version 1.49 unbuildable on linux

[wdoekes]

Many issues with 1.49. Was this release intentional?

• 1.49 is a lot higher than 1.20.0
• 1.49 does not have a 1.49.0 patch version
• #define LOGX is done twice
• OPAL_DiskInfo is now DTA_DEVICE_INFO
• void DtaDevOS::identify should be bool ...

[RossiSoftLtd]

Just found out the same issue. Wondering the same as @wdoekes

Thanks

[cranderson]

g++ -DHAVE_CONFIG_H -I. -Wall -Werror -std=c++11 -I./linux -I./Common -I./Common/pbkdf2 -I./linux -I./LinuxPBA -O2 -flto=auto -ffat-lto-objects -fexceptions -g -grecord-gcc-switches -pipe -Wall -Werror=format-security -Wp,-U_FORTIFY_SOURCE,-D_FORTIFY_SOURCE=3 -Wp,-D_GLIBCXX_ASSERTIONS -specs=/usr/lib/rpm/redhat/redhat-hardened-cc1 -fstack-protector-strong -specs=/usr/lib/rpm/redhat/redhat-annobin-cc1 -m64 -march=x86-64 -mtune=generic -fasynchronous-unwind-tables -fstack-clash-protection -fcf-protection -mtls-dialect=gnu2 -fno-omit-frame-pointer -mno-omit-leaf-frame-pointer -c -o LinuxPBA/UnlockSEDs.o LinuxPBA/UnlockSEDs.cpp
make[1]: Leaving directory '/builddir/build/BUILD/sedutil-1.49'
In file included from ./linux/os.h:28,
from LinuxPBA/UnlockSEDs.cpp:20:
./Common/log.h:338:9: error: "LOGX" redefined [-Werror]
338 | #define LOGX(level)
| ^~~~
./Common/log.h:330:9: note: this is the location of the previous definition
330 | #define LOGX(level)
| ^~~~
In file included from LinuxPBA/UnlockSEDs.cpp:22:
./linux/DtaDevOS.h:58:19: error: ‘OPAL_DiskInfo’ has not been declared
58 | void identify(OPAL_DiskInfo& disk_info);
| ^~~~~~~~~~~~~
./linux/DtaDevOS.h:60:24: error: conflicting return type specified for ‘virtual long long unsigned int DtaDevOS::getSize()’
60 | unsigned long long getSize();
| ^~~~~~~
In file included from ./linux/DtaDevOS.h:21:
./Common/DtaDev.h:134:36: note: overridden function is ‘virtual const long long unsigned int DtaDev::getSize()’
134 | virtual const unsigned long long getSize() = 0;
| ^~~~~~~
LinuxPBA/UnlockSEDs.cpp: In function ‘uint8_t UnlockSEDs(char*)’:
LinuxPBA/UnlockSEDs.cpp:34:39: error: ‘getDtaDevOS’ is not a member of ‘DtaDevOS’
34 | if (DTAERROR_SUCCESS == DtaDevOS::getDtaDevOS(devref, &d, true) && d!=NULL) {
| ^~~~~~~~~~~
cc1plus: all warnings being treated as errors
make[1]: *** [Makefile:782: LinuxPBA/UnlockSEDs.o] Error 1

[youk]

Hey @scottcmarks would you manage to look into your problematic commit?

[vt-alt]

Can authors please explain meaning of new versioning scheme?

So far, no Linux distribution builds 1.49.x https://repology.org/project/sedutil/versions perhaps the maintainers don't understand what this tag means. And distributions are most common way for end users to get the software. Should distributions release 1.49.4 instead of 1.20.0.

Also, can you please explain meaning of linux_works tag? Is it work in progress or usable source state?

Curiously, both linux_works and 1.49.4 exist only in feature/windows-opensource branch, not in the master branch (which is with 1.20.0), so this may be a bit confusing for people.

[youk]

@r0m30, apologies for pinging you, but do you still have access to this repo? Since the author of the problematic commit went AWOL, there is a need to fix this somehow.

[r0m30]

@youk I don't know if I still have access to the repo. I haven't checked since the first of the year.

[scottcmarks]

This was on a feature branch, so ymmv. Those tags are now on the develop branch, still entirely tentative.
Insofar as not building on linux, afaik the usual

cd ~/sedutil
autoreconf -i
./configure --enable-silent-rules
make all

with optional

sudo make install

has been percolating for me quite reliably. Linux, libc etc versions on request as I update them and rebuild several times a day.

The idea of a 1.49 version is that it is just under a 1.5 version, which in turn is halfway to a 2.0 version, which is what I am heading toward. The idea, which seems to be panning out, is that by creating an explicit DtaDrive layer adhering to the standards for Ata (HD), Sata (SAT passthrough), Scsi ("direct" Scsi or SAS), and NVMe there is almost no need for separate linux, etc. (OS-specific) subfolders full of duplicative code. (Although it might be time to retire HD Ata code, the non-SCSI-intermediated NVMe code presents essentially the same problems, so two birds with one stone.). Instead, each will contain the implementation of a singleton of its subclass of DtaOS (e.g. DtaLinux) with its OS-specific implementation of eight or so methods, e.g. PerformSCSICommand, which itself takes standard-set (OS-independent) parameters. The 1.5 version is just that; in the 2.0 architecture, the interface to DtaDrive and its kin will be simplified to identify, IF_SEND, and IF_RECV, and the "command-writing" code of the many utility functions will be trimmed down by using the new closure-based wrappers for the common cases.