Activity for Stephane
-
Stephane posted a comment on discussion Open Discussion
Nobody want to dictate anything to you and certainly not me neither Apple. You made a mistake by naming types like you did, Apple will not change his naming that's for sure, because its at compiler level for them not at applicaion level like lzma, this conflict can be easily be fixed in less than 1 second, you constantly give false excuses not to do it, and you are searching an illusory quick fix, its your choice, good luck with that. Concerning LzmaEnc_Finish i already explain to you many times...
-
Tino Reichardt , If you define your own type in an interface, it is for inherit them under certain circumstance. It is no more an interface. Its a device ready to explode soon or later. Rename these type takes less than one second. Igor wanted to use typedef to define his own type, his has to accept the consequence of his mistake (bad naming).
-
If you define your own type in an interface, it is for inherit them under certain circumstance. It is no more an interface. Its a device ready to explode soon or later. Rename these type takes less than one second. Igor wanted to use typedef to define his own type, his has to accept the consequence of his mistake (bad naming).
-
.
-
When i asked to you why you cannot rename the types you said "Its everywhere" So it took me 3 hours to write a c program to rename these types , and it works and it takes less than 1 second to execute it. In less than 1 second the problem is solved, (68 files modified). After that you said "Its a bad solution" P_uint32 i use as an example "is too large". So i re run the programm with "uint32" (It can be anything) the same length and it works. After that you said "UInt32 is comfortable for me." "what...
-
When i asked to you why you cannot rename the types you said "Its everywhere" So it took me 3 hours to write a c program to rename these types , and it works and it takes less than 1 second to execute it. In less than 1 second the problem is solved. After that you said "Its a bad solution" P_uint32 i use as an example "is too large". So i re run the programm with "uint32" (It can be anything) the same length and it works. After that you said "UInt32 is comfortable for me." "what if ..." etc...."...
-
If you define your own type in an interface, it is for inherit them under certain circumstance. It is no more an interface. Its a device ready to explode soon or later. Rename these type takes less than one second. Igor wanted to use typedef to define his own type, his has to accept the consequence of his mistake (bad naming).
-
Swift can be bridged with objective C or C. You create a bridge file who is a h file where you put your #include h files of the c files. The Swift compiler compile the swift file, clang compiler compile the c program and link the whole thing into one binary. For lzma i put the h file of the c wrapper in the bridge file and that wrapper call your interface with lzmalib.h declare in it, the Swift pgm call the wrapper. Like that your interface does not inherit the Swift definition avoiding conflict....
-
Swit can be bridged with objective C or C. You create a bridge file who is a h file where you put your #include h files of the c files. clang compile the swift and the c program and link that into one binary. For lzma i put the h file of the c wrapper who call your interface with lzmalib.h declare in it. Swift pgm call the wrapper. Like that your interface does not inherit the Swift definition avoiding conflict. On the other interface i use (cryptography or jpeg ...) i put the include of the h file...
-
I use the standard interface Lzmalib.h who call LzmaEnc NOT LzmaEnc2.c where LzmaEnc_Finish is called again, so i don't compile LzmaEnc2.c and using 7ZIPST i don't need LzmaEncFinish.
-
Again you play with fire. Apple has his definition and you have yours. The same name does not means the same definition. This is an interface , and in no way a dependance must be created between 2 languages, you don't know what apple will do next, its too dangerous. This is the first time i saw UIntxx in any interface i use, see stdint.h. Seems like a modern revamping is needed. And concerning the name you use you have the solution if you want to rename your definition.
-
You play with fire, there no guarantee that Swift (internal definition) use the same definition as you. You create a dependance (same definition) between you and Swift. Remember the boolean problem. One way to resolve this is to create a c wrapper called from Swift (which i used ) who call you interface Lzmalib.h LzmaCompress, like that the 2 definitions are not in conflict. (an interface is created). You can define UInt32 as unsigned long , there will be no problem with Swift, Swift will use UInt32...
-
We need some #ifdef solution for 7zTypes.h. In the file LzFind.h you have : typedef struct _CMatchFinder { Byte *buffer; UInt32 pos; UInt32 posLimit; .. In the file LzFind.c you have : define kMaxValForNormalize ((UInt32)0xFFFFFFFF) How do you solve that with #ifdef in 7zTypes.h ?
-
You can use whatever name you want, "P_uint32" is an example. You can only lowercase your type int32, uint32 int32 is not Int32. Int16 -> int16 Int32 -> int32 Int64-> int64 UInt16 -> uint16 UInt32 -> uint32 UInt64-> uint64 Like that there is no conflict with Swift, i tested it , it works.
-
Laziness... It took me 3 hours to write a c pgm described below who changes the types name, there is 68 files modified.
-
I wrote a c pgm who resolve the conflict between Swift and Mr Pavlov definition. The c progam has a directory as argument, the "C" directory in our case. It change all c and h files recursively (Utils/7z etc). Execution time is less than 1 second. The definitions hardcoded in the pgm are : char oldTypes[] = "UInt16,UInt32,UInt64,Int16,Int32,Int64", newTypes[] = "P_uint16,P_uint32,P_uint64,P_int16,P_int32,P_int64"; Compile like this : cc changeType.c -Ofast -o changeType Save the directory first....
-
Why compile and execute an empty function ? What's the purpose ? LzmaEnc_Finish is not use with _7ZIP_ST.
-
It does not work that way. If you don't want to rename your type definition, i can do nothing for you and you will stay in conflict with Swift. There is no changes in your software with the solution i use. Concerning changes in your software there one i posted 4 years ago : In LzmaEnc.c : ifndef _7ZIP_ST void LzmaEnc_Finish(CLzmaEncHandle pp) { .. .. } endif and ifndef _7ZIP_ST LzmaEnc_Finish(p); endif This is a modification i rewrite at every release.
-
You put yourself in a very bad situation and you don't accept the consequence. I solved this problem months ago. This topic is to inform you that your are in conflict with Swift, that's all.
-
Why you can't ?
-
After renaming Bool in 7zTypes.h to BoolInt because swift used Bool as type and program cannot compile, there are other definitions Swift already used : Int16, UInt16, Int32, UInt32, Int64, UInt64 So once again can you rename these definitions like you did with Bool -> BoolInt to avoid confusion with Swift. The definition (right click on Int64 for example points to 7zTypes.h and NOT the swift definition) Here the Swift definition : https://developer.apple.com/documentation/swift/swift_standard_library/numbers_and_basic_values...
-
Hello, Here : https://superuser.com/questions/852141/john-the-ripper-crack-zipcrypto-password This will help you. It works with AES 7zip encrypted files. Dont forget to edit john.conf (Min and max length, alpha, num etc...) to reduce the process.
-
Finally ... 7 months later..... this boolean bug is solved. So I no longer need the wrapper and can now use the lzma apis directly with Swift. the second problem raised three years ago is still not solved.
-
1 - clang 2 - Its a problem i posted 3 years ago. https://sourceforge.net/p/sevenzip/discussion/45797/thread/9e68012c/?limit=25&page=2#0ef0/0697/a6a9/960f/8fa7/d4c7/a013 I don't use Lzma2Enc.c but LzmaEnc.c from LzmaLib.c, the LZMA library interface as you call it. At every release i have to fix it. Another thing. i cannot use LzmaLib.c functions from Swift because you define "typedef int Bool" in 7zTypes.h and Swift already use Bool but not as int. So i have to call a C pgm ( a wrapper) who call...
-
LzmaEnc.c:1467:69: warning: for loop has empty body [-Wempty-body] for (len = 3; len < limit && data[len] == data2[len]; len++); i have to put -Wno-empty-body to silence this warning. And i add this condition : ifndef _7ZIP_ST LzmaEnc_Finish(p); endif and ifndef _7ZIP_ST void LzmaEnc_Finish(CLzmaEncHandle pp) { CLzmaEnc p = (CLzmaEnc )pp; if (p->mtMode) MatchFinderMt_ReleaseStream(&p->matchFinderMt); } endif
-
Its better to not calll a function instead of calling it and do nothing as its done...
-
Yes as i told the warning is in LzmaEnc.c and LzmaLib.c and i just propose a solution...
-
Just an idea : In LzmaEnc.c Before : void LzmaEnc_Finish(CLzmaEncHandle pp) { #ifndef...
-
clang
-
Apple LLVM version 6.0 (clang-600.0.56) (based on LLVM 3.5svn)
-
Hello, In LzmaEnc.c function name LzmaEnc_Finish, there is a warning for p = p; "explicitly...
1