SLE-11 SP1: Use stdin, not tmp files for passwords (bsc#986971) #226
Conversation
| @@ -256,6 +288,11 @@ SystemCmd::doExecute( string Cmd ) | |||
| y2error( "dup2 stderr child failed errno=%d (%s)", errno, | |||
| strerror(errno)); | |||
| } | |||
| if( close( sin[1] )<0 ) | |||
| { | |||
| y2error( "close child failed errno=%d (%s)", errno, | |||
There was a problem hiding this comment.
NP (I wouldn't change already tested code for this): the original patch used "close (stdin) child failer errno", which is probably convenient for easily distinguish from the error closing sout below.
There was a problem hiding this comment.
Right, but I wanted to keep it as similar to the surrounding code as possible -- and there it's similarly impossible to tell which close() went wrong - stdout or stderr? Now it's one more to guess from. But when that happens, something went badly wrong already, so it's just cosmetic which of those close() calls failed.
|
After a close inspection, I confirm this patch looks 100% equivalent to the SP4 one, so LGTM. |
|
I have compared it to the https://github.com/yast/yast-storage/pull/223/files change and nothing was forgotten. So again, small LGTM from me 😉 |
Backport to SLE-11 SP1.
Since there were a lot of changes between SP1 and SP3, the SP3/SP4 patch did not apply at all, so this is very much a manual effort.
Still, it is quite straightforward to compare this PR manually to the PR against SP4:
https://github.com/yast/yast-storage/pull/223/files
I also created RPMs in the IBS with this version, built a DUD, integrated that DUD into the SLES-11-SP1 ISO and tested with that ISO: It installs without problem, and the created encrypted LVM can successfully be decrypted so that system boots without any problems.