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

Re: [XaraXtreme-dev] Ping

Alex Bligh wrote:

> Jed Frechette wrote:
>> Doesn't this imply that, over time, the "in the wild" binary
>> would become obsolete and likely eventually fail to function
>> as the OS and related bits  advance?
> CDraw doesn't talk the OS. I'm not sure it depends on any
> library at all - I think it's static linked and doesn't depend
> on any library at all.

To set the record straight here:

While it doesn't talk to the OS, due to mistakes (and perhaps
some non-mistakes) made by the developers, it uses and requires
GNU GCC-compatible C++ implementation (meaning compiler _and_
library) for at least the following reasons:

- static constructor calls
- C++ exception handling
- vftable layout (as some "hidden" code in GDraw uses virtual
  function calls to virtual functions overriden and implemented
  in the gdraw header file, which gets compiled by everyone).

Had the source been available, I could have fixed these mistakes
in a jiffy. With no source available; while I can fix (and have
fixed some of) this locally, I can't distribute these
corrections. As the non-availability of the source prevents me
from signing the agreement, I have been requested by Neil to not
even display such source code (to Xara people).

> However, in practical terms it is only a few thousand lines of
> code (see the presentation I gave in Kent), and it's pretty
> easy to work out WHAT it's meant to do (if not how it does it)
> from how it's called, and though the API isn't the cleanest in
> the world, it is an API.

Funny how it keeps getting repeated how small GDraw is, yet
there seems to be complete unwillingness to truly make it
available to make "Xara LX"/"Xtreme" a truly and fully
Free Software, GPL-compatible project.

> So potential developers have the assurance that if Xara never
> released the code, Magix got bought by MS or other pointy-
> horned events occurred, it would be quite possible (not
> trivial, but possible) to convert it to use code from another
> drawing engine - it might run slower, but it wouldn't be
> wasted code.

Pretty lame "assurance" if you ask me, as it assumes someone or
a group of people to think it was actually worth the effort to
do this. I think we all know what assumption is the mother of.
I think it'd be a mistake to assume this, especially with at
least one functional Free Software, cross platform, alternative.

> Note that I am not advocating people do this, merely pointing
> out it's a "security blanket" for developers.

One could recreate Windows XP too when MS abandons it (which is
quite soon)... Falling for this I think would be a mistake.