This version has a handful of build fixes and array size fixes, but does
not yet have the speed fix Tom Wilson posted.
Well, my best guess at the proper fixes. I made the compiler warnings go
away, but I'm not 100% sure they are really the right way to fix each item.
Some arrays looked like they were one byte too small so they couldn't
hold their terminating null. But maybe the arrays were the right size
and the variables that used them needed to be reduced by one instead.
Others were fuzzier like just needing more room for path strings, but no
specific number, just, more, so I just picked a bigger value arbitrarily.
This shows everything in one diff
https://github.com/bkw777/VirtualT/commit/fe6df94725a5fcf1989964b22cb79848dc778a6d
Oh yeah I see some fltk fixes for newer versions too I forgot those.
Particularly I'm not sure if removing the Cancel button in display.cpp
was correct. It makes the too-many-params error go away, but I don't
know if that option was actually needed and so it needs some other fix
that lets you still have the button.
I think Ken said he addressed the speed problem in his repo by now. I
don't remember if he posted a github or other public link to his new git
repo.
The diff for Tom's speed fix from my local copy is:
(not a real diff, I collapsed all the spaces for the email)
bkw@negre:~/src/VirtualT$ git diff HEAD^
diff --git a/src/io.c b/src/io.c
index 28bbdb4..ef4cf42 100644
--- a/src/io.c
+++ b/src/io.c
@@ -1262,14 +1262,15 @@ int inport(uchar port)
{
if (fullspeed == 0)
{
+ // Disabled per Tom Wilson <wilso...@gmail.com> post to m100 list
/* Loop through all LCD driver modules */
- for (c = 0; c < 10; c++)
- {
- /* Check if this driver is enabled */
- if (lcdbits & (1 << c))
- if (lcdTime[c]+.000014 > hirestimer())
- return (0x80);
- }
+ //for (c = 0; c < 10; c++)
+ //{
+ // /* Check if this driver is enabled */
+ // if (lcdbits & (1 << c))
+ // if (lcdTime[c]+.000014 > hirestimer())
+ // return (0x80);
+ //}
return 64;
}
else
--
bkw
On 3/24/20 4:22 PM, Kevin Becker wrote:
For what it's worth, I just built VirtualT on Fedora 31 using Brian's
github. I didn't build fltk myself, I used the packaged versions.
git clone https://github.com/bkw777/VirtualT.git
sudo dnf install fltk fltk-devel fltk-static fltk-fluid
libXinerama-devel libjpeg-devel
export FLTKDIR=/usr/lib64
make
I haven't done any real testing but it launches and appears to work
On Mon, Jan 6, 2020, at 10:17 PM, Brian K. White wrote:
I was just now able to build the latest FLTK and VirtualT on linux with
the following changes from the v1.7 sourceforge version:
https://github.com/bkw777/VirtualT/commit/fe6df94725a5fcf1989964b22cb79848dc778a6d
I don't know if every change is quite the most correct way to resolve
the compiler warnings or errors, and I don't know if one or two changes
don't also break compiling on Windows or Mac.
But this commit link shows everything all together and it's not very
much, and at least for me, it builds without even any compiler warnings
let alone errors, and the resulting binary runs.
I started with a clean copy of the src zip file from sourceforge, not a
fork from one of the other copies of VirtualT already on github.
For fltk I didn't I didn't have to change anything. I just cloned the
current fltk repo from github (from the fltk web site) and followed the
CMake directions for building on linux.
--
bkw
--
bkw
