Dealing with library version mismatch

 Note : this article was initially redacted as an answer to David Jud's comment, but it became long enough to be worth converting into a full blog entry.
In previous article, I attempted to introduce a few challenges related to designing an extensible API.

In this one, I'll cover an associated but more specific topic, on how to handle a library version mismatch.

Version mismatch is a more acute problem in a DLL scenario. In a static linking scenario, the programmer has several advantages :

• Compiler will catch errors (if a type, or a prototype, has changed for example). This gives time to fix these errors. Of course, the application maintainer will prefer that a library update doesn't introduce any change in existing code, but worst case is, most errors should be trapped before shipping the product.
• Compiler will automatically adapt ABI changes : if an updated type is larger/smaller than previous version, it will be automatically converted throughout the code base. Same thing happens in case of enum changes : adaptation to new enum values is automatically applied by compiler.
• Library is available during compilation, which means programmer has a local copy that he can update (or not) according to its requirements.

Well, this last property is not always true : in larger organisations, library might belong to a "validated" pool, which cannot be easily adapted for a specific project. In which case, the user program will either have to host its own local copy, or adapt to the one selected by its organisation.

But you get the idea : problematic version mismatches are likely to be trapped or automatically fixed by the compiler, and therefore should be corrected before shipping a binary. Of course, the less changes, the better. Program maintainers will appreciate a library update as transparent as possible.

For a dynamic library though, the topic is a lot harder.

To begin with, user program typically does not have direct control over the library version deployed on target system. So it could be anything. The library could be more recent, or older, than expected during program development.

Now these two types of mismatches are quite different, and trigger different solutions :

Case 1 - library version is higher than expected

This one can, and should, be solved by the library itself.

It's relatively "easy" : never stop supporting older entry points.
This is of course easier said than done, but to be fair, it's first and foremost a question of respecting a policy, and therefore is not out of reach.
Zstandard tries to achieve that by guaranteeing that any prototype reaching "stable" status will be there "forever". For example, ZSTD_getDecompressedSize(), which has been recently superceded by ZSTD_getFrameContentSize(), will nonetheless remain an accessible entry point in future releases, because it's labelled "stable".

A more subtle applicable problem is ABI preservation, in particular structure size and alignment.
Suppose, for example, that version v1.1 defines a structure of size 40 bytes.
But v1.2 add some new capabilities, and now structure has a size of 64 bytes.
All previous fields from v1.1 are still there, at their expected place, but there are now more fields.

The user program, expecting v1.1, would allocate the 40-bytes version, and pass that as an argument to a function expecting a 64-bytes object. You can guess what will follow.

This could be "manually" worked around by inserting a "version" field and dynamically interpreting the object with the appropriate parser. But such manipulation is a recipe for complexity and errors.
That's why structures are pretty dangerous. For best safety, structure definition must remain identical "forever", like the approved "stable" prototypes.

In order to avoid such issue, it's recommended to use incomplete types. This will force the creation of underlying structure through a process entirely controlled by current library, whatever its version, thus avoiding any kind of size/alignment mismatch.

When above conditions are correctly met, the library is "safe" to use by applications expecting an older version : all entry points are still there, behaving as expected.

Whenever this condition cannot be respected anymore, an accepted work-around is to increase the Major digit of the version, indicating a breaking change.

Case 2 - library version is lower than expected

This one is more problematic.
Basically, responsibility is now on the application side. It's up to the application to detect the mismatch and act accordingly.

In David Jud's comment, he describes a pretty simple solution : if the library is not at the expected version, the application just stops there.
Indeed, that's one way to safely handle the matter.

But it's not always desirable. An application can have multiple library dependencies, and not all of them might be critical.
For example, maybe the user program access several libraries offering similar services (encryption for example). If one of them is not at the expected version, and cannot be made to work, it's not always a good reason to terminate the program : maybe there are already plenty of capabilities available without this specific library, and the program can run, just with less options.

Even inside a critical library dependency, some new functionality might be optional, or there might be several ways to get one job done.
Dealing with this case requires writing some "version dependent" code.
This is not an uncommon situation by the way. Gracefully handling potential version mismatches is one thing highly deployed programs tend to do well.

Here is how it can be made to work : presuming the user application wants access to a prototype which is only available in version v1.5+, it first tests the version number. If condition matches, then program can invoke target prototype as expected. If not, a backup scenario is triggered, be it an error, or a different way to get the same job done.

Problem is, this test must be done statically.
For example, in Zstandard, it's possible to ask for library version at runtime, using ZSTD_versionNumber(). But unfortunately, it's already too late.
Any invocation of a new function, such as ZSTD_getFrameContentSize() which only exists since v1.3.0, will trigger an error at link time, even if the invocation itself is protected by a prior check with ZSTD_versionNumber() .

What's required is to selectively remove any reference to such prototype from compilation and linking stages, which means this code cannot exist. It can be excluded through pre-processor.
So the correct method is to use a macro definition, in this case, ZSTD_VERSION_NUMBER . 

Example :

#if ZSTD_VERSION_NUMBER >= 10300

size = ZSTD_getFrameContentSize(src, srcSize);

#else

size = ZSTD_getDecompressedSize(src, srcSize);

/* here, 0-size answer can be mistaken for "error", so add some code to mitigate the risk */

#endif

That works, but requires to compile binary with the correct version of zstd.h header file.
When the program is compiled on target system, it's a reasonable expectation : if libzstd is present, zstd.h is also supposed to be accessible. And it's reasonable to expect them to be synchronised. There can be some corner case scenarios where this does not work, but let's say that in general, it's acceptable.

The detection can also be done through a ./configure script, in order to avoid an #include error during compilation, should the relevant header.h be not even present on target system, as sometimes the library is effectively optional to the program.

But compilation from server side is a different topic. While this is highly perilous to pre-compile a binary using dynamic libraries and then deploy it, this is nonetheless the method selected by many repositories, such as aptitude, in order to save deployment time. In which case, the binary is compiled for "system-provided libraries", which minimum version is known, and repository can solve dependencies. Hence, by construction, the case "library has a lower version than expected" is not supposed to happen. Case closed.

So, as we can see, the situation is solved either by local compilation and clever usage of preprocessing statements, or by dependency solving through repository rules.

Another possibility exists, and is, basically, the one proposed in ZSTD_CCtx_setParameter() API : the parameter to set is selected through an enum value, and if it doesn't exist, because the local library version is too old to support it, the return value signals an error.

Using safely this API feels a lot like the previous example, except that now, it becomes possible to check library version at runtime :

if (ZSTD_versionNumber() >= 10500) {

   return ZSTD_CCtx_setParameter(cctx, ZSTD_p_someNewParam, value);

} else {

   return ERROR(capability_not_present);

}

This time, there is no need to be in sync with the correct header.h version. As the version number comes directly at runtime from the library itself, it's necessarily correct.

Note however that ZSTD_CCtx_setParameter() only covers the topic of "new parameters". It cannot cover the topic of "new prototypes", which still requires using exclusion through macro detection.

So, which approach is best ?

Well, that's the good question to ask. There's a reason the new advanced API is currently in "experimental" mode : it needs to be field tested, to experience its strengths and weaknesses. There are pros and cons to both methods.
And now, the matter is to select the better one...

The art of designing advanced API

 A library API (Application Programming Interface) is even more important than its implementation.

There are many reasons for this statement :
- An API exposes a suitable abstraction. Should it prove broken, unclear or just too complex, the library will be misused, which will ultimately be paid by users' time (multiplied by nb of users).
- An API is a contract. Break it, and existing applications can no longer work with your library. Adaptation cost is once again paid by users' time (if it ever happens).
- Because of this, API tend to stick around for a long time, much longer than underlying implementation.

If an implementation is modified to provide, say, a 5% speed improvement, it's all free, every user can immediately benefit about it without further hassle. But if one has to add a single parameter, it's havoc time.

Because it's so stressful to modify an API, one can be tempted to look very hard once, in order to design and expose a perfect API, one that will stand the test of time, and will never need to be changed. This search is (mostly) a delusion.
- perfect API, embedding such a strong restriction to never change in the future, can take forever to build, all under intense stress, as there is always a question mark hanging around : "is there a use case that it does not cover ?". Eventually, it's only a matter of time before you (or your users) find one.
- perfect API lean towards "complex API", as the requirement to support everything makes it add more parameters and control, becoming more and more difficult to manage by users.
- "complex" quickly leads to "misleading", as supporting some "future scenarios" for which there is no current implementation, and maybe no current need, will be categorised bad ideas after all, but side-effects of this useless abstraction will remain in the API.

So, the next great idea is to plan for API changes.
The way Zstandard library tries to achieve this is by quickly converging towards some very simple prototypes, which offer "core" functionalities at a low complexity level.
Then, more complex use cases, not covered by simple API, do show up, and the need to serve them introduce the creation of an "experimental section", a place where it's still possible to play with API, trying to find an optimal abstraction for intended use case, before moving into "stable" status (aka: this method will no longer change).

A consequence of this strategy is the creation of more and more prototypes, dedicated to serving their own use case.
Need to compress with dictionary ? Sure, here comes ZSTD_compress_usingDict() .
Need to process data in a streaming fashion ? Please find  ZSTD_compressStream() .
In a streaming fashion with a dictionary ? ZSTD_compressStream_usingDict() .
Need control over specific parameters ? Go to _advanced() variants.
Preprocess dictionary for faster loading time ? Here are _usingCDict() variants.
Some multithreading maybe ? ZSTDMT_*()

Combine all this ? Of course. Here is a list of a gazillion methods.

As one can see, this doesn't scale too well. It used to be "good enough" for a dozen or so methods, but as combinatorial complexity explodes, it's no longer suitable.

In latest release of Zstandard, we try to get a fresh look to this situation, and provide an API simpler to manage. The result of which is the new advanced API candidate, which actually stands a chance to become stable one day.

It features 2 core components : 

ZSTD_compress_generic() is the new main compression method. It's designed to support all other compression methods. It can do block, streaming, dictionary, multithreading, and any combination of those. We have plan for even more extensions, and they all seem to fit in.

This is possible because now sending parameters is a separate operation, which can be completed in as many steps as necessary.

The main vehicle to set these parameters is ZSTD_CCtx_setParameter() .

It uses an enum based policy, where the parameter is selected in an enum list, and new value is provided as an unsigned type.

This design has been favoured over previous one, which was using a struct to pass all parameters in a single step. The struct was inconvenient as it forced user to select a value for each and every parameter, even out-of-scope ones, in order to change just one of them. Perhaps even more importantly, the struct is problematic for future changes : adding any new parameter would change the struct size, which is an ABI break. It can quickly get ugly when the program and library work on common memory areas using different sizes.

The enum policy allow us to add any new parameter while preserving API and ABI, so it looks very flexible.

However, it comes with its own set of troubles.

To begin with, enum values can very easily change : just add a new enum in the list, and see all enum values after that one slide by one.
It can be a problem if, in a version of the library, ZSTD_p_compressionLevel is attributed a 2, but in a future version, it becomes a 3. In a dynamic library scenario, where version mismatch can easily happen, it means the caller is changing some other random parameter.

To counter that, it will be necessary to pin down all enum values to a manually assigned one. This will guarantee the attribution.

Another issue is that the value of the parameter is provided as an unsigned type, so the parameter must fit this type. That's not always possible.

For example, there is a dedicated method to set pledgedSrcSize, which is essentially a promise about how much data is going to be processed. This amount can be very large, so an unsigned type is not enough. Instead, we need an unsigned long long, hence a dedicated method.

Another even more obvious one happens when referencing a prepared dictionary in read-only mode : this parameter is a const ZSTD_CDict* type, so it is  set through a dedicated method, ZSTD_CCtx_refCDict().

And we have a few other exceptions using their own method, as the argument cannot fit into an unsigned.

But the large majority of them uses ZSTD_CCtx_setParameter().

In some cases, the adaptation works though it's not "best".

For example, a few parameters are selected among a list of enums, for example ZSTD_strategy . The enum is simply casted to an unsigned and passed as argument. It works. But it would have been even nicer to keep the argument type as the intended enum, giving the compiler a chance to catch potential type mismatch (example).

So this design could be in competition with another one : define one method per parameter. The most important benefit would be that each parameter can have its own type.

But this alternate design has also its own flaws :

adding any new parameter means adding a method. Therefore, if a program uses a "recent" method, but links against an older library version, this is a link error.
In contrast, the enum policy would just generate an error in the return code, which can be identified and gracefully dealt with.

Creating future-proof API is hard. There is always a new unexpected use case which shows up and would require another entry point or another parameter. The best we can do is plan for those changes.

The new Zstandard's advanced API tries to do that. But since it is a first attempt, it likely is perfectible. 

This is design time, and it will cost a few revisions before reaching "stable" status. As always, user feedback will be key to decide if the new design fits the bill, and how to amend it.

Follow up : Dealing with library version mismatch

Edit :
Arseny Kapoulkine made an interesting comment, arguing that specialized entry points make it possible for compiler's DCE (Dead Code Elimination) optimisation to kick in, removing useless code from the final binary :
https://twitter.com/zeuxcg/status/882816066296172550

In general this is true. Calling
specialized_function1(...)
is clear for the linker,
then it's possible to remove any potentially unused specialized_function2() from binary generation.

In contrast calling
generic_function(mode=1, ...)
with
void generic_function(int mode, ...)
{
   switch(mode) {
      case 1 : return specialized_function1(...);
      case 2 : return specialized_function2(...);
}

makes it much more difficult. In general, for the second case, specialized_function2() will remain in the binary.
(an exception being usage of inline, associated with -flto, and non-ambiguous selection parameter, but that's no longer a "simple" setup).

For Zstandard though, it doesn't make a lot difference.
The reason is, whatever "specialized" entry point is invoked, (ZSTD_compress(), or ZSTD_compress_usingDict() for example), it's just an entry point. The compression code is not "immediately behind", it's reached after several indirection levels. This design make it possible for a single compression code to address multiple usage scenarios with vastly different set of parameters, which is vital for maintenance. But disentagling that for DCE is a lot more difficult.
If required, -flto makes it possible to optimize size better, and some difference become visible, but remain relatively small.