Bugs item #3028241, was opened at 2010-07-12 06:25
Message generated for change (Comment added) made by bagder
You can respond by visiting:
https://sourceforge.net/tracker/?func=detail&atid=100976&aid=3028241&group_id=976

Please note that this message will contain a full copy of the comment thread,
including the initial issue submission, for this request,
not just the latest update.
Category: client module
Group: crash
Status: Open
>Resolution: Fixed
Priority: 5
Private: No
Submitted By: Jon Sargeant (nordock)
Assigned to: Daniel Stenberg (bagder)
Summary: Retry causes access violation with URL part sets cont'd

Initial Comment:
I can't figure out how to modify my old bug report so I'm submitting a new one. This bug is very hard to reproduce. I left out some important details in my old bug report. First, you have to specify an output file. Second, one download (i.e., one part of the part set) must succeed in order for a failed download to crash. Third, the failed download must not retrieve any data before failing. In the end, I had to hack main.c in order to satisfy the third condition. I also added trace messages to make the problem clearer. I'm attaching these changes as a patch diff'ed against curl 7.21.0.

Here is the command line I entered:
curl --retry 2 "http://www.google.{com,xxx}" -o "temp#1.csv"

And here is the output from the program:
RESETTING OUTS.BYTES
RESETTING OUTS.STREAM
OUTS.BYTES=0

[1/2]: http://www.google.com --> tempcom.csv
  % Total % Received % Xferd Average Speed Time Time Time Current
                                 Dload Upload Total Spent Left Speed
100 8199 0 8199 0 0 14989 0 --:--:-- --:--:-- --:--:-- 14989
CLOSING OUTS.STREAM
RESETTING OUTS.STREAM
OUTS.BYTES=8199

[2/2]: http://www.google.xxx --> tempxxx.csv

GOT RESPONSE 1245032 BUT LET'S PRETEND IT'S 400
Warning: Transient problem: FTP error Will retry in 1 seconds. 2 retries left.
Throwing away 64d bytes
ACCESSING OUTS.STREAM: (nil)
>>> CRASH <<<

This should be enough information to pinpoint the problem.

----------------------------------------------------------------------

>Comment By: Daniel Stenberg (bagder)
Date: 2010-07-14 23:53

Message:
Great recipe, thanks a bunch!

Using this, I completely understand the problem and I've pushed a fix with
commit c8d42b2f1c1.

I consider this case closed now. Again, thanks for your help and patience!

----------------------------------------------------------------------

You can respond by visiting:
https://sourceforge.net/tracker/?func=detail&atid=100976&aid=3028241&group_id=976
Received on 2010-07-14