[Bug default/31313] New: infinitish run for abidw on libgs-9.27-1.el8.x86_64.rpm /usr/lib64/libgs.so.9.27

[Bug default/31313] New: infinitish run for abidw on libgs-9.27-1.el8.x86_64.rpm /usr/lib64/libgs.so.9.27

[fche at redhat dot com]

https://sourceware.org/bugzilla/show_bug.cgi?id=31313

            Bug ID: 31313
           Summary: infinitish run for abidw on
                    libgs-9.27-1.el8.x86_64.rpm /usr/lib64/libgs.so.9.27
           Product: libabigail
           Version: unspecified
            Status: NEW
          Severity: normal
          Priority: P2
         Component: default
          Assignee: dodji at redhat dot com
          Reporter: fche at redhat dot com
                CC: libabigail at sourceware dot org
  Target Milestone: ---

Created attachment 15342
  --> https://sourceware.org/bugzilla/attachment.cgi?id=15342&action=edit
binary

git/master abidw on a RHEL binary results in days-long run with no apparent
progress.
The binary is /usr/lib64/libgs.so.9.27 from RHEL 8.4.0's
libgs-9.24-1.el8.x86_64.rpm.
Attaching stack traceback and gzipped binary.

-- 
You are receiving this mail because:
You are on the CC list for the bug.

[Bug default/31313] infinitish run for abidw on libgs-9.27-1.el8.x86_64.rpm /usr/lib64/libgs.so.9.27

[fche at redhat dot com]

https://sourceware.org/bugzilla/show_bug.cgi?id=31313

--- Comment #1 from Frank Ch. Eigler <fche at redhat dot com> ---
Created attachment 15343
  --> https://sourceware.org/bugzilla/attachment.cgi?id=15343&action=edit
eu-stack -n0 -s -m -v -p $PID   output, after days of runtime

-- 
You are receiving this mail because:
You are on the CC list for the bug.

[Bug default/31313] infinitish run for abidw on libgs-9.27-1.el8.x86_64.rpm /usr/lib64/libgs.so.9.27

[dodji at redhat dot com]

https://sourceware.org/bugzilla/show_bug.cgi?id=31313

dodji at redhat dot com changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
             Status|NEW                         |ASSIGNED

--- Comment #2 from dodji at redhat dot com ---
Hello,

I could reproduce the issue, thanks for reporting the issue.

This appears to be one these embarrassing quadratic behaviour that the current
type canonicalization process has :-(

I tried the work-in-progress type-hashing branch on the binary and here is what
I am getting:

$ time ~/git/libabigail/hash-types/build/tools/abidw -d usr/lib/debug
usr/lib64/libgs.so.9.27 > libgs.so.9.27.hashed.abi

real    0m45,004s
user    0m44,734s
sys     0m0,142s
$ 

In other words, the current type-hashing work on the branch
https://sourceware.org/git/?p=libabigail.git;a=shortlog;h=refs/heads/users/dodji/hash-types
seems to address precisely this kind of issues.

I'll look into this a bit more to see if I can find a work-around before I am
done with the type-hashing work.  But we might as well wait for type-hashing
work to land.

-- 
You are receiving this mail because:
You are on the CC list for the bug.