|
|
Same applies to https://weblate.gnunet.org/translate/gnu-taler/wallet-android/en/?checksum=baf2c5eb3055f3ab with %1$d × %2$s representing <quantity> × <token name> (wallet/src/main/res/values-*/strings.xml, string 173). Translators have no clue what to do with %1$d × %2$s, if they are clever enough they leave the string as is but that's not always expectable. Even in languages with writing direction right-to-left those variables make no sense to translate. |
|
|
avalos
2025-09-27 21:25
developer
~0026046
Last edited: 2025-09-27 21:27
|
This is incorrect, this can indeed change based on language, e.g. RTL languages such as Arabic and Hebrew.
Besides, who knows if some language has its own quantity indicator other than "×".
Also, any incorrect translation of a string with placeholders (regarding correct syntax) should be detected during review, right?
|
|
|
Stefan
2025-09-29 11:54
developer
~0026062
Last edited: 2025-09-29 12:25
|
The argument that variables in strings should be object for translations is definitely right in a case like https://weblate.gnunet.org/translate/gnu-taler/taler-bank-spa/en/?checksum=dbc6088dc9e7daa7 where payto://iban/[receiver-iban]?message=[subject]&amount=[%1$s:X.Y] contains some expressions to translate like in German payto://iban/[IBAN des Empfängers]?message=[Buchungsvermerk]&amount=[%1$s:X.Y]. We had this case already discussed in https://bugs.gnunet.org/view.php?id=9524 a long time ago. In Spanish, the translation was a bit awkward until today with payto://iban/[iban-destinatario]?message=[asunto]&amount=[%1$s:X.Y] which I changed to payto://iban/[IBAN del destinatario]?message=[asunto]&amount=[%1$s:X.Y] in https://weblate.gnunet.org/translate/gnu-taler/taler-bank-spa/es/?&offset=51. Please go for a quick review, if I may please you. Thanks.
Otherwise, we MUST avoid translations of variables in Weblate. In the given two cases
https://weblate.gnunet.org/translate/gnu-taler/wallet-android/en/?checksum=3f0518f73d0b67ba#translations (%1$s × %2$s which stands for <quantity> × <price>) and
https://weblate.gnunet.org/translate/gnu-taler/wallet-android/he/?checksum=baf2c5eb3055f3ab (%1$d × %2$s which stands for <quantity> × <token name>)
the necessity to translate %1$d × %2$s is none as there is nothing to translate maybe except for the mathematical operand × which might be displayed in a different way across languages like e.g. Japanese, Korean, Chinese... (where deepl renders a big cross instead of an x). Indeed, in some really weird dialects like Ethiopian an operand like the multiplier x could somehow be written in a very strange way, though.
If you look on the Hebrew translation done by a mother-tongue speaker, you will find out that he left the order of variables unchanged in https://weblate.gnunet.org/translate/gnu-taler/wallet-android/he/?checksum=baf2c5eb3055f3ab and
https://weblate.gnunet.org/translate/gnu-taler/wallet-android/en/?checksum=3f0518f73d0b67ba, even though Hebrew is an RTL language!
Anyways, the actual string deals with a multiplication. It seems superfluous to argue in which order the variables appear in a multiplication. A different case would be the one mentioned above. In the cases with %1$s × %2$s, the × should be described on Weblate as a mathematical operand so that translators know what the x is meant to be. But I tend to avoid those two translations.
|
|
