4 ms·
Amusing, but after looking up on the actual reasons why they came up with this scheme, I'm very impressed by the amount of work that goes into backwards compati
by magic_haze 13y ago
Amusing, but after looking up on the actual reasons why they came up with this scheme, I'm very impressed by the amount of work that goes into backwards compatibility:
> Batch files ran programs from C:\Windows\System32 with the expectation that the resulting program would match the native OS. These expectations were not explicit, but they were implied by the nature of the activity. If the System32 directory were filled with 32-bit programs, then a batch file that ran the C:\Windows\System32\REG.EXE program to upset a system registry setting would be running the 32-bit version of REG.EXE, which means that it would be updating the 32-bit simulated version of the registry instead of the real 64-bit version. Other types of scripting files (such as REG files) have the same problem.
http://technet.microsoft.com/en-us/magazine/ff955767.aspx http://technet.microsoft.com/en-us/magazine/ff955767.aspx
How did Linux handle this problem?
- byroot 13y agoAFAIK it just don't have the problem. 32bits binaries executed on a 64bits system are not isolated in any way.
- _ZeD_ 13y agothe problem is not in the 32/64 reg.exe executable. It's that "32-bit simulated version of the registry". Why do I need something like that? (I expect some binary compatibility, but we're just opening a can of worms...) By comparison, most of the configuration files / registry equivalent info in linux are saved in plain text, and AFAIK all of them did not care about the 32/64 bit version of the executables.
- workbench 13y ago> 32-bit simulated version of the registry > How did Linux handle this problem? Not having something as terrible as the registry in the first place.