Re: [AD] Plans for Allegro 4.9.5 (resend)

[Thomas Fjellstrom <tfjellstrom@xxxxxxxxxx>]

On Saturday 13 September 2008, Thomas Fjellstrom wrote:
> On Saturday 13 September 2008, Peter Wang wrote:
> > On 2008-09-13, Thomas Fjellstrom <tfjellstrom@xxxxxxxxxx> wrote:
> > > Ok, I've been working on it a little tonight, and I wonder, should I
> > > rename the vtable stucts to be more like the other ALLEGRO_*_INTERFACE
> > > structs?
> >
> > Sure.
> >
> > > Also, if you're wondering why I haven't made it all part of the
> > > ALLEGRO_SYSTEM_INTERFACE stuff, its mainly because the fshooks need to
> > > be setup before the system driver gets initialized, does allegro still
> > > load a system driver name from a config file?
> >
> > Not right now.  Currently there is exactly one system driver per
> > operating system.  We might need an equivalent of SYSTEM_NONE later.
> > Also, theoretically, the Linux console port might come back, in which
> > case we need multiple system drivers.
> >
> > Something like file system handling, on any one OS, will have the same
> > implementation no matter what system driver is active, so I think you're
> > right to separate it from the system driver.
> >
> > Peter
>
> I've merged trunk with my branch and done a bunch of fixups, the code
> should work now, it just isn't used anyplace besides al_get_path which I
> also made sure to setup for win32 and unix.
>
> If you feel like merging it to trunk feel free, but theres still plenty of
> things left to do (like merge in trunks changes made after I did the
> original merge this monrning), most are noted in FSHOOK-TODO.
>

Most of the base lib fshook work is done now. I've also converted the parts of 
allegro that get compiled here on X to use the new fshook api, as well as the 
AL_PATH stuff (except for config.c, haven't done that yet, easy fix though, its 
all stdio, mostly search&replace). examples and the demo now use it all as 
well. My branch is 100% compilable on linuX.

I'd like people to test it on their available platforms and to take a look at 
the api and code to tell me if theres anything seriously wrong.

I'd like to get the code working on all three platforms and merged into 4.9 
before the next WIP release :)

I also have some concerns about my AL_PATH api, and al_get_path function... I 
decided to split the AL_PATH api into the drive, path and filename, which 
"automagically" parses the path into the 3 components, which means if you want 
the path all dirs, the last character has to be a / or \. if you don't that 
path component becomes the AL_PATH's filename component. When working on the 
demo Misc.cpp file I noticed that that is somewhat inconvenient at times, but 
not having the split components can be very inconvenient at other times. And 
al_get_path on unix is a little iffy, I wanted USER_DATA_PATH to include the 
"program" name to save the extra append later by the user, but that needs a 
way to get that from the program, it did use the exe name, but that does not 
work right so I've switched to just making it a clone of USER_HOME_PATH and 
the user has to append the dir they want.


p.s. sorry if you got a duplicate, seems it didn't get to some people at 
all...

-- 
Thomas Fjellstrom
tfjellstrom@xxxxxxxxxx