Apache OpenOffice (AOO) Bugzilla – Issue 19060
Zip dictionaries package organization
Last modified: 2013-02-24 20:34:13 UTC
Hi, First, some CC's will be added when a zip language file will be modified I ask the concerned readers to check the package and post a reply to tell if it is correct or not I'll try to put a detailled Log on what has been modified for each package I post a small document concerning DicOOo but sumaryzing the zip file structure Thanks Laurent
Created attachment 9002 [details] Some guidelines on zip file architecture
Created attachment 9003 [details] French SpellChecker
*fr_FR.zip - sgauti@openoffice.org merged GPL-fr to README_fr_FR.txt
Created attachment 9005 [details] uk_UA spellchecker
*uk_UA.zip - lvm@mystery.lviv.net merged copyright to readme
Created attachment 9006 [details] sv_SE spellchecker
*sv_SE.zip - stekman@sedata.org merged following files to readme - contributors - copying - copyright -news *sv_SE.zip - stekman@sedata.org Removed hyph_sv_SE.dic from package
Created attachment 9007 [details] sv hyphenation
*hyph_sv.zip - stekman@sedata.org creation need a README_hyph_sv.txt file need be to added in list
Hi, I think we need to think this through a bit more before racing ahead here. Since we have to open up every zip and repackage it we should: 1. always make sure the package name has both language and region information (so hyph_sv.zip should really be hyph_sv_SE.zip) since there may be cases when one language is spoken in two differnt regiosn that have different hyphenation rules due to different accents and pronounciation of the same word. This happens with en since US and GB do hyphenate differently. We should also think about forcing the README_* to be really two files since mixing encodings in any one file will create problems for most text editors! So my idea os simply catting the files together messes up on some of these. Perhaps we need to think about the correct format for the README file since mixing encodings is going to mess things up. Perhaps if we go with pure UTF-8 we could then have a text file that would be readable under ascii/latin1 as is but that could support multiple encodings in the same file. Are there any/many text editors that will properly handle utf-8 chars for all platforms (Windows, Linux, Solaris, MacOSX, FreeBSD, Irix, etc). I think emacs and kate under Linux will proeprly handle utf-8 files. What do you think? Should we convert every *.txt file to utf-8 first before concatenating them? Kevin
HI Kevin I agree with you Sorry if my question is dummy, but, what about an swx file handling (if possible) multiple charset for README ? Laurent
Hi, Would and swx version work for your "preview contets of Readme"? Perhaps the right way to do this is to leave all of the peices as they are but create a new ReadMe_*.swx which contains all of them properly converted into a writer document. Would DicOOo be able to show that file when the box to review the readme is selected? Kevin
Hi, Adding myself to CC on this one. Kevin
Hi DicOOo can handle sxw files as it uses OOo to display the readme files I just have to now it before - sxw is a solution (need changing 2 lines of code) an other - I can dive in OOO Zip package api in order to retreive all files except .aff, .dic ... and display them In the next coming version, all files (for different dictionaries) are merged into one only file when edited - I need to invitagate a more deeper though ... This last solution would work but i expect some problems on character sets when doing it dynamically and we will the not standardize the Zip content Let me know Laurent
Hi, I think since the installer can only be used inside of OOo and it is only 2 lines of code to change we should keep all files in all zips the same as they are now except for the **addition** of a README_lang_country.swx file that combines all of the extra pieces. If that sounds like a good idea, let's do that instead of just merging differnt text files with differnt encodings into one. Richard and Laurent are we in agreement? Kevin
it's ok for me I wait for Richard agreement before beginning to modify my current building version (Beta5.5) Should it be possible to make a call on lingucomponent list to have people doing these sxw files (especially for "strange" character sets) do we expect more than one sxw file/zip archive or all will be merged into only one file ? I prefer only one, it should be sufficient Laurent
Downloaded the Architecture notes, and the first 4 dicts. I will upload to stardiv this afternoon. About the README_xx_XX.sxw. I can handle it but it seems it would be more accessible as a text file. I prefer using a text editor for small files compared to OOo. Maybe my system doesn't have enough horsepower. Small files are slow. But, on the other hand, if you guys feel it's a good move, I can do it. Just need to redo my group of files (dicts) to create an sxw. Perhaps we should agree on a layout for the sxw? I'll take a look at the architecture.sxw. By the way, it makes downloading easier if you don't include your local path in the filename, leave only the filename.ext. One other note, I'm on a dialup line, flatrate, but I don't stay online constantly. Therefore, I don't check by issue number for messges unless I receive an email. Just in case I seem slow in answering. :-)) Richard.
DicOOo is redundant now given super-dooper extensions, right?
closing