From: Pawel K. <pk...@be...> - 2004-04-14 21:21:36
|
Hi, I have just tried current CVS joe. Syntax highlightning really rocks. Few notes though: 1. joe -syntax XXX does not work. It ignores the syntax parameter. 2. The option menu (under ctrl-t) is a bit broken. See http://tfuj.pl/joe-menu.png 3. How to disable utf8? Joe seems to assume all files are utf8 and I can't read my iso-8859-2 files. 4. Incorrect jsf file causes joe segfaults, but I can understand it as it is the beta version. Sorry if it is all known. I have prepared (ugly) simple syntax higlightning file for XML: http://tfuj.pl/xml.jsf. It will change very soon as this is a working version. Feel free to use it and correct it. take care, pkot -- mailto:pk...@be... http://www.gnokii.org/ |
From: <ja...@av...> - 2004-04-15 02:45:32
|
From: Pawel Kot <pk...@be...> X-X-Sender: pk...@ur... To: joe...@li... Message-ID: <Pin...@ur...> MIME-Version: 1.0 Content-Type: TEXT/PLAIN; charset=US-ASCII X-Spam-Score: 0.0 (/) X-Spam-Score: 0.0 (/) X-Spam-Report: Spam Filtering performed by sourceforge.net. See http://spamassassin.org/tag/ for more details. Report problems to http://sf.net/tracker/?func=add&group_id=1&atid=200001 Subject: [joe] joe-current: syntax and utf8 Sender: joe...@li... Errors-To: joe...@li... X-BeenThere: joe...@li... X-Mailman-Version: 2.0.9-sf.net Precedence: bulk List-Unsubscribe: <https://lists.sourceforge.net/lists/listinfo/joe-editor-general>, <mailto:joe...@li...?subject=unsubscribe> List-Id: <joe-editor-general.lists.sourceforge.net> List-Post: <mailto:joe...@li...> List-Help: <mailto:joe...@li...?subject=help> List-Subscribe: <https://lists.sourceforge.net/lists/listinfo/joe-editor-general>, <mailto:joe...@li...?subject=subscribe> List-Archive: <http://sourceforge.net/mailarchive/forum.php?forum=joe-editor-general> X-Original-Date: Wed, 14 Apr 2004 23:21:23 +0200 (CEST) Date: Wed, 14 Apr 2004 23:21:23 +0200 (CEST) X-Avici-MailScanner: Found to be clean, Found to be clean X-MailScanner-From: joe...@li... X-Avici-MailScanner-Information: Please contact the ISP for more information Status: R Pawel Kot <pk...@be...> wrote: >Hi, >1. joe -syntax XXX does not work. It ignores the syntax parameter. I will look into this. I'm thinking of revamping the option logic a bit, and in any case there needs to be a method of setting the current syntax from within the editor. >2. The option menu (under ctrl-t) is a bit broken. See >http://tfuj.pl/joe-menu.png It's a feature! Really, I made menus four lines, mainly for file selection. But the same code is used for ^T options. There will be a setting for the number of lines. The yellow selection bar is temporary: it was just a test for color and will be removed (back to white). >3. How to disable utf8? Joe seems to assume all files are utf8 and I can't >read my iso-8859-2 files. ^T U flips it on and off, but I'm thinking that you need to set '-asis' in the joerc file (I haven't added code yet to turn utf-8 on in utf-8 locales). I had thought that someone had added code to do this (-asis) automatically, but perhaps not: I will look into this. >4. Incorrect jsf file causes joe segfaults, but I can understand it as it >is the beta version. Yeah, this needs work. Just don't make any mistakes :-) >Sorry if it is all known. I have prepared (ugly) simple syntax >higlightning file for XML: http://tfuj.pl/xml.jsf. It will change very >soon as this is a working version. Feel free to use it and correct it. No no, keep sending bug reports. In particular, I have difficulty testing joe for non-ascii character sets. >take care, >pkot >-- >mailto:pk...@be... >http://www.gnokii.org/ |
From: Pawel K. <pk...@be...> - 2004-04-15 08:23:59
|
On Wed, 14 Apr 2004 ja...@av... wrote: > >2. The option menu (under ctrl-t) is a bit broken. See > >http://tfuj.pl/joe-menu.png > > It's a feature! Really, I made menus four lines, mainly for file selection. > But the same code is used for ^T options. There will be a setting for the > number of lines. The yellow selection bar is temporary: it was just a test > for color and will be removed (back to white). I mean -- it is not nicely aligned. IMHO the layout should be: X option X option X option Sometimes X and/or following space is missing. Look at the middle column. Or being more specific. Currently we have: T Overtype OFF I Autoindent OFF Word wrap ON D Tab width 8 Left margin 1 Right margin 77 X Rectangle mode OFF Indent char 32 Indent step 1 french spacing OFF Highlighting OFF no tabs OFF IMHO it should look like: T Overtype OFF I Autoindent OFF Word wrap ON D Tab width 8 Left margin 1 Right margin 77 X Rectangle mode OFF Indent char 32 Indent step 1 french spacing OFF Highlighting OFF no tabs OFF > >3. How to disable utf8? Joe seems to assume all files are utf8 and I can't > >read my iso-8859-2 files. > > ^T U flips it on and off, but I'm thinking that you need to set '-asis' in > the joerc file (I haven't added code yet to turn utf-8 on in utf-8 locales). > I had thought that someone had added code to do this (-asis) automatically, > but perhaps not: I will look into this. I tried ^T U, but it doesn't help. Using -asis also doesn't help. I tried it already. > No no, keep sending bug reports. In particular, I have difficulty testing > joe for non-ascii character sets. I can be non-ascii beta-tester ;-) pkot -- mailto:pk...@be... http://www.gnokii.org/ |
From: <ja...@av...> - 2004-04-15 14:34:56
|
>IMHO it should look like: >T Overtype OFF I Autoindent OFF Word wrap ON >D Tab width 8 Left margin 1 Right margin 77 >X Rectangle mode OFF Indent char 32 Indent step 1 > french spacing OFF Highlighting OFF no tabs OFF Ok, I've improved this: only things like 'Word wrap' start in the first column because I'm keying off of the 'W'. Otherwise we need to say: 'W Word wrap ON'. >I tried ^T U, but it doesn't help. Using -asis also doesn't help. I tried >it already. I checked in a fix: I was assuming that the terminal was always UTF-8. Now I assume that the terminal is UTF-8 if you are editing a UTF-8 file, but really I have to do the iconv() conversion stuff, so that the terminal can be UTF-8 and the file non-UTF-8 and visa-versa. |
From: Pawel K. <pk...@be...> - 2004-04-15 18:49:41
|
On Thu, 15 Apr 2004 ja...@av... wrote: > >I tried ^T U, but it doesn't help. Using -asis also doesn't help. I tried > >it already. > > I checked in a fix: I was assuming that the terminal was always UTF-8. Now > I assume that the terminal is UTF-8 if you are editing a UTF-8 file, but > really I have to do the iconv() conversion stuff, so that the terminal can > be UTF-8 and the file non-UTF-8 and visa-versa. Thank you. It works now. pkot -- mailto:pk...@be... http://www.gnokii.org/ |
From: <ni...@re...> - 2004-04-15 15:11:20
|
>I checked in a fix: I was assuming that the terminal was always UTF-8. Now >I assume that the terminal is UTF-8 if you are editing a UTF-8 file, but >really I have to do the iconv() conversion stuff, so that the terminal can >be UTF-8 and the file non-UTF-8 and visa-versa. > > The proper way to find out the current terminal encoding would be to read the value of nl_langinfo(CODESET). |
From: <ja...@av...> - 2004-04-15 15:42:56
|
<ni...@re...> wrote: >The proper way to find out the current terminal encoding would be to >read the value of nl_langinfo(CODESET). I know, but it is not enough information if the answer is UTF-8. Suppose that LC_ALL=de_DE.UTF-8 What you really want is (a) the terminal is UTF-8, and (b) the most likely non-UTF-8 coding in use is ISO-8859-1. I'm thinking of calling setlocale twice as a hack to get this information: s=getenv("LC_ALL"); /* Actually try LC_ALL, LC_CTYPE, LANG */ remove .UTF-8 from end of string if it's there setlocale(LC_TYPE,s); non_utf8_codeset = nl_langinfo(CODESET); setlocale(LC_TYPE,""); terminal_codeset = nl_langinfo(CODESET); So when you hit ^T U, you flip between UTF-8 and non-utf8-codeset for the file, and JOE will translate the file format to the terminal format with iconv(). What do you all think? |
From: Koblinger E. <eg...@cs...> - 2004-04-15 15:54:48
|
On Thu, 15 Apr 2004 ja...@av... wrote: > What you really want is (a) the terminal is UTF-8, and (b) the most likely > non-UTF-8 coding in use is ISO-8859-1. I can't see the reason for (b). I have LANG=hu_HU (which is latin-2), and I might want to edit latin-1, latin-2, utf-8, koi8-r, euc-jp etc. files. Those who live in Germany may also want to edit a Latin-2 file. > remove .UTF-8 from end of string if it's there > > setlocale(LC_TYPE,s); > non_utf8_codeset = nl_langinfo(CODESET); This stuff (e.g. turning de_DE.UTF-8 into Latin-1, hu_HU.UTF-8 into Latin-2 is a good for some heuristics offering the guessed charset of the file, but still I'd rather choose from a long-long list that does not depend on my locale. <intermezzo> > setlocale(LC_TYPE,""); > terminal_codeset = nl_langinfo(CODESET); This is correct, seems we all agree that this is the way to determine the terminal's behaviour, and the user should not be able to modify it within joe. It can be assumed that the LANG and LC_* variables are set up correctly to reflect this. </intermezzo> > So when you hit ^T U, you flip between UTF-8 and non-utf8-codeset for the > file, and JOE will translate the file format to the terminal format with > iconv(). Back to the previous topic... I can't really see the point in this two-value utf8 vs. non-utf8 story. non-utf8 is not a charset :-))) It should be an option named "file charset" or similar, which could have many values. The heuristics seen above is good to make a guess, but if it's not okay than I want to be able to choose the right charset from a long list, not only from utf8 and my old-fashioned-non-utf8 locale. bye, Egmont |
From: <ja...@av...> - 2004-04-15 18:20:04
|
Koblinger Egmont <eg...@cs...> wrote: >Back to the previous topic... I can't really see the point in this >two-value utf8 vs. non-utf8 story. non-utf8 is not a charset :-))) It >should be an option named "file charset" or similar, which could have many >values. The heuristics seen above is good to make a guess, but if it's not >okay than I want to be able to choose the right charset from a long list, >not only from utf8 and my old-fashioned-non-utf8 locale. Well you're right, it would be the default. I assume the most popular file type for a given location is the old-fashioned-non-utf8 locale of that location. Or is that not a correct assumption? So there's another issue: do you choose just the charset at the prompt, or the entire locale string (hu_HU....)? I'm thinking that I need to call setlocale() again to get isalpha() to work properly for the file. Oh yuck... each buffer could have a different locale... |
From: Koblinger E. <eg...@uh...> - 2004-04-15 21:22:30
|
> Well you're right, it would be the default. I assume the most popular file > type for a given location is the old-fashioned-non-utf8 locale of that > location. Or is that not a correct assumption? Yes, if it's only the default but I can override it to any charset, then it's perfect. > So there's another issue: do you choose just the charset at the prompt, or > the entire locale string (hu_HU....)? I'm thinking that I need to call > setlocale() again to get isalpha() to work properly for the file. Oh > yuck... each buffer could have a different locale... My god... tough question... My idea at this moment is to try to avoid isalpha() completely and try to avoid using locales as much as possible, rather iconv() all the stuff to the internal utf8 representation and have an isalpha-like function call that works for utf8. I'll think on it a little bit more... -- Egmont |