Hi all, I tried to build the latest Shoes code with ruby1.9.2p136 for Windows. Look at the commits of ashbb/shoes and ashbb/rubyinstaller on GitHub. There are some restricts, e.g. - just updated ruby but not update other dependencies - omitted shoes/cache and shoes/shybuilder so far - line 9-12 in shoes/shoes.rb is a monkey patch just to eliminate an error for require 'open-uri'. But anyway, succeeded to build shoes.exe included ruby1.9.2p136 and it can open Shoes opening window. ;-) I'm not sure using ruby1.9.2 will resolve any issues. Just tried. :-P ashbb
Ash! You're the best. I've been working over this lately. Did you see my pull request to rubyinstaller? I was having an issue where shoes would build, but not open. I'll give this a shot too! On Wed, Jan 5, 2011 at 10:56 AM, ashbb <ashbbb@gmail.com> wrote: > Hi all, > > I tried to build the latest Shoes code with ruby1.9.2p136 for Windows. > > Look at the commits of ashbb/shoes and ashbb/rubyinstaller on GitHub. > > There are some restricts, e.g. > - just updated ruby but not update other dependencies > - omitted shoes/cache and shoes/shybuilder so far > - line 9-12 in shoes/shoes.rb is a monkey patch just to eliminate an > error for require 'open-uri'. > > But anyway, succeeded to build shoes.exe included ruby1.9.2p136 and it can > open Shoes opening window. ;-) > > I'm not sure using ruby1.9.2 will resolve any issues. Just tried. :-P > > ashbb > >
Little endian means that the binary bits are read in one direction (front to back) big endian is the oposite. U should confirm the directions tho as its been a bit since i have had to mess with it. (in response to origin on the endian) cw -----Original Message----- Date: Saturday, January 15, 2011 2:35:03 am To: "shoes@librelist.com"<shoes@librelist.com>, From: "Dave Wilkinson" <wilkie05@gmail.com> Subject: Re: [shoes] First step for WalkAbout (Shoes 4) UTF16LE (yes, little endian... it is 16bit unicode, so it could be either, so you have to specify) is the native unicode of Windows. Is it using code given from a text box or something? There's no way source files are being stored as utf16. In the former case, it makes sense that it would be converting from the windows native utf16 to utf8. Although, normally, that conversion isn't really common, so maybe such a package was removed?
Hi all,
As Steve mentioned, this is a problem of Shoes itself built with
ruby1.9.2p136. I think we have to update other dependencies too.
BTW. Replace this line:
https://github.com/shoes/shoes-contrib/blob/master/browser.rb#L14
to
@example_box.items =
Dir.glob("#{box.text}/*.rb".force_encoding('US-ASCII'))
You can run shoes-contrib/browser.rb with the Shoes.
This is not a correct solution, though... :-P
Regards,
ashbb
> Did you see my pull request to rubyinstaller?
Yes, merged.
https://github.com/ashbb/rubyinstaller/commit/49c215f62b8316d365581ef4a1edd0ec88c7b3a7
Thanks! :)
ashbb
I don't know if you get notes when someone comments on one of your commits, but I just made on on RubyInstaller. I thought that that other line needed changing, too? Does this work? I can't check until tomorrow.
Hi Steve, > I don't know if you get notes when someone comments on one of your commits, GitHub sent me your comment by email. :) > I thought that that other line needed changing, too? No. Look at the reply: https://github.com/ashbb/rubyinstaller/commit/e8cdcbe677820b7f8cb7fdf21a65e0dc109ea891 > Does this work? Yes. I could build shoes.exe with ruby1.9.2p136 and it could open Shoes opening window. :) ashbb
So, yeah. After getting my Windows 7 issues straightened out, it seems that this builds. Fun. :D When I try to build Hackety, though... ( by following https://github.com/ashbb/shoes_hack_note/blob/master/md/hack028.md ) I get this strange, strange error: code converter not found (UTF-16LE to ASCII-8BIT) After doing some googling, it seems like this means that Rubygems is failing to load? It happens on the line that says 'require hpricot' Anyone have any thoughts on this? I'm going to dig deeper into it over the weekend.
If I open up the shoes-contrib browser, a similar error appears whenever I try to pick a sample: code converter not found (UTF-16LE to UTF8) So this seems to be a problem with the build itself, not Hackety. Hm.
On Fri, 2011-01-14 at 22:23 -0800, Steve Klabnik wrote: > > code converter not found (UTF-16LE to UTF8) > it could be your new Win 7/Ruby install is, maybe, possibly missing the UTF16LE to UTF8 converter because of incorrect env paths or it may LE means little endian, a suspicious name. Where did that name come from?
The first is coming from the line in Hackety where it does a 'require hpricot' and the second is from the Shoes Contrib browser where it's using input from a textbox to glob a directory, as you can see from Ash's example. On Fri, Jan 14, 2011 at 10:54 PM, Cecil Coupe <ccoupe@cableone.net> wrote: > On Fri, 2011-01-14 at 22:23 -0800, Steve Klabnik wrote: > > > > > code converter not found (UTF-16LE to UTF8) > > > it could be your new Win 7/Ruby install is, maybe, possibly missing the > UTF16LE to UTF8 converter because of incorrect env paths or it may LE > means little endian, a suspicious name. Where did that name come from? > > >
Doing some more research, this seems to be helpful: http://tech.groups.yahoo.com/group/vimdev/message/59227 :/ Also, of note, my Ruby 1.9.2 that's installed works fine. So... yeah. Seems to be an issue with the Special Stuff we're doing in Shoes. Yay.
I've upgraded hpricot from 0.8.1 to 0.8.3 here: https://github.com/zacharyscott/shoes/tree/upgrade_hpricot Particularly because I saw this commit (which was not included in 0.8.1): https://github.com/hpricot/hpricot/commit/8e815c2 Hope this helps! Give it a try and let me know what you guys think! On Sat, Jan 15, 2011 at 6:35 PM, Steve Klabnik <steve@steveklabnik.com> wrote: > Doing some more research, this seems to be helpful: > http://tech.groups.yahoo.com/group/vimdev/message/59227 > :/ > Also, of note, my Ruby 1.9.2 that's installed works fine. So... yeah. Seems > to be an issue with the Special Stuff we're doing in Shoes. > Yay. -- Zachary Scott
zzak's commit includes .o files which will break on anything that's not linux, if you guys try them out... So, the windows problems still happen, even with the new hpricot. This small shoes script causes the error: Shoes.app do para Gem::VERSION end So... yeah. Even just loading anything apparently has problems...
UTF16LE (yes, little endian... it is 16bit unicode, so it could be either, so you have to specify) is the native unicode of Windows. Is it using code given from a text box or something? There's no way source files are being stored as utf16. In the former case, it makes sense that it would be converting from the windows native utf16 to utf8. Although, normally, that conversion isn't really common, so maybe such a package was removed?