largefile problems

Through actual problems with handling files larger than 2 Gigabyte, I had to face the fact that there are serious problems with the largefile implementation around shared-libraries. This is bound to the effect that 64on32 platforms allow a preprocessor #define _FILE_OFFSET_BITS=64 which will shift the integretal type of off_t from a 32bit to a 64bit entity.

That will in fact lead to problems if an application is compiled with a different off_t size than the shared-library it will be linked to. Among the problems are different sizes of the callframe for those functions that take an argument of off_t type. Note how the "seek" call uses an off_t in the middle of its arguments which may be 32bit or 64bit depending on a preprocessor define. And you know that zziplib wraps up "seek"-like calls as well.

The observations were largely made independent of the zziplib however, as a zip-file does uses 32bit offsets in its header fields anyway, and therefore a single zip-archive (or any of its wrapped files) can not be larger than 2 Gigabyte anyway. I would to refer you for deeper information to my website about this problem-space at

zziplib related

Still however the problems hold for zziplib usage when the functions are linked dynamically as a shared-library. Here we can face the fact that an application (or a higher level library) uses a different off_t size than the underlying zziplib shared-library.

If you read the zzip/zzip.h header file then you will surely see a number of off_t usages around, here they are wrapped in the form of zzip_off_t to get away with platforms not predefining off_t in the first place. Those functions are (at the time of writing):

• zzip_telldir (return type)
• zzip_seekdir (second param)
• zzip_tell (return type)
• zzip_seek (second param of three)

What might be not as obvious however: you will find also the off_t type being used in the plugin-handlers callback functions. That is based on the fact that the plugin-structures is filled by default with the posix-functions from the C library. A 64on32 platform however offers quite usually a mixedmode C library exporting two symbols for tell/seek calls to match either 32bit off_t or 64bit off_t

• zzip_plugin_io->seeks (return type and second param)
• zzip_plugin_io->filesize (return type)

The problem here: the application might not make use of zzip_seek/zzip_tell explicitly, but may be the internal implementation of a zzip call uses io->seeks or io->filesize. When an application uses plugin-io with these callbacks overridden then surely problems will arise.

zziplib mixedmode option

I have extended the zziplib implementation to allow itself to live fine on 64on32 systems. The 64on32 system are like linux and solaris where the default off_t is 32bit and only by the preprocessor hint they shift into 64bit. The C library on these systems is a mixedmode one offering a pair for each of the problematic functions - lseek and lseek64 for example.

The zziplib header file detects when it is present on a 64on32 system (through hints in configured zzip/conf.h) and that _FILE_OFFSET_BITS has been set to 64bit. In that case it does automatically issue #defines that shift the symbol-name from zzip_seek into zzip_seek64. Likewise, all the *_ext_io functions are renamed into *_ext_io64 calls

The zziplib library itself will also pick up the renamings when it is compiled with 64bit off_t - in effect an application with a 64bit-off_t dependency can only link with a zziplib compiled in 64bit-off_t mode. If the application does not use any call symbol with an off_t dependency then it does not matter and the link will succeed. That's simply because function calls without an off_t dependency will not be renamed and they are the same for a 32bit-off_t zziplib or a 64bit-off_t zziplib.

As an extra, the zziplib exports a few of its common calls like being a mixedmode library when you compile it both in 64bit mode and as a shared library. In that case, the resulting shared library will export symbol pairs for the calls with an off_t dependency, i.e. both zzip_seek and zzip_seek64 are present.

Note that for reasons of being a lightweight library, the zziplib library does not export mixmode call pairs for the *_ext_io family of functions. The current generation of zziplib does call io->seeks unconditionally of any case'ing flag and so far there are no problems with the current design.

Implementation details

In the header file zzip/zzip.h you will find the define ZZIP_LARGEFILE_RENAME which triggers the renaming process. See zzip/conf.h on the conditions where it is being triggered.

For the implementation of the mixedmode symbol pairs, see zzip/dir.c for an example for the zzip_seekdir/zzip_seekdir64 pair - here we use libtools -DPIC to detect the situation of being compiled as shared-library, we use the preprocessor #def ZZIP_LARGEFILE_RENAME to know we are on a 64on32 system compiled in 64bit-off_t, and we check the transitional largefile API to be present by looking for EOVERFLOW errno.

When all the three are present then we simply #undef the renaming preprocessor macro and define a function symbol (without the renaming) and call the renamed symbol already compiled a few lines before. We use the pre-off_t type "long" for the 32bit entity of these calls. While we mostly let the compiler do the shrink/expand of these integer types, we do also sometimes check for overflows of the seekvalue.

rpm extras and pkg-config

The provided .spec file shows how to compile both variants of the zziplib shared library and to install them in parallel in the system. Also we provide doubled sets of .pc files for pkg-config installation. That should make it a lot easier for applications to link to the correct library they want.

Here are all the variants that you can find after installing the vanilla rpm files from zziplib.sf.net:

$ pkg-config --list-all | sort | grep zzip
zziplib32                   zziplib32 - ZZipLib - libZ-based ZIP-access Library
zziplib64                   zziplib64 - ZZipLib - libZ-based ZIP-access Library
zziplib                     zziplib - ZZipLib - libZ-based ZIP-access Library
zzip-sdl-config             zzip-sdl-config - SDL Config (for ZZipLib)
zzip-sdl-rwops              zzip-sdl-rwops - SDL_rwops for ZZipLib
zzipwrap                    zzipwrap - Callback Wrappers for ZZipLib
zzip-zlib-config            zzip-zlib-config - ZLib Config (for ZZipLib)