'svn up' locks up in svn_diff__lcs()

[wayne roberts <wr...@cox.net>]

running subversion-1.4.2 on fedora:
I was doing an svn update, and found that the client was blocking in 
svn_diff__lcs() in this loop:
do {
    ... calls svn_diff__snake() from here ..
} while (fp[d].position[1] != &sentinel_position[1]);

i added a printf to print the values of both pointers fp[d].position[1] 
and &sentinel_position[1]
and found they were never becoming equal and the loop would run forever.

the files that its hanging on are not source-code files.  it is a text 
list of filenames which is about 260,000 lines long.  The workaround is 
to delete my local copies, lose my local changes.

---------------------------------------------------------------------
To unsubscribe, e-mail: users-unsubscribe@subversion.tigris.org
For additional commands, e-mail: users-help@subversion.tigris.org

RE: Re: 'svn up' locks up in svn_diff__lcs()

[we...@tigris.org]

This issue still exists in 1.5.0 .
svn_diff__lcs keeps calling svn_diff__snake forever.
Issue reproduces good.
Any things I can try ?

-- Hans Lambermont
   hlambermont at virtualproteins ddotcom

------------------------------------------------------
http://subversion.tigris.org/ds/viewMessage.do?dsForumId=1065&dsMessageId=1728582

To unsubscribe from this discussion, e-mail: [users-unsubscribe@subversion.tigris.org].

Re: 'svn up' locks up in svn_diff__lcs()

[Ed Price <ed...@gmail.com>]

On 11/23/06, wayne roberts <wr...@cox.net> wrote:
> running subversion-1.4.2 on fedora:
> I was doing an svn update, and found that the client was blocking in
> svn_diff__lcs() in this loop:
> do {
>     ... calls svn_diff__snake() from here ..
> } while (fp[d].position[1] != &sentinel_position[1]);
>
> i added a printf to print the values of both pointers fp[d].position[1]
> and &sentinel_position[1]
> and found they were never becoming equal and the loop would run forever.
>
> the files that its hanging on are not source-code files.  it is a text
> list of filenames which is about 260,000 lines long.  The workaround is
> to delete my local copies, lose my local changes.

Sounds like an interesting bug... No followup?

Could you come up with a simple reproduction recipe?

How long did you wait before declaring it "forever"?
(Or can you tell "deductively" from the code that it won't halt?)

-ed

---------------------------------------------------------------------
To unsubscribe, e-mail: users-unsubscribe@subversion.tigris.org
For additional commands, e-mail: users-help@subversion.tigris.org