6:43 a.m.
As far as looking into other editors, xemacs works great for me - I
don't intend to change it. The problem of a good Unicode
editor is very much present.
I have been tracking XEmacs for this purpose on Win32 for quite some time.
Actually, I would even settle for a decent subset of Unicode. Unfortunately
Win32 Unicode editing is simply not supported in XEmacs stable, and the 21.5
branch does not produce binaries - being beta.
As Stephen put it:
> Replacing the current Mule internal representation with Unicode
is
planned, but not
> going to happen soon.
As it looks right now, simple unicode editing should be done with GNU Emacs,
and to get good unicode support, you have to use (g)vim. It works well even
on Win32, but using it is a pain for someone who has emacsen in their nervous
system.
I'm sorry Stephen, but the amount of UTF-8 encoded messages in newsgroups and
mail is going up all the time. Win32 XEmacs stable has problems even with
ISO-8859-15, not to mention UTF-8. We need a release of 21.5 to get Mule
support on Win32. Even then, any advanced unicode usage would be beyond
XEmacs.
As I have noted earlier, I can't build 21.5 myself, but I could provide test
results if a (net)installer would be provided. I've tried out some of the
unofficial 21.5 builds that have been available, but submitting bugs on old
versions is useless.
I honestly find it fairly tragic that XEmacs running on a Unicode-native OS
fails to provide any decent support for Unicode. As it stands, I can't
recommend XEmacs on Win32. And as I tend to support cross-platform
applications, that means I can't recommend XEmacs on any platform.
I wish this would change. But I don't see that on the horizon.
Camillo
7:19 a.m.
New subject: Characters lost when saving UTF-8 file on win32 (21.5.16)
I was able to build 21.5 easily on Win32 - after some initial wrestling
with config.inc, I was fine. I'm using MSVC.
I did some more testing with the version I built on a Debian box. Saving
the mentioned file produces exactly the same results on both MSVC Win32
and Debian (Cygwin build does not report any errors, but running it just
goes back to command line after some time - this might be another thing
to report).
This seems to be consistent with what Stephen already said.
Now, on the subject of Unicode support in current editor world. I deal
with Unicode every day (http://oss.software.ibm.com/icu) and one of the
most painful parts of the work is dealing with Unicode text data. I
don't need sophisticated editing tools, but rather need the editor to be
able to swallow utf-16, utf-8 and (ideally) backslash U format. Unipad
can do that, but is too expensive. Also, I frequently need to identify a
codepoint under the cursor (properties etc...). This is a great task to
do with xemacs - if it only supported Unicode well. Other Unicode
enabled Win32 editors are either clumsy and ugly or not multi-platform
or are no better than Notepad.
To summarize - I was really looking forward to using Unicode enabled
xemacs. Guess I'll have to wait some more.
Regards,
v.
Sars, Camillo wrote:
>As far as looking into other editors, xemacs works great for me -
I
>don't intend to change it. The problem of a good Unicode
>editor is very much present.
>
>
I have been tracking XEmacs for this purpose on Win32 for quite some time.
Actually, I would even settle for a decent subset of Unicode. Unfortunately
Win32 Unicode editing is simply not supported in XEmacs stable, and the 21.5
branch does not produce binaries - being beta.
As Stephen put it:
>>Replacing the current Mule internal representation with Unicode is
>>
>>
planned, but not
>>going to happen soon.
>>
>>
As it looks right now, simple unicode editing should be done with GNU Emacs,
and to get good unicode support, you have to use (g)vim. It works well even
on Win32, but using it is a pain for someone who has emacsen in their nervous
system.
I'm sorry Stephen, but the amount of UTF-8 encoded messages in newsgroups and
mail is going up all the time. Win32 XEmacs stable has problems even with
ISO-8859-15, not to mention UTF-8. We need a release of 21.5 to get Mule
support on Win32. Even then, any advanced unicode usage would be beyond
XEmacs.
As I have noted earlier, I can't build 21.5 myself, but I could provide test
results if a (net)installer would be provided. I've tried out some of the
unofficial 21.5 builds that have been available, but submitting bugs on old
versions is useless.
I honestly find it fairly tragic that XEmacs running on a Unicode-native OS
fails to provide any decent support for Unicode. As it stands, I can't
recommend XEmacs on Win32. And as I tend to support cross-platform
applications, that means I can't recommend XEmacs on any platform.
I wish this would change. But I don't see that on the horizon.
Camillo
10:12 a.m.
New subject: Characters lost when saving UTF-8 file on win32 (21.5.16)
>>>> "Vladimir" == Vladimir Weinstein
<vweinste(a)earthlink.net> writes:
Vladimir> I was able to build 21.5 easily on Win32 - after some
Adrian will be pleased to hear that. He's done a fair amount of
fiddling with the Makefile.
Vladimir> initial wrestling with config.inc, I was fine. I'm using
Vladimir> MSVC.
If you think we can improve the default config.inc, please send a diff
(or the whole thing) of config.inc and your Installation file to
xemacs-beta. Adrian Aichner builds regularly on Win32 with VC, and
(probably because he does it a lot ;-) never complains about
config.inc issues.
Vladimir> Also, I frequently need to identify a codepoint under
Vladimir> the cursor (properties etc...). This is a great task to
Vladimir> do with xemacs - if it only supported Unicode well.
Ah, then you might be interested in XEmacs/UTF-2000/CHISE. It is an
independently maintained fork of XEmacs for linguistics research,
specifically for working with character set databases. The home page
is in Japanese, but the links to screen shots and download pages are
easily identified.
http://www.m17n.org/chise/xemacs
I might be able to provide English translation for the installation
documents if the quantity is not too huge.
NO WARRANTEE:
0. I haven't built this fork for three years.
1. I don't know if it builds on Windows, Cygwin or MSVC. Cygwin
probably comes close. If you're willing to go so far as to build it
on Unix for evaluation, and attempt a build on either or both Windows
platforms, then I'll match your Windows builds and tell you whether
I'm willing to actually diagnose and fix any problems that show up.
2. It is possibly _huge_; at one time the dumped binary was 35MB on
my Linux box. I can definitely help you with that; there is no need
to dump the character databases, and it should be a trivial change to
either load them all at run-time, or at least reduce the set you need.
3. I have no idea about the quality of the code; the original design
and implementation were terribly flawed for a production editor,
that's why it forked---it was good enough for its developers, they
weren't interested in dealing with nitpicks. My guess is that it's
probably very stable for day-to-day work in Oriental languages,
especially Japanese, and YMMV for Devanagari and Arabic.
4. It may expect you to have a wide variety of unusual fonts installed.
--
Institute of Policy and Planning Sciences http://turnbull.sk.tsukuba.ac.jp
University of Tsukuba Tennodai 1-1-1 Tsukuba 305-8573 JAPAN
Ask not how you can "do" free software business;
ask what your business can "do for" free software.
9:04 a.m.
New subject: Characters lost when saving UTF-8 file on win32 (21.5.16)
>>>> "Camillo" == Camillo Sars
<Camillo.Sars(a)F-Secure.com> writes:
Camillo> I'm sorry Stephen, but the amount of UTF-8 encoded
Camillo> messages in newsgroups and mail is going up all the time.
I'm aware of that. However, I've never seen any, and the majority of
active XEmacs developers can live fine in 7 bits. For the core of
developers it's currently a fairly academic issue.
Camillo> I honestly find it fairly tragic that XEmacs running on a
Camillo> Unicode-native OS fails to provide any decent support for
Camillo> Unicode.
Sure. But it's also tragic that there are basically no developers
coming forward from that community to help out with XEmacs, compared
to the number of downloads we're seeing on the netinstaller URL.
While I'm interested in getting the Unicode support right because it's
the right thing to do, I can't put high priority on it myself. It
won't do anything to improve the project as a whole. I'd like to do
it for the sake of all the people who are using and supporting XEmacs
through discussion and bug reports and feature requests, and need
Unicode as soon as possible. But we have to put priority on the
things that the current developers, and the people who will provide
the next generation of XEmacs developers, care about. My best bet is
that Unicode is not high on their list, not yet. And of course most
core developers are even less altruistic than that; they just work on
the things that make their lives easier.
--
Institute of Policy and Planning Sciences http://turnbull.sk.tsukuba.ac.jp
University of Tsukuba Tennodai 1-1-1 Tsukuba 305-8573 JAPAN
Ask not how you can "do" free software business;
ask what your business can "do for" free software.
7515
days inactive
7515
days old
3 comments
3 participants
participants (3)
-
"Särs, Camillo" -
Stephen J. Turnbull -
Vladimir Weinstein