I have to convert a large legacy application to Delphi 2009 which uses strings, AnsiStrings, WideStrings and UTF8 data all over the place and I have a hard time to understand how the new string types work and how they should be used.
The application fully supported Unicode using TntUnicodeControls and there are 3rd party DLLs which require strings in specific encodings, mostly UTF8 and UTF16, making the conversion task not as trivial as one would suspect.
I especially have problems with the C DLL calls and choosing the right type.
I also get the impression that there are many implicit string conversions happening, because one of the DLL seems to always receive UTF-8 encoded strings, no matter how the Delphi string is encoded.
Can someone please provide a short overview about the new Delphi 2009 string types UnicodeString and RawByteString, perhaps some usage hints and possible pitfalls when converting a pre 2009 application?
See Delphi and Unicode, a white paper written by Marco Cantù and I guess
The Absolute Minimum Every Software Developer Absolutely, Positively Must Know About Unicode and Character Sets (No Excuses!), written by Joel.
One pitfall is that the default Win32 API call has been mapped to use the W (wide string) version instead of the A (ANSI) version, for example ShellExecuteA If your code is doing tricky pointer code assuming internal layout of AnsiString, it will break. A fallback is to substitute PChar with PAnsiChar, Char with AnsiChar, string with AnsiString, and append A at the end of Win32 API call for that portion of code. After the code actually compiles and runs normally, you could refactor your code to use string (UnicodeString).
Watch my CodeRage 4 talk on "Using Unicode and Other Encodings in your Programs" this friday, or wait until the replay of it is available online.
I'm going to cover some encodings and explain about the string format.
The slides will be available shortly (I'll try to get them online today) and contain a lot of references to stuff you should read on the internet (but I must admit I forgot the link to Joel on Unicode that eed3si9n posted).
Will edit this answer today with the uploads and the links.
Edit:
If you have a small sample where you can show that your C/C++ DLL receives the strings UTF8 encoded, but thought they should be encoded otherwise, please post it (mail me; almost anything at the pluimers dot com gets to me, especially if you use my first name before the at sign).
Session materials can be downloaded now, including the "Using Unicode and Other Encodings in your Programs" session.
These are links from that session:
Read these:
Marco Cantu, Whitepaper “Delphi and Unicode”
Marco Cantu, Presentation “Delphi and Unicode”
Nick Hodges, Whitepaper “Delphi in a Unicode World”
Relevant on-line help topics:
What's New in Delphi and C++Builder 2009
String Types: Base: ShortString, AnsiString, WideString, UnicodeString
String Types: Unicode (including internal memory layouts of the string types)
String Types: Enabling for Unicode
String Types: RawByteString (AnsiString with CodePage $ffff)
String Types: UTF8String (AnsiString with CodePage 65001)
String <-> PChar conversions: PChar fundamentals
String <-> PChar conversions: Returning a PChar Local Variable
String <-> PChar conversions: Passing a Local Variable as a PChar
Hope this gets you going. If not, mail me and I'll try to extend the answer here.
Note that it does not only hit real string code. It also hits code where PCHAR is used to trawl through buffers, or interface with APIs.
E.g. initialization code of headers that load the DLL dynamically (getprocedureaddress/loadlibray)
It seems almost all my problems come from the automatic conversion on assignments to UTF8String.
I already had old code using UTF8String just to help me think which type of string a variable should contain.
When starting to port my application, I replaced AnsiString with UTF8String for the same reason, but the code depended on UTF8String being just an alias to (classic) AnsiString
Now with the automatic conversion that assumption is no longer true, which created many problems.
Be careful if you use UTF8String when porting from pre-2009 Delphi code!
Another thing to watch out for when passing string between dlls built with different versions of Delphi or C++ Builder is that, starting with 2009, the StrRec part of AnsiStringBase gained two extra fields; codePage and elemSize. They are 2 bytes each (short ints), so the size of StrRec is now 12 bytes instead of 8. This can cause invalid pointer exception problems with memory allocation and destruction, even when the data part of the string seems to transfer ok.
Related
So much information about Unicode but hard for for me to get a conclusion.
I'm working on an multi-language Delphi XE5 application and now I face this problem with this unicode characters. Honestly I don't want to understand the magic behind, I just want to see them work in my application.
Before it was simple. In general use String data type. Now I've read about WideString, UnicodeString, AnsiString and the fact that String in XE5 is compliant with UTF-16
I've tested with WideString and the lating characters like (șțăîâ) are working, but it's still not clear if WideString is the best one or not. Should I use UnicodeString or else?
So, If I should make a multi-language application that support all languages, in the end, what kind of data type should I use? Is it any possibility to maintain String type and get the same results like WideString?
Remark: I use inside my application FireDac components, but this should not matter.
In modern Delphi "string" is a shortcut to "UnicodeString" real data type. Use it unless have to forced to use other types.
WideString is but Delphi pseudonym for Microsoft OLE BSTR type and lacks reference counting. That disables copy-on-write optimizations and makes those string work slower than UnicodeString in general (the very data buffer is copied time and again instead of just passing new pointer to it). Unless you need exactly those features - better use usual strings. However for i18n both those types work enough.
I finally upgraded to Delphi XE. I have a library of units where I use strings to store plain ANSI characters (chars between A and U). I am 101% sure that I will never ever use UNICODE characters in those places.
I want to convert all other libraries to Unicode, but for this specific library I think it will be better to stick with ANSI. The advantage is the memory requirement as in some cases I load very large TXT files (containing ONLY Ansi characters). The disadvantage might be that I have to do lots and lots of typecasts when I make those libraries to interact with normal (unicode) libraries.
There are some general guidelines to show when is good to convert to Unicode and when to stick with Ansi?
The problem with general guidelines is that something like this can be very specific to a person's situation. Your example here is one of those.
However, for people Googling and arriving here, some general guidelines are:
Yes, convert to Unicode. Don't try to keep an old app fully using AnsiStrings. The reason is that the whole VCL is Unicode, and you shouldn't try to mix the two, because you will convert every time you assign a Unicode string to an ANSI string, and that is a lossy conversion. Trying to keep the old way because it's less work (or some similar reason) will cause you pain; just embrace the new string type, convert, and go with it.
Instead of randomly mixing the two, explicitly perform any conversions you need to, once - for example, if you're loading data from an old version of your program you know it will be ANSI, so read it into a Unicode string there, and that's it. Ever after, it will be Unicode.
You should not need to change the type of your string variables - string pre-D2009 is ANSI, and in D2009 and alter is Unicode. Instead, follow compiler warnings and watch which string methods you use - some still take an AnsiString parameter and I find it all confusing. The compiler will tell you.
If you use strings to hold bytes (in other words, using them as an array of bytes because a character was a byte) switch to TBytes.
You may encounter specific problems for things like encryption (strings are no longer byte/characters, so 'character' for 'character' you may get different output); reading text files (use the stream classes and TEncoding); and, frankly, miscellaneous stuff. Search here on SO, most things have been asked before.
Commenters, please add more suggestions... I mostly use C++Builder, not Delphi, and there are probably quite a few specific things for Delphi I don't know about.
Now for your specific question: should you convert this library?
If:
The values between A and U are truly only ever in this range, and
These values represent characters (A really is A, not byte value 65 - if so, use TBytes), and
You load large text files and memory is a problem
then not converting to Unicode, and instead switching your strings to AnsiStrings, makes sense.
Be aware that:
There is an overhead every time you convert from ANSI to Unicode
You could use UTF8String, which is a specific type of AnsiString that will not be lossy when converted, and will still store most text (Roman characters) in a single byte
Changing all the instances of string to AnsiString could be a bit of work, and you will need to check all the methods called with them to see if too many implicit conversions are being performed (for performance), etc
You may need to change the outer layer of your library to use Unicode so that conversion code or ANSI/Unicode compiler warnings are not visible to users of your library
If you convert to Unicode, sets of characters (can't remember the syntax, maybe if 'S' in MySet?) won't work. From your description of characters A to U, I could guess you would like to use this syntax.
My recommendation? Personally, the only reason I would do this from the information you've given is the memory use, and possibly performance depending on what you're doing with this huge amount of A..Us. If that truly is significant, it's both the driver and the constraint, and you should convert to ANSI.
You should be able to wrap up the conversion at the interface between this unit and its clients. Use AnsiString internally and string everywhere else and you should be fine.
In general only use AnsiString if it is important that the Chars are single bytes, Otherwise the use of string ensures future compatibility with Unicode.
You need to check all libraries anyway because all Windows API functions in Delhpi XE replaced by their unicode-analogues, etc. If you will never use UNICODE you need to use Delphi 7.
Use AnsiString explicitly everywhere in this unit and then you'll get compiler warning errors (which you should never ignore) for String to AnsiString conversion errors if you happen to access the routines incorrectly.
Alternately, perhaps preferably depending on your situation, simply convert everything to UTF8.
Stick with Ansi strings ONLY if you do not have the time to convert the code properly. The use of Ansi strings is really only for backward compatibility - to my knowledge C# does not have an equiavalent to Ansi strings. Otherwise use the standard Unicode strings. If you have a look on my web-site I have a whole strings routines unit (about 5,000 LOC) that works with both Delphi 2007 (non-Uniocde) and XE (Unicode) with only "string" interfaces and contains almost all of the conversion issues you might face.
I'm upgrading some ancient (from 2003) Delphi code to Delphi Architect XE and I'm running into a few problems. I am getting a number of errors where there are incompatible types. These errors don't happen in Delphi 6 so I must assume that this is because things have been upgraded.
I honestly don't know what the difference between PAnsiChar and PWideChar is, but Delphi sure knows the difference and won't let me compile. If I knew what the differences were maybe I could figure out which to use or how to fix this.
The short: prior to Delphi 2009 the native string type in Delphi used to be ANSI CHAR: Each char in every string was represented as an 8 bit char. Starting with Delphi 2009 Delphi's strings became UNICODE, using the UTF-16 notation: Now the basic Char uses 16 bits of data (2 bytes), and you probably don't need to know much about the Unicode code points that are represented as two consecutive 16 bits chars.
The 8 bit chars are called "Ansi Chars". An PAnsiChar is a pointer to 8 bit chars.
The 16 bit chars are called "Wide Chars". An PWideChar is a pointer to 16 bit chars.
Delphi knows the difference and does well if it doesn't allow you to mix the two!
More info
Here's a popular link on Unicode: The Absolute Minimum Every Software Developer Absolutely, Positively Must Know About Unicode and Character Sets
You can find some more information on migrating Delphi to Unicode here: New White Paper: Delphi Unicode Migration for Mere Mortals
You may also search SO for "Delphi Unicode migration".
A couple years ago, the default character type in Delphi was changed from AnsiChar (single-byte variable representing an ANSI character) to WideChar (two-byte variable representing a UTF16 character.) The char type is now an alias to WideChar instead of AnsiChar, the string type is now an alias to UnicodeString (a UTF-16 Unicode version of Delphi's traditional string type) instead of AnsiString, and the PChar type is now an alias to PWideChar instead of PAnsiChar.
The compiler can take care of a lot of the conversions itself, but there are a few issues:
If you're using string-pointer types, such as PChar, you need to make sure your pointer is pointing to the right type of data, and the compiler can't always verify this.
If you're passing strings to var parameters, the variable type needs to be exactly the same. This can be more complicated now that you've got two string types to deal with.
If you're using string as a convenient byte-array buffer for holding arbitrary data instead of a variable that holds text, that won't work as a UnicodeString. Make sure those are declared as RawByteString as a workaround.
Anyplace you're dealing with string byte lengths, for example when reading or writing to/from a TStream, make sure your code isn't assuming that a char is one byte long.
Take a look at Delphi Unicode Migration for Mere Mortals for some more tricks and advice on how to get this to work. It's not as hard as it sounds, but it's not trivial either. Good luck!
I use delphi 7.
I need to read a utf-8 file line by line, each line contain a word and its weight (a number)
So I need to read every next line, then divide a line by a separator (tab char) and save this in memory.
So,
1) is there a library to work with utf-8 files in Delphi (3-rd party maybe)
2) will functions operate ok with widestring? I use PosEx. So, if they won't, can you also give a link to 3-rd party library to work with widestrings?
If it is really UTF-8 that you are dealing with, then you should not need anything special as far as reading and processing them. You should be able to treat them as pchar or even as a normal Delphi 7 string. If you try to show the contents in some kind of message box, then you may need to do some conversions. For example, I don't believe the Delphi 7 message box method would display UTF-8 strings correctly if the string contained any byte values over 127 (0x7f). For something like that, you would need to convert to UTF-16 and call the Windows API MessageBoxW or something similar. Otherwise, though, UTF-8 strings can be treated in many situations the same as single byte ANSI strings.
I don't think UTF-8 is typically referred to as "widestring". I might be wrong, but I think that typically means UTF-16.
If your file is encoded as UTF-8, and the characters you're looking for are ASCII, then there's no need to use WideString at all. ASCII is a subset of UTF-8, and any ASCII character is guaranteed not to interfere with the special encoding used for other characters in UTF-8. The number characters 0 through 9 and the tab character are all ASCII.
The JCL comes with various functions and classes for dealing with Unicode, if you find you really need to use them.
If most of your input is UTF-8, it might be worthwhile to change your codepage on startup from the "default" to utf8 (codepage 65001). This will make all ansistring->widestring conversions effectively become a lossless utf-8->utf-16.
With D7, you will need a set of so called "unicode" components, components that base themselves on the winapi -W functions. Delphi's own components only do this with the watershed D2009 release that switches the default string type to UTF-16.
If you want to heavily invest in Unicode support, upgrading might be a smart thing to do
WideString is an UTF-16 implementation (a COM BSTR compatible one), it can't store UTF-8 strings, if you assign an 8 bit string it will be converted to UTF-16. But unless you use explicitly the proper conversion function, Delphi will interpret the 8 bit string using the current codepage.
An UTF-8 string can be stored in a Delphi AnsiString (the default string type in Delphi 7), but string manipulation functions are designed for ANSI codepages, not UTF-8. The difference is that UTF-8 is a multi byte character set. But the first 127 ANSI characters, more than one byte is needed to encode a given "character", while many ANSI codepages (especially those for European languages) only require one byte, encoding only 255 "characters" (while UTF-8 can encode the whole Unicode set).
If you're just looking for the tab character AFAIK you could use simply an AnsiString, but you have to ensure that any byte above $80 you may need to look for is not part of a multibyte sequence. If you have more complex processing needs, it may be easier to find libraries working on UTF-16 strings than UTF-8. As Rob Kennedy said, JCL is a good starting point as a free library implementing UTF string manipulation.
You could simply read the file as-is into a normal TStringList via its LoadFrom...() methods, then loop through the list as needed. If loading the entire file into memory at one time is not an option, then you can open the file using a TFileStream and then use the TStreamReader.ReadLine() method to read the stream line-by-line.
If you need to decode a given UTF-8 sequence to UTF-16 for processing, then I would suggest using the Win32 API MultiByteToWideChar() function directly, only because the RTL's UTF8Decode() function has a broken UTF-8 implementation in older Delphi versions (not sure about D7, but it definately does in D6).
The nice thing about either loading approach is that they are both encoding-aware in D2009 and later, which means that if you ever upgrade, you can make a couple of very small code changes to tell the RTL that the data is UTF-8, and it will decode it to UTF-16 for you automatically, and then the rest of your processing code can remain the same (assuming you are not doing anything that is Ansi-specific).
I'm porting an isapi (pageproducers) application from delphi 7 to delphi 2009, the pages are based on html files in UTF8.
Everything goes well except when Onhtmltag is fired and I replace a transparent tag with any value with special characters like accented characters (áé...) Those characters are replaced in the output with an � character.
What's wrong?
As part of your debugging procedure, you should go find out exactly what byte value(s) the browser receives for the question-mark character.
As you should know, Delphi 2009's string type is Unicode, whereas all previous version were ANSI. Delphi 7 introduced the Utf8String type, but Delphi 2009 made that type special. If you're not using that type for holding strings that are encoded as UTF-8, then you should start doing so. Values held in Utf8String variables will be converted to UnicodeString values automatically when you assign one to the other.
If you're storing your UTF-8-encoded strings in ordinary AnsiString variables, then they will be converted to Unicode using the default system code page if you assign them to a UnicodeString. That's not what you want.
If you're assigning UTF-8-encoded literals to variables of type string, stop that. That type expects its values to be encoded as UTF-16, just like WideString always has.
If you are loading your files into a TStrings descendant with LoadFromFile, then you need to start using that method's second parameter, which tells it what encoding to use. UTF-8-encoded files should use TEncoding.UTF8. The default is TEncoding.Unicode, which is little-endian UTF-16.
This is probably a character encoding issue.
The Delphi IDE usually uses Windows-1252 or UTF-16 to encode source code.
HTML often uses UTF-8.
You probably need some transliteration between those encodings.
For that you need to find out what encodings are used exactly (like Rob mentions).
Or revert to HTML escaping accented characters (like Ralph mentions)
Can you post a small app that shows the problem? (you can email me, about anything that has jeroen in the username and pluimers.com in the domain name will arrive in my mailbox).
--jeroen
Thank you for your help, after some test the problem was very very simple (or stupid also)
response.contenttype := 'text/html charset=UTF-8'
No need to translate manually between unicodestring utf8string ansistring widestring. Delphi 2009 string usage is near to perfect.