The current wine (1.6.1-1) uploaded to cauldron on Nov 18 appears to have broken large HTTP downloads using InternetReadFile(). The only reproducible case Ii have involves using the EZDownloader function of the commercial UnfairAdvantage product, which downloads stock and mutual fund data. A wireshark trace shows that after about 2MB of a 7MB file is downloaded, we start seeing TCP Dup ACKs, packet tretransmission, out-of-sequence packets, and reassembly errors. Eventually, the application gets an ERROR_NO_MORE_FILES, and aborts. I tried the same operation on an older cauldron with a wine executable date of Jul 18, and it works fine. Also, the same download using a URL in firefox works correctly. I couldn't find anything similar in Google. If we don't have local wine patches that would cause this, I request that 1.6.1-1 be withdrawn until upstream can correct this. Reproducible: Steps to Reproduce:
I downloaded and built the upstream versions of 1.6 (placed in cauldron around last July), 1.6.1 (placed in cauldron around mid-November), and the just-released development version, 1.7.7. The problem does occur with upstream 1.6.1, so this is an upstream problem. It does *not* happen with either 1.6 or 1.7.7, so we have the choice of either going back or going forward.
Keywords: (none) => TriagedAssignee: bugsquad => mageia
Keywords: (none) => UPSTREAMURL: (none) => http://bugs.winehq.org/show_bug.cgi?id=35053
Did you reported it upstream? We can't go with 1.7.7 as is unstable release but we can wait for 1.6.x to have a fix for your issue (as 1.7.7 is not affected, it should be a good backport to have from upstream).
Status: NEW => ASSIGNED
Thank you for your bug report and the link. Please test wine-1.6.1-2.mga4 from updates_testing.
Hardware: x86_64 => AllSource RPM: wine => wine-1.6.1-1.mga4.src.rpm
(In reply to Damien Lallement from comment #3) > Thank you for your bug report and the link. > Please test wine-1.6.1-2.mga4 from updates_testing. Working fine with this package, thanks !
Thanks for the feed back, package will arrive in core/release soon!
Done.
Status: ASSIGNED => RESOLVEDResolution: (none) => FIXED
But it's not yet pushed (could be great to not close before it's done, something similar happens with mga3, and it was never really fixed until release, event some months after with a release) (will copy this next time I saw similar closing ;) )
Sorry, I will wait! Don't bite me please.
Status: RESOLVED => REOPENEDResolution: FIXED => (none)
CC: (none) => tmb
now pushed.
Status: REOPENED => RESOLVEDResolution: (none) => FIXED