There are Shoes 3.3.3 beta’s to download from and the Change Log is here.
What you don’t know, behind the scenes. It was a lot of work from core contributers and myself. And then some core panties got twisted too tight to work together and now it’s just me working on the C side of Shoes. I miss them. I’ll work through their bugs reports and suggestions as best I can because they are good things to do.
After @backorder moved the Shoes C code into a more logical configuration and I changed the rakefile mess to leverage that – It’s quite developer friendly now if you are of the C persuasion. Most Shoes folk will never see those changes and that’s OK with me. It was something that just needed to be done.
Turns out there is another task that ‘just needs to be done’ – cleaning up my computers by removing all the useless junk I have gathered over many years. Some junk was shoes related and some were just grand ideas I got involed with Shoes. I have a soft spot in my heart for the Raspberri Pi so I ordered an SSD and that board didn’t work, I ordered some more parts to make the ssd work in the main box because good money needs more good money and and then some more parts and then some head scratching on making all these Chinese adapters work and I was pushing GB’s of data around which is kind of slow.
I’m not done yet but I can see the light at end of the tunnel. It just needed to be done.
Yes, it took forever and it’s not perfect but we could wait forever for perfection. Shoes 3.3.2 has some fun new features and some bug fixes. Available via at the usual place.
You may have noticed that it’s been a while since Shoes 3.3.1 was released and 3.3.2 keeps getting pushed into the future. What’s going on?
Well, some of the new features are very difficult to write and time consuming to test. We’ve discovered some very difficult issues with packaging for Windows 10 and so far, there is no way to get around what Microsoft wants and what I’m willing to do to Shoes, or what your willing to do. It’s called “code signing” which requires the annual purchase of a certificate (and the price can be very high with some certificate authorities). I won’t pay for that and I don’t think you should either.
The ground shifted and all we can do is adapt. Don’t package your scripts in an exe! Just tell your Windows users to install Shoes first. Then run the .shy you created.
Ruby also shifts under our feet. As Ruby evolves, they change things. Some of those changes will involve a lot of work for the Shoes C code in the future. I just found another problem with Ruby 2.2.6 (can’t cross compile the fiddle gem for Windows) – that will slow down building and testing even more if it can’t be resolved. I also found a new dependency problem last night for Windows but that should not be too difficult.
And then there are the Shiny Things that may not work, but consume lots of time to discover if it’s a failure or a brilliant idea. Like ebook
This one might cause some problems because it fixes a problem where the code didn’t do what the documentation said for some click and release events. The docs say to deliver |btn, left, top| to the proc. They didn’t. From what I can tell, they never did.
We also accept/present more button numbers instead of is always “1”. Button numbers start at one and increase from there in some pattern only you and your mouse device know. Cross platform? No! Your problem to sort out if the wheel click on OSX is a different number on Linux. Most people will stick with not caring which button. Now you have options.
If you’ve worked around the lack of |btn,l,t| on click and release then you may want to test your scripts with this beta. In this case, the documentation says the right thing and the code was incorrect. Long time wrong. See here
The beta also include the latest, almost nearly OK, plot widget.
I thought Shoes should have a plot widget and I happened to have some ancient 1989 Java code (still works) that I could re-purpose into a C/Cairo based ‘plot’ widget for Shoes. Shiny! New! Fun! Might as well fix some know problems with that old code. So I did that It wasn’t the easiest thing in to do in Shoes but not the hardest either. There are many sharp edges to make round so it might not be in Shoes 3.3.2 – currently it’s issue 264.
It’s never going to be gnuplot replacement. If you need that level of control, you’re already using gnuplot or matlab. This would only be a gateway into the world of visualization).
Toy Data:
Real data:
I like it and there is the possibility to add to the graph in real time with new data or add new data series to plot,
It’s also a long road to release quality. I did mention the edge case problems, Yes, I did.
Shoes 3.3.2 may not redefine Ruby exit and quit like it used to. Instead your scripts should use Shoes.quit. That should work on on Shoes 3.3.1 (and perhaps earlier). I might be the only person that calls exit or quit on a button press.
The window decorator icons and the osx global menu work as before. Avoid the panic and change your calls to Ruby quit and exit to Shoes.quit. Assuming you have any. Of course, you’ll try it and report bugs if it doesn’t work. Right? Please.
I know it seems like a long time since the last release – it has been a long time. Here is the deal, we were mucking around with command line stuff which is not cross platform by definition and is difficult to debug and test. A lot of that difficulty went away with the new cshoes launch scripts for OSX – I think I told you how great fix is?
Pretending to be a xterm console will get you into some dark places. Digging new tunnels in dark places is usually not a successful strategy for finding daylight. Dig I did.
Byebug was never well tested or used and with the old clunky osx launcher not all that usable. Shoes 3.3.2 is going to fix that. From the cshoes command line, `-d` will put you in byebug, pointing to some Shoes startup Ruby code. Set breakpoints in whatever file you intend to debug. ‘c’ and continue to it. Any args after -d are subject to normal shoes handling – so -d knows nothing about those extra args. You can debug the Ruby code of Shoes startup and processing this way or set a breakpoint on the file you want and let Shoes load it normally.
I’ve already mentioned the Shoes.terminal and it’s documented in the Shoes Manual now in case you need to change some appearance settings. Beta Time!
Yes, there are some known bugs. It’s a beta. You might find it useful.
I’ve uploaded the Shoes 3.3.2 binaries.
What’s new – damn near nothing. A font loading fix on Windows. If you happen to be a command line commando however, you might see some things that matter: A new Shoes.terminal to replace that brain dead Shoes.console. Many man hours (lot of many) were put into that creature. Which caused a long overdue rewrite of OSX terminal launch scripts. If you are an OSX command line person you’ll say “Thank you, Jesus!” But my name is Cecil and I have a lot of work to before 3.3.2 gets beyond this early beta. We also added a Fun but rudimentary profiler (caution: may not work on Windows).
With the new terminal code, you can run test/unit tests of your non-gui code without the command line if your willing to do some Shoes/Ruby level of work. I call it pre-beta because it is not complete enough to be a beta. If you’re using Shoes from the OSX command line – you want this not completely baked version of Shoes – you really do. It also fixes the multiple script open bug on OSX.
For OSX, it is not perfect but it is So Much better. You just get to the Cobbler (Maintain Shoes or -c) and create the cshoes script. If you’re driving Shoes from a textmate or applescript, you’ll need to update it to use cshoes instead of open’
Never fear. There is work going on for Shoes 3.3.2 – but it’s edge case things, and bug fixes for stuff most people don’t care about. Yet, it has to be done.
There are so many OSX issues to fix. Since I don’t know what I’m doing in Cocoa and Objective-C, it’s taking a lot time. That said, I have made progress. Presumably, I’ve learned a lot about OSX. It is better on OSX.