Thread: [Lily4jedit-devel] coLinux
Brought to you by:
fodber
From: Joshua K. <jos...@gm...> - 2005-03-03 02:34:29
|
Hi... I wonder if you have come across or use coLinux. Its said to be able to run linux applications to native speeds, without modifications on windows. You could modify your existing linux partition and run in on windows. The webby is at http://colinux.org and lots of documentation is at wiki.colinux.org Anyway I installed it at my workplace, and its seems to run fine. I could install lilypond 2.4.2 on debian in colinux easily at quite a plesant speed using a petium 500 pc. I, however have not yet try it at home. i seem to screwed up, aftering switching my hdd from primary to slave, and when booting my debain causes kernel panic. Well, i'm thinking, what if colinux becomes the prefered choice for running lilypond on windows (over cygwin or qemu or vmware or virtualpc). 2 options for lilytool users. 1. Running in X lilytool could be installed on colinux and used via X ( because colinux dun support x displaying natively, so that would be either running cygwin/X or VNC or NXclient). 2. Using Lilytool in windows then, since the colinux would be like another terminal or machine, lilytool could (s)ftp to colinux, upload/update the lilypond file, then ssh(2) in and run lilypond on the file. but this might complicate on how to view the files, so maybe colinux place the file using samba on windows shared folder? ha, just thoughts again ;) Joshua |
From: Bertalan F. <fo...@fr...> - 2005-03-03 08:42:39
|
I also was thinking about something like this. There could be a server in colinux, which could work with a network connection, sending messages and files. If it is much faster than running lilypond in cygwin, it would worth creating it. Bert |
From: Johannes S. <Joh...@gm...> - 2005-03-03 11:34:14
|
Hi, On Thu, 3 Mar 2005, Bertalan Fodor wrote: > I also was thinking about something like this. There could be a server > in colinux, which could work with a network connection, sending messages > and files. If it is much faster than running lilypond in cygwin, it > would worth creating it. IIRC coLinux only works on W2K and WXP, not on 95-ME. Another option would be MinGW32. In contrast to cygwin, MinGW32 does not provide a complete POSIX interface (no fork()...), and therefore is much faster starting up. AFAIK the main problem regarding speed was that the bash initializes a few things the POSIX way, which gets terribly slow on 95-ME, and to a certain extent also on NT-XP. Another advantage is that we don't have to care about cygwin installation, but can distribute the whole package in one .zip together with jEdit and java and lilypond and ghostscript. To my knowledge, the only reason to choose cygwin over MinGW32 was TeX, which works best with a complete Unix like environment. Having said this, there exists a good alternative to cygwin's TeX on Windows: MiKTeX. However, I only heard good things about it, never tried it. Of course, the easiest method would be to just disable the TeX backend (which is the default anyway). From the CVS logs I remember that in the last days, work was done to support point-and-click with the PDF backend via pseudo hypertext links. Unfortunately, I didn't have enough time yet to investigate more (would be nice to ask ghostscript to give a list of links together with their hot spots on the page, and I'm sure this is possible), because I am still writing away like crazy on my PhD. Ciao, Dscho |
From: Bertalan F. <fo...@fr...> - 2005-03-03 12:27:55
|
Well, porting to MinGW would be a very rational solution. However, it would be tougher than using cygwin and would require strong knowledge of posix development. >To my knowledge, the only reason to choose cygwin over MinGW32 was TeX, >which works best with a complete Unix like environment. Having said this, >there exists a good alternative to cygwin's TeX on Windows: MiKTeX. >However, I only heard good things about it, never tried it. > > I do use it, and is very good. >From the CVS logs I remember that in the last days, work was done to support point-and-click with the PDF backend >via pseudo hypertext links. > Yes it is. My plan is to integrate the multivalent (http://multivalent.sf.net) browser's PDF module into lily4jedit, providing a better point-and-click experience than with jDvi. Bert |
From: Johannes S. <Joh...@gm...> - 2005-03-03 12:53:47
|
Hi, On Thu, 3 Mar 2005, Bertalan Fodor wrote: > Well, porting to MinGW would be a very rational solution. However, it > would be tougher than using cygwin and would require strong knowledge of > posix development. The biggest problems IMHO are python and guile. Well maybe not Python, because there exists pyMinGW, but rather glib and pango. If I find the time, I will look into it next week. > >From the CVS logs I remember that in the last days, work was done to support point-and-click with the PDF backend > >via pseudo hypertext links. > > > Yes it is. My plan is to integrate the multivalent > (http://multivalent.sf.net) browser's PDF module into lily4jedit, > providing a better point-and-click experience than with jDvi. Yes, multivalent is a very good browser. And then we can kick out jDvi. Ciao, Dscho |