[Xastir] Getting rid of external libraries (opinion)

[Lee Bengston]

> On Fri, 8 Aug 2008, Curt, WE7U wrote:
>
>> We've been suffering from problems with IM (somewhat more), GM
>> (somewhat less), Motif, and Lesstif for some time, with a few
>> problems here and there with wget and libcurl.  Sometimes a few
>> other libs but probably not enough to worry about.
>>
>> When I think about external libs now I mostly think of problems
>> instead of solutions.
>>
>> I've been interested for some time in dumping some of our libraries
>> and finding alternatives, but not interested enough yet to actually
>> start doing it.
>>
>> 1) IM/GM might be replaced for our purposes with some package that
>> could do GIF/JPG/PNG, was in source form, and was GPL'ed.  We could
>> add it to our package like we did the Shapelib package.
>>
>> Tom will probably break in here and suggest that we not add things
>> to our sources like this, but really we've had so much trouble with
>> IM/GM that I'm suggesting that replacing them with a known working
>> subset is a good thing.  It'd solve a major problem that we have had
>> for years:  Unstable image libraries.  There's quite a bit of image
>> processing code in Xastir so this would take a while to do.

I think I have to side with Curt on this one.  I don't think having
Shapelib integrated has really hurt anything.  Even though it is
present internally, the user still has the option of building with his
own library.  If that same approach was taken with IM, GM, or an
alternative, then I am hard pressed to understand what the
disadvantage(s) would be.  It appears that many of the recent problems
have been associated with the way the packages for the specific Linux
distributions were built, and I would think all of those type of
problems would be avoided with an internal image library.  Perhaps an
alternative image processor is better suited for Xastir-NG, and just
adding an internal IM or GM coded using the settings that work best
with Xastir would be an easier value-add for the current version - or
is there a licensing issue with doing that?

>> 2) I've had some trouble with non-working wget in the past, and very
>> occasionally with libcurl.  Our code is already isolated pretty well
>> here so that we could find a package that did ftp/http fetches and
>> integrate it into our code easily.  That would solve another problem
>> and would be easy/fast to do.

I've played with a lot of linux distro's and Xastir, and both wget and
libcurl seem to work OK for me.  Therefore I don't have much of an
opinion here.

>> 3) The last problem is the OpenMotif/Lesstif thing.  These libraries
>> are rapidly aging now, everyone else having moved on to the Qt
>> and/or Gtk+ libraries.  Switching our GUI over to one (or both?) of
>> these would be a more major task, as there's GUI code through most
>> of our source files.

I don't think re-doing the GUI is worth the effort in the current
Xastir version.  It would be nice, but realizing resources don't grow
on trees, I agree that a new GUI is better suited for the next
generation Xastir.  A motif based library internal to the current
Xastir, though, would be very interesting, but I don't know how
feasible that would be.  For example, I recently created a VM using
openSolaris, and I didn't see a LessTif or openMotif package for that,
so it didn't look like I was going to get Xastir running on it.  Ditto
for Solaris 10.

Regards,
Lee - K5DAT
Murphy, TX