I noticed that some of the ewftools tools have code to check for the maximum number of open file descriptors.
These checks are within #if defined( HAVE_GETRLIMIT ) blocks.
There is however currently no autoconf code which actually checks for getrlimit() and sets HAVE_GETRLIMIT.
Also, the code which includes sys/resource.h, which is needed for getrlimit(), is contained within a #if defined( HAVE_SYS_RESOURCE_H ) block, and there is no autoconf code which checks for the presence of sys/resource.h and sets HAVE_SYS_RESOURCE_H.
Looking through the git history, I see that at one point there were such checks in configure.ac, so I can imagine that this was disabled on purpose for some reason. On the other hand, they might just have fallen between the cracks during a past refactoring.
I have not followed the code all the way through to see how the results of these getrlimit() checks, which are passed through multiple abstraction layers to be stored for later use, are ultimately used.
These checks stood out to me though because I have encountered the situation in the past that a collection of EWF files could not be mounted because the maximum number of open file descriptors was too small, without a helpful error message.
I noticed that some of the ewftools tools have code to check for the maximum number of open file descriptors.
These checks are within
#if defined( HAVE_GETRLIMIT )
blocks.There is however currently no autoconf code which actually checks for
getrlimit()
and setsHAVE_GETRLIMIT
.Also, the code which includes
sys/resource.h
, which is needed forgetrlimit()
, is contained within a#if defined( HAVE_SYS_RESOURCE_H )
block, and there is no autoconf code which checks for the presence ofsys/resource.h
and setsHAVE_SYS_RESOURCE_H
.Looking through the git history, I see that at one point there were such checks in
configure.ac
, so I can imagine that this was disabled on purpose for some reason. On the other hand, they might just have fallen between the cracks during a past refactoring.I have not followed the code all the way through to see how the results of these
getrlimit()
checks, which are passed through multiple abstraction layers to be stored for later use, are ultimately used.These checks stood out to me though because I have encountered the situation in the past that a collection of EWF files could not be mounted because the maximum number of open file descriptors was too small, without a helpful error message.