Change Details

When I connect gpg-agent running with LANG=ja_JP, when an error, it can be like this with longer error message: ``` $ gpg-connect-agent "SCD getattr KEY-STATUS" /bye ERR 100663364 オブジェクト内に項目が欠如してい\343 <SCD> ``` It's "Missing item in object" but Japanese translation text is 18 character long, 54 in bytes in UTF-8. The message is truncated in libassuan/src/assuan-handler.c (in the function assuan_process_done), formatting with `%.50s`, thus, it is truncated at the middle of character in the UTF-8 representation. There are two problems here: (1) How the error message is truncated (2) It would be good if an error message is generated at the client side (not server side), and it is client side which does the translation, if needed.

When I connect gpg-agent running with LANG=ja_JP, when an error, it can be like this with longer error message: ``` $ gpg-connect-agent "SCD getattr KEY-STATUS" /bye ERR 100663364 オブジェクト内に項目が欠如してい\343 <SCD> ``` It's "Missing item in object" but Japanese translation text is 18 character long, 54 in bytes in UTF-8, it is truncated in the middle of character, resulted `\343`. The message is truncated in libassuan/src/assuan-handler.c (in the function assuan_process_done), formatting with `%.50s`, thus, it is truncated at the middle of character in the UTF-8 representation. There are two problems here: (1) How the error message is truncated (2) It would be good if an error message is generated at the client side (not server side), and it is client side which does the translation, if needed.

When I connect gpg-agent running with LANG=ja_JP, when an error, it can be like this with longer error message: ``` $ gpg-connect-agent "SCD getattr KEY-STATUS" /bye ERR 100663364 オブジェクト内に項目が欠如してい\343 <SCD> ``` It's "Missing item in object" but Japanese translation text is 18 character long, 54 in bytes in UTF-8, it is truncated in the middle of character, resulted `\343`. The message is truncated in libassuan/src/assuan-handler.c (in the function assuan_process_done), formatting with `%.50s`, thus, it is truncated at the middle of character in the UTF-8 representation. There are two problems here: (1) How the error message is truncated (2) It would be good if an error message is generated at the client side (not server side), and it is client side which does the translation, if needed.