# Another ABI breakage

**URL:** <https://forum.hdfgroup.org/t/another-abi-breakage/5503>\
**Category:** HDF5 Library\
**Created:** [March 6, 2019, 2:19pm UTC](https://forum.hdfgroup.org/t/another-abi-breakage/5503 "2019-03-06T14:19:29Z")\
**Posts on this page:** 2\
**Page:** 1

<div class="post-metadata">

**Author:** ![paramon](https://forum.hdfgroup.org/letter_avatar_proxy/v4/letter/p/8491ac/32.png) [@paramon](https://forum.hdfgroup.org/u/paramon)\
**Post date:** [March 6, 2019, 2:19pm UTC](https://forum.hdfgroup.org/t/another-abi-breakage/5503/1 "2019-03-06T14:19:29Z")

</div>

Hello HDF5ers,  
Some bad news for you today! ☹

In HDF5 1.10.1, H5Fpublic.h:

```auto
/* Library's file format versions */
typedef enum H5F_libver_t {
    H5F_LIBVER_EARLIEST, /* Use the earliest possible format for storing objects */
    H5F_LIBVER_LATEST /* Use the latest possible format available for storing objects*/
} H5F_libver_t;

```

`H5F_LIBVER_LATEST` is `1`.

In HDF5 1.10.2, H5Fpublic.h:

```auto
/* Library's format versions */
typedef enum H5F_libver_t {
    H5F_LIBVER_ERROR = -1,
    H5F_LIBVER_EARLIEST = 0, /* Use the earliest possible format for storing objects */
    H5F_LIBVER_V18 = 1, /* Use the latest v18 format for storing objects */
    H5F_LIBVER_V110 = 2, /* Use the latest v10 format for storing objects */
    H5F_LIBVER_NBOUNDS
} H5F_libver_t;

#define H5F_LIBVER_LATEST H5F_LIBVER_V110

```

`H5F_LIBVER_LATEST` is `2`.  
`1` is `H5F_LIBVER_V18`.

A call to [H5Pset\_libver\_bounds](https://support.hdfgroup.org/HDF5/doc/RM/H5P/H5Pset_libver_bounds.htm) with `H5F_LIBVER_EARLIEST, H5F_LIBVER_LATEST` in a program built against HDF5 1.10.1 and executed against HDF5 1.10.5 is effectively understood as `H5F_LIBVER_EARLIEST, H5F_LIBVER_V18`, which will lead to errors (e.g. it’s not possible to copy latest-format objects into such file).

Please:

1. Submit a JIRA item.
2. Consider ways to fix it in a hot-patch release.

Maybe there is some way of elegant work-around?

Best wishes,  
Andrey Paramonov

---

<div class="post-metadata">

**Author:** ![epourmal](https://forum.hdfgroup.org/letter_avatar_proxy/v4/letter/e/c67d28/32.png) [@epourmal](https://forum.hdfgroup.org/u/epourmal)\
**Post date:** [March 7, 2019, 5:34am UTC](https://forum.hdfgroup.org/t/another-abi-breakage/5503/2 "2019-03-07T05:34:31Z")

</div>

Hi Andrey,

Thank you for reporting, but we will not fix this issue.

To be honest, HDF5 1.10.0 and 1.10.1 ARE NOT stable “fully-baked” releases of HDF5 1.10. One of the issues with those releases is exactly the issue you brought up: one couldn’t distinguish between the features that were LATEST in 1.8.\* and the features that are LATEST in 1.10.\*.

The issue was addressed only in HDF5 1.10.2. We discussed it in our [Blog](https://www.hdfgroup.org/2018/04/why-should-i-care-about-the-hdf5-1-10-2-release/) and for more information referred our users to the [design document](https://www.hdfgroup.org/wp-content/uploads/2018/04/RFC-Setting-Bounds-for-Object-Creation-in-HDF5-1.10.0-Update.pdf).

We understand the importance of ABI compatibility and are putting in place a new process to avoid ABI issues mentioned on this FORUM.

Thank you and please continue your excellent reporting!

Elena
