kempelen
@Robin Wow, very great finding there. Also all your answers makes the picture clearer, thank you!
I cloned your repo now, and tested with “cat -A” (that I asked from you, but my questions were overwhelming, sorry). Attached picture shows this contains the DOS CR LF linefeeds, and this may be the cause of having all those #13 codes in the translation process. I checked out the original repo too, and anticapitalista’s files in FAQ folder also indeed in DOS format.
I don’t know whether Transifex will “like it” if you convert these files to Unix line ending (LF, instead of CR+LF)!
Easiest way to convert it is the dos2unix package, but pay attention this converts by overwriting the file! I tried dos2unix icewm.html and it indeed removed the “^M” from line ends (that is CR, #13). When you have time, please make a test that you fix a file and upload to TESTING AREA of Transifex to see whether that improves the strings (hopefully eliminates #13).
If that allows uploading and does not break existing translations and removes #13’s, that already improves how the guys can continue on the translation.
The .txt files that you found are indeed the .adoc files, named “txt” for correct mime/type I guess. I also don’t know AsciiDoc, but I used similar “HTML builders” before, e.g. for building gallery out of just images and a template. Similar idea. You have a master template (e.g. .conf), you have all your content articles, and they get merged together. So if you need to add a menu, you add it at one place (or if you need to change something in all pages). (For examine, adding a “languages” menu at top… 😉 )
This is the program anticapitalista uses for making that process: https://asciidoctor.org/
(This was noticed by Claude Code in META GENERATOR of the files, while we were working on the development, you can see if you find “asciidoc” here (fix character encoding in the View menu!), it also explains the #13 problem there: https://c64.rulez.org/~lion/antiX/antix-faq-program.txt )
I’ll get back to Claude asking whether the process that “I” developed, in which HTML files are in the translation work, is still sufficient if we have access to the source “adoc” files. But I suspect the process is OK. I’m 99% sure this is OK. Especially because “poa4” workflow is invented for that directly.
Please take a brief read at po4a’s “what is this”, just so that we are on the same page (I mentioned this program to you 10 times so far 😉 Including antixscreensho2 help discussion): https://www.po4a.org/
“Makefile” is a cool thing. It’s very basic (yet advanced), it defines: what steps to take to do a task. E.g. “make spell” will find section spell and execute it, most of it is BASH script, that you know well (first line is a comment here with @ and #)
spell:
@# ispell -m -S -p ispell_english *.txt
for f in $(TEXT_FILES); do aspell -c $$f; doneBut if we can stay on HTML side of the workflow, we are totally cool without all this. We just need to be able to
– update the EN text (adoc txt files),
– rebuild the EN HTML,
– re-push into the translation work.
– (then see how that re-push works. This is not well tested in either of the two processes. On my side this is entirely up to po4a and GNU gettext, on your side, this is up to Transifex. Whichever is better or whichever we trust better. 🙂 )
I’ll need to read back this thread to re-ask questions that were forgotten, I’ll do that soon. I wanted to work on the HU translation, but cannot reach that ever. 🙂
One request I have @Robin: please test the Unix conversion and what Transifex says when you update the file with that (backup translation first, because it may be disastrous if every line is a new structure for it due to line ending change). My expectation is that #13 trash will be gone. That will make translation easier.
“My process” is 100% compatible with existing Transifex HTML files, so any further or previous work on TX can be loaded to my project if we later decide to switch. (Assuming we don’t add HTML sections on translation.)
@marcelocripe I can do the English text, I work for a foreign company since 2000, alone from Hungary by now, so I don’t even speak or write a Hungarian word in work hours, except to my family (working from home). 🙂 So I have some practice, even if I have never been on any English course in any of my schools. 🙂 (We had to learn Russian here in Hungary in elementary and middle school, you know, back then we had the Russian army still here. :-))) )
It’s not me who makes the process so rigid, it’s the tool set that I’m using. Please take a brief look, may be useful for any translator, just to know this project exists: https://www.po4a.org/
Regarding additions or deletions to the files, I’ll think about that. I’m sure you’ll need that, even if we update the English text.
I agree with you regarding the IceWM page.
(@Robin: I’ll try the adoc and makefiles and stuff. But probably that does not affect our translation process, only the EN updates, which we don’t do yet. We collect ideas in en_GB and this thread. So nothing is lost from what you or translators made.)
