On Friday September 30 2016 17:37, you wrote to me:
MvdV>> I understand RC50 is asking for just that.
ES> Yes. He said if there will be at least 5 NC's who needs UTF-8 in ES> segments, he will set his nodelist checkers up.
There are some 65 NCs in the R50 section of the nodelist. So to get five, you only need to convince another 4 out of 64 or 6% of the remaining NCs. Should be doable... ;-)
MvdV>> But... now we are talking message editors. This thread started MvdV>> out with the nodelist. And for the nodelist, we do not need to MvdV>> call upon the
ES> But what is the main goal of UTF-8 nodelist distrubution?
The main goal is to give every sysop in Fidonet the opportunity to get his name in the nodelist as spelled in his own native language.
ES> Using it in UTF-compatible software like FTN editors, no?
That is a logical next step. But then we run into a chicken and egg problem.
How does one deal with a chicken and egg problem? Well, certainly not by doing nothing and wait for someone else to take the first step. Teh way to deal with a chicken and egg situatation is to not sit and wait, but do that part that can be done now.
The part that can be done now is that we use the UTF-8 nodelist. For that we have the tools. We have UTF-8 capable text processors and the lastest version of MakeNl can handle UTF-8 encoded nodelits. And the ZC2 has it setup...
So convince as many of your fellow NC to particpate as you can and get your RC to produce a segment for the DAILYUTF along with the regular nodelist segment.
Once the UTF-8 nodelist is widely used by many sysops, there will be demand for FTN readers and other FTN tools that can make use of it. And then who knows what will happen. MvdV>> In the meantime, lacking a fully UTF-8 version of golded, what MvdV>> I do is that I convert it to the local character set for local MvdV>> use. In my case CP850. For that there is iconv. The R50 sysops MvdV>> could do the same and convert their part of the UTF-8 nodelist MvdV>> back to CP866 for local use. Golded has no problem with MvdV>> compiling a nodelist containing eight bit characters. Ik is a MvdV>> "╨║╨╛╤Б╤В╤Л╨╗╨╡╨╣", but we have to start somewhere.
ES> Can't convert it.
You can, but you have to compromise.
ES> iconv "stumbles" on the first non-russian myltibyte symbol with ES> "iconv: (stdin):54:28: cannot convert" error, so I need to ES> use -c option.
I see you are familiar with iconv. Good that saves a lot of explaining. ;-)
ES> And what we will see there? "Blijf_Tonijn" in best case ES> scenario or "Blf_Tnn" and "Bl?f_T?n?n" in the worst cases (like iconv) ES> instead of "Bl─│f_T├╕n─│n".
Indeed. The concatenated Dutch 'ij' does not fit in any of the common code pages. And the o with slash does not fit in CP866.
What I do is that I do some preprocessing. For thet I use sed, the serial editor. You are probably familiar with that as well.
Before calling iconv I run sed with the following script to convert some "impossible" characters to CP850:
Of course your script should look different since you are converting to CP866 and should take care of the characters not present in CP866. Maybe someting like:
