >From the hand of PKS
