Some typos in strings (Weblate)

l10n
#1

Hi All

as usual, starting the new session of translations, I found some little typos, as follows:

Comment: Fca5T
File: 03102101.xhp
Context: 03102101.xhp\n par_id711996\n help.text

Now it reads:

"is redimensioned".

Should correct to read:

"is resized". (or downsized? reduced?)

#2

If this is a serious question and not a rhetorical one you already
know the answer to (string appears in different context/in different
files, ....) then you should give concrete examples.
Copy the links from weblate please (the chain-symbol in the
Translation box with "copy permalink" hover text)

ciao
Christian

#3

In data venerdì 3 luglio 2020 11:58:29 CEST, Christian Lohmaier ha scritto:

> PS: I'm still wondering why strings "already translated" in 6.3 and 6.4
> are again new to translate in Master... The examples are many, our time
> short.

If this is a serious question and not a rhetorical one you already
know the answer to (string appears in different context/in different
files, ....) then you should give concrete examples.
Copy the links from weblate please (the chain-symbol in the
Translation box with "copy permalink" hover text)

Hi Christian

yes, this is a serious question and I know also the answer. But:

- I use an offline CAT Tool (I suppose many people use one)
- The CAT Tool created, during many years of localisation, a Translation memory and
suggest to me the translated strings (let's say the fuzzy ones, from 100% to 70% of
matches)
- From the TM I'm getting *hundreds* of 100% matching strings, which means that they
are exactly the same strings we got from the previous versions of LibreOffice (consider
from 6.2 to 6.4)

So, my question is: why all this strings cannot be *auto-inserted* in Weblate, eventually
with a "confirmation needed" (for security reasons), in order to facilitate the translator's
job? Is it not possible?

In any case, as I still haven't finished my "previous" check of the strings in order to
eliminate all these 100% matches, I can provide some simple examples, percentage on
side (but I get also long sentences, even with tags):

https://weblate.documentfoundation.org/translate/libo_help-master/textswriter01/it/?
checksum=3797f7853b8bfc5d[1] (100%)

https://weblate.documentfoundation.org/translate/libo_help-master/textswriter01/it/?
checksum=9a9f28b8bdb29e21[2] (100%)

https://weblate.documentfoundation.org/translate/libo_help-master/textswriter01/it/?
checksum=1184304f7e055808[3] (100%)

https://weblate.documentfoundation.org/translate/libo_help-master/textswriter01/it/?
checksum=f7c9fdfb264defb9[4] (100%)

https://weblate.documentfoundation.org/translate/libo_help-master/textswriter01/it/?
checksum=bc1fbf2098f0003e[5] (99.33%)

https://weblate.documentfoundation.org/translate/libo_help-master/textswriter01/it/?
checksum=f4c90c5313674f22[6] (100%)

https://weblate.documentfoundation.org/translate/libo_help-master/textswriter01/it/?
checksum=e5b25ffe5c7de87e[7] (100%)

What I mean is: if I have these full matches in my TM, why cannot they appear
automatically in Weblate?

Thanks for understanding and for your work.

Ciao

#4

Hi Valter,

yes, this is a serious question and I know also the answer. But:
[…]
So, my question is: why all this strings cannot be *auto-inserted* in Weblate, eventually
with a "confirmation needed" (for security reasons), in order to facilitate the translator's
job? Is it not possible?

You might have picked bad examples, but for all you cited, the old
translations are showing up on the Machine Translation tab and are
offered for insertion. In this case they come mostly from the amagama
based TM, i.e. the one that was done based on the translations
imported from the old branches from pootle.

https://weblate.documentfoundation.org/translate/libo_help-master/textswriter01/it/?
checksum=3797f7853b8bfc5d[1] (100%)
[some more examples]

What I mean is: if I have these full matches in my TM, why cannot they appear
automatically in Weblate?

Well, they do, on the machine translation tab. For old strings (that
already were translated on pootle) from the "tmserver" source (the
local amagama server), and for translations done on weblate it would
be from weblate's tranlsation memory. (that in old version was rather
slow, but with the version we deployed last weekend also those results
should be quick to load in.

https://translations.documentfoundation.org/translate/libo_help-master/textswriter01/it/?checksum=3797f7853b8bfc5d#machine

Is one of those strings that were changed during weblate times and
thus have a match from weblate's TM as well as similar matches from
amagama..

So maybe I'm completely missing your point.

ciao
Christian