I've launched debug_info test at the end of the last week. I modified
dwarflint & libdw to print out every memory location considered for
relocation. The test itself then is to use readelf -winfo and dwarflint
to collect reloc prints, and expect them to be the same.
The test takes ages. I had to restart it on Monday due to trivial bug
in method that I've made, and it's still running as of now. I had to
tweak the method a couple times to get rid of some false positives (e.g.
absence of .debug_info), but there were no real failures so far, and I
expect that hooks are doing the right thing WRT .debug_info.
PM
Show replies by date
Sounds good! I did a final grep 'read_[48][us]byte' on the branch,
and didn't see anything still using it that should be relocatable.
Please merge master into your branch so diff master..pmachata/reader_hooks
looks pure. After that, Uli can do a review and if he has no issues
we can merge it into master.
Thanks,
Roland
Roland McGrath пишет:
Please merge master into your branch so diff
master..pmachata/reader_hooks
looks pure.
Done.
PM