Amebis/stdex
1
0
Fork 0
Code Issues Pull Requests Actions Packages Projects Releases Wiki Activity

Stop throwing in destructors

Browse Source 
When exception processing is unwinding the stack, any exception thrown
in destructors end up std::terminate()-in our process.

Since all workarounds to report errors from destructors seems an
overkill, and the only important place where we should notify the user
about a failure is the stdex::stream::cache::~cache() (data loss occurs
when failure happens here), and if we'd throw in stdex::stream::cache::
~cache(), our process would get terminated anyway so the data loss is
inevitable, let's just silence this for now and come up with a better
solution later if we get smarter anytime in the future.

Signed-off-by: Simon Rozman <simon@rozman.si>
This commit is contained in:
Simon Rozman 2025-03-06 22:09:50 +01:00
parent 69f2c639cd
commit bfd8aaff65
3 changed files with 33 additions and 25 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

7
include/stdex/stream.hpp
View File

@ -86,7 +86,7 @@ namespace stdex
public:
basic(_In_ state_t state = state_t::ok) : m_state(state) {}
virtual ~basic() noexcept(false) {}
virtual ~basic() {}
///
/// Reads block of data from the stream
@ -1852,9 +1852,10 @@ namespace stdex
#endif
{}
virtual ~cache() noexcept(false)
virtual ~cache()
{
if (m_source) {
try {
flush_cache();
if (!ok()) _Unlikely_
throw std::system_error(sys_error(), std::system_category(), "failed to flush cache"); // Data loss occurred
@ -1864,6 +1865,8 @@ namespace stdex
m_source->set_mtime(m_mtime);
#endif
}
catch (...) {} // TODO: Never throw in destructors. If we'd throw here exception stack unwinding would std::terminate() our process anyway. Is there a way to catch this?
}
}
virtual _Success_(return != 0 || length == 0) size_t read(

8
include/stdex/system.hpp
View File

@ -171,10 +171,12 @@ namespace stdex
return *this;
}
virtual ~basic_sys_object() noexcept(false)
virtual ~basic_sys_object()
{
if (m_h != TR::invalid_handle)
TR::close(m_h);
if (m_h != TR::invalid_handle) {
try { TR::close(m_h); }
catch (...) {} // Failure to close a handle should not be that devastating as throwing in a destructor may be.
}
}
///

3
include/stdex/zlib.hpp
View File

@ -61,6 +61,7 @@ namespace stdex
virtual ~zlib_writer()
{
try {
m_zlib.avail_in = 0;
m_zlib.next_in = NULL;
do {
@ -76,6 +77,8 @@ namespace stdex
// deflateReset(&m_zlib);
deflateEnd(&m_zlib);
}
catch (...) {} // TODO: Never throw in destructors. If we'd throw here exception stack unwinding would std::terminate() our process anyway. Is there a way to catch this?
}
virtual _Success_(return != 0) size_t write(
_In_reads_bytes_opt_(length) const void* data, _In_ size_t length)