I'm running Windows XP and I switch between 3 languages: English, French and Russian. Language for non-Unicode programs is set to Russian. When I try to backup any file with French characters, it says that the file was deleted before it could be backed up. I think the same thing happened when I tried to backup files with Russian characters and the non-Unicode language was set to English. Oh, and I'm running 1.10.2.
Any way to fix this? Thanks!
Problem with foreign characters
Unicode filenames
We introduced the capability to backup files with Unicode characters in version 1.9. As we use the standard zip file format we use the zip file’s comment option to hold the actual filename. This works in our tests but there may be problems where the language is switched on-the-fly. Please define an example where this is going wrong, along with what the filename should be and the language settings on the PC so that we can duplicate the problem here.
This forum does not allow UNICODE characters, so you may have to send a word document containing the filenames.
Barry
This forum does not allow UNICODE characters, so you may have to send a word document containing the filenames.
Barry
I have 3 languages installed: English (United States), French (Canada) and Russian and I can switch between the 3 with the language bar.
The language for non-Unicode programs in Regional Settings is Russian.
I create a file with French characters by switching to French and using the filename testé.txt, where the last e has an accent aigu like here: http://www.rudyjk.com/aisbackup/ . BTW you can type that letter using the /? key at the bottom right of the keyboard (in French mode of course). When I try to backup the folder containing that file, it says "File deleted before backup.". And BTW, in the session log shown within AISBackup the accent is not shown, I only see a normal 'e'.
Also, the language is switched to English (Canada) when I run AISBackup but I don't think this makes any difference.
The language for non-Unicode programs in Regional Settings is Russian.
I create a file with French characters by switching to French and using the filename testé.txt, where the last e has an accent aigu like here: http://www.rudyjk.com/aisbackup/ . BTW you can type that letter using the /? key at the bottom right of the keyboard (in French mode of course). When I try to backup the folder containing that file, it says "File deleted before backup.". And BTW, in the session log shown within AISBackup the accent is not shown, I only see a normal 'e'.
Also, the language is switched to English (Canada) when I run AISBackup but I don't think this makes any difference.
Code pages
Although we seem to have unicode working we appear not to have AISBackup working with "Language for non-Unicode programs". The filename tést.txt is Hex 5400E900730074002E00740078007400 unicode
and Hex 546573742E747874 ansi in other word é got translated to e by AISBackup. We should be able to work around this, but AISBackup may still display test.txt or the equivalent Russian letter teйt.txt.
Barry
and Hex 546573742E747874 ansi in other word é got translated to e by AISBackup. We should be able to work around this, but AISBackup may still display test.txt or the equivalent Russian letter teйt.txt.
Barry
Ah, I see (I think). I'm curious, if you have the time for some explanation, how exactly are file names stored in Windows? 2 versions for each file, ANSI and Unicode? Sometimes only in ANSI? And if it is in ANSI, then that's where the 'Language for non-Unicode programs' is used to decide which characters to display?
Seems to be a huge mess, because for example I just downloaded a Word file via GMail with an 'é' in its filename but it got converted to 'и' when I saved it to disk. Or maybe that's just because GMail stores it in ANSI.
Seems to be a huge mess, because for example I just downloaded a Word file via GMail with an 'é' in its filename but it got converted to 'и' when I saved it to disk. Or maybe that's just because GMail stores it in ANSI.
Unicode filenames
Microsoft hold all filenames as Unicode in FAT and NTFS drives, additionally they hold the small filename (8.3) as ANSI characters. There are 65535 possible Unicode characters plus $0000 which is often used as a string delimiter. There are 255 possible ANSI characters, again $00 is often used as a string delimiter. (not all of the characters can be used in filenames, e.g. *, ?)
Most programs written for Windows only support ANSI characters as the programming environments we use usually only support ANSI by default.
We incorrectly assumed that non translated characters were set to ?, so any filename with a ? was a Unicode named file. We did not consider filenames like testé.txt which translate to testé.txt when using English code pages and teste.txt when using Russian code pages.
We have updated AISBackup and firsts tests are looking good, so the next upload of version 2.0 beta will support (or have better support) for Unicode.
Barry
Most programs written for Windows only support ANSI characters as the programming environments we use usually only support ANSI by default.
We incorrectly assumed that non translated characters were set to ?, so any filename with a ? was a Unicode named file. We did not consider filenames like testé.txt which translate to testé.txt when using English code pages and teste.txt when using Russian code pages.
We have updated AISBackup and firsts tests are looking good, so the next upload of version 2.0 beta will support (or have better support) for Unicode.
Barry
Works great, except for this one file which I zipped here: http://www.rudyjk.com/aisbackup/AGATHODA.zip . My settings are as before. This is just a font that I downloaded and never used, really, so I have no idea what charset it's in. Looks like English characters but I guess not.