New dictionary format of firmware 2.14

[AlPe]

Doesn't the current code already do that? Other scripts will use characters with ord() > 127.

If ord(x) > 127, then character x is considered ok.

In other words: right now, a keyword goes into 11.html if and only if it its 2-character prefix contains an ascii char x with ord(x) <= 127 which is not a letter. And I have tested that, in this case, the keyword is correctly retrieved.

Am I missing something?

[tshering]

I did not look at your code. I was only guessing your code would implement the idea

move all keywords containing a non-letter (defined as [A-Za-z]) in the first two characters into 11.html

[AlPe]

Ah, ok. No, the code is supposed to do what you asked for.

1) If the keyword is only 1 character long, the code appends an "a" to it.
2) Then it looks at the first two characters of the keyword. If both are (an ASCII letter or have ord() > 127), then the keyword goes to a suitable ??.html file. Otherwise, it goes to 11.html.

There might be still a small "slack", for example in the mentioned case of "o'clock". But I tested that putting "o'clock" into 11.html works, so I guess the current code is fine.

[tshering]

I would like to mention that entries with identical keywords (and different definitions) have to follow directly on each other. If not, the Kobo displays only the first occurrence. For "economy" reasons, I did not check for this with my Japanese-English dictionary. Therefore, I have to do it anew.

[AlPe]

Penelope sorts the keys (i.e., keywords) before processing and outputting them, so the code should be fine w.r.t. this issue.

Also, in the current version I wrapped the content of a word+definition into a <div>, so that two identical keywords with different definitions will be cleanly displayed one below the other.

[tshering]

Today I run into a problem. I made a French dictionary. To my disappointment, no word with an accented vowel in the first two letters did show up. The reason for this was evidently the encoding of the filenames inside the zip file. This was a surprise for me, since my Japanese dictionaries work perfectly, so why should not the French dictionary? Anyhow, I tried zipping the file under Ubuntu, rather than under Windows, and the dictionary was working fine.
The culprit is a strange default behavior of 7-zip under Windows (or maybe I should rather say the culprit was me not being aware of it). 7-zip encodes only those filenames in utf-8 that contain characters not supported by the local codepage. Therefore, one has to enforce utf-8 encoding, under Windows.

7z a -tzip -mcu=on dicthtml.zip *.html words

[gouni]

Tshering

I give you my mini table of letter for french

Open with notepad++


http://www.mediafire.com/?ptmn10cplqr7kaa

[gouni]

I need your help to use 7zip

I use the windows console

i have 300 fichiers xx.html xx= ab ac ad ae

I want to compress each into gzip

The name of my *.html is in ansi code windows and i will transform to utf-8


to have

my 300 files xx.html.gz

and have accented-utf8 characters

What command should I use?

Thank

aa.html = aa.html.gz
.
aé.html = aé.html.gz
.
ûf.html = ûf.html.gz
.
.
zz.html = zz.html.gz

[ShellShock]

Originally Posted by tshering

Today I run into a problem. I made a French dictionary. To my disappointment, no word with an accented vowel in the first two letters did show up. The reason for this was evidently the encoding of the filenames inside the zip file. This was a surprise for me, since my Japanese dictionaries work perfectly, so why should not the French dictionary? Anyhow, I tried zipping the file under Ubuntu, rather than under Windows, and the dictionary was working fine.
The culprit is a strange default behavior of 7-zip under Windows (or maybe I should rather say the culprit was me not being aware of it). 7-zip encodes only those filenames in utf-8 that contain characters not supported by the local codepage. Therefore, one has to enforce utf-8 encoding, under Windows.

7z a -tzip -mcu=on dicthtml.zip *.html words

Great find!

[tshering]

Gouni,

there is no need to convert the filenames already when creating the gz-files. Do

for %i in (*.html) do 7z a -tgzip "gz\%i" "%i"

then change to the gz sub-directory and do

7z a -tzip -mcu=on dicthtml.zip *.html words

If the location of 7z.exe is not known to the system give the full path, for instance "C:\Program Files\7-Zip\7z" (including the quotation marks).

[tshering]

Originally Posted by gouni

Tshering

I give you my mini table of letter for french

Open with notepad++


http://www.mediafire.com/?ptmn10cplqr7kaa

Thank you for the table. I am not sure what I am supposed to do with it. Is this a part of the description of your problem?

[gouni]

Thank you Tshering, yes is a part of description of my problem.

I do not know it but I listen for the Council.

I've done with notepad + my text is in UTF8, but I saved my files in utf8 names.

I work with windows, I don't know unix, I do not know programming.

All my file names éa éb ...... ûm html are windows character.


I read the text 7-zip.chm to try to understand, its difficult for me.

[tshering]

Originally Posted by gouni

All my file names éa éb ...... ûm html are windows character.

This is fine. Make a back up copy of your files for safety reasons and execute the two commands given here. Tell us if it worked.

[gouni]

Thanks for giving your time tshering. It is very difficult, never I did that. I get to the second command (a day to understand but I get)


I know do this :

7z a -tzip -mcu=on e:\zzzarch e:\zztest\*.html

give me : all files in zztest go corectly in file zzzarch.zip



but I do not succeed the first command

My organisation folder in e:
e:\zztest = my folder with all files *.html

7z" for %i in (*.html) do 7z a -tgzip" "gz\%i" "%i"

(*.html) = e:\zztest\*.html do not correctly ?

[tshering]

Do not replace "*.html" by "e:\zztest\*.htm". Instead change to this directory by executing

cd e:\zztest

and then execute

for %i in (*.html) do 7z a -tgzip "gz\%i" "%i"

Edit:
If you are not already on e: then execute first

e:

before you change to "e:\zztest".