Hi,
here is the third iteration of the retrace server design.
News:
- added the synchronous mode
- no more AES for generating passwords, random number generator is
used instead of that
- the HTTP interface must accept .tar and .tar.gz files in addition to
.tar.xz
- QEMU user mode emulation is mentioned as an alternative to
cross-target GDB builds
- 30 MB .xz limit not mentioned much
- better paths: /var/spool/abrt-{retrace,repo}
- added a requirement to explain an error code in the response body to
distinguish between various limits
- HTTP persistent connections
- a note about prelink
Michal Toman has already been writing some retrace worker code. I think
now it's the right time for that.
I also attached an HTML version.
Karel
Show replies by date