User-Agent: Mozilla/5.0 (Windows; U; Win98; en-US; rv:1.0.1) Gecko/20020823 Netscape/7.0 Build Identifier: ftp.mozilla.org/pub/mozilla/releases/mozilla1.2.1/mozilla-win32-1.2.1-installer.exe When attempting to download a file from http://www.blindwrite.com/, the file was corrupted on each try. The download went well when i used Internet Explorer Reproducible: Always Steps to Reproduce: 1.Go to blindwrite download page and click on one of the links for "English only-no manuals" 2.Download file using download manager 3.Try to run file after download Actual Results: The file was much smaller than it should have been. When i tryed to run the file it said that it was corrupted OS is Windows ME

I have a similar problem with the nightly build for Solaris 2.7 Build ID: 0000000000 (Mozilla/5.0 (X11; U; SunOS sun4u; en-US; rv:1.3b; MultiZilla v1.1.33 (b)) Gecko/20030127) When I try to download ftp://ftp.berlios.de/pub/cdrecord/alpha/win32/cdrtools-1.10-win32-bin.zip with the download manager, I get a corrupted file with the correct file size. The same file downloaded with wget has the right CRC.

waschk, please don't set flags if you don't know how.

I just wanted to say that I too am having a problem with corrupted files from the Download Manager. For example, when I download the mozilla.zip file from http://ftp.mozilla.org/pub/mozilla/nightly/latest/ I get a zip file of the same size as when I use Internet Explorer. But when I unzip it, winzip reports three CRC errors. But the one from IE unzips without any errors. I have tried it three different times. Win XP, direct internet connection, no proxy.

Exactly the same problem. I run Moz 1.3.1 on both Windows 98 and Linux boxes. Downloads are *unreliable* on both OS's - to the extent that I have to use Ie5.

Mozilla/5.0 (Macintosh; U; PPC Mac OS X Mach-O; en-US; rv:1.4) Gecko/20030624 Mac OSX 10.2.6 Darwin Kernel Version 6.6 I just "upgraded" from Mozilla 1.3 to 1.4, hit a bug in mail, tried twice (now with 1.4) to download the latest build before reporting the mail bug, but I get a file full of zeros. % ls -al mozilla-mac-MachO.dmg.gz -rw-r--r-- 1 ### ### 15993602 Aug 26 19:10 mozilla-mac-MachO.dmg.gz % hexdump mozilla-mac-MachO.dmg.gz 0000000 0000 0000 0000 0000 0000 0000 0000 0000 * 0f40b00

Mozilla/5.0 (Windows; U; Windows NT 5.1; en-US; rv:1.5) Gecko/20031007 Windows XP SP1 On several occassions I encountered corrupted downloads of multimedia files. The files (WAV and MPG) were beyond recognition. This happened when doing concurrent downloads (multiple files at a time). When opening one file from the web right into the player app, the download was ok.

In addition to my previous posting: download corruption also happens when doing a single download.

I have the same problem with files that download (sometimes download does not even start - separate bug) being invalid files. I have not successfully downloaded a file (other than "save image as") for several months with Mozilla 1.6 and now 1.7 RC and I am forced to open Explorer for all download activity. issue: file download "completes" but file is bad reproduce: always (if download starts - separate bug) Mozilla 1.7 RC Windows 98 SE description: file to download (any type) will be smaller when "completed" than reported by server and therefore invalid. (ex. .avi file to download 1874kb reported by webserver, 1851kb recieved but download dialog shows "complete", invalid file... unable to open file)

I can confirm the same bug...Happens all the time and I have to use IE to d/l the files I want. For example go to http://www.math.uga.edu/~djb/html/dvd.html and try to d/l (save to disk) FreeZoneLG6053.exe which is 52,551 bytes long. The saved file is something less and thus corrupt.

spyros: i just downloaded FreeZoneLG6053.exe from the above site with firefox 0.9.3 and mozilla 1.7b and was unable to reproduce the issue. i did see it long ago, though. perhaps you are running into some sort of network issue between you and the server. maybe dumping the packets with ethereal would help you get to the bottom of this? you could compare the bad download to a good one with IE...

*** This bug has been confirmed by popular vote. ***

One of the ways that music firms try to discourage pirated music downloads is by uploading corrupt files to illicit servers. I don't know if bad software is sent to illicit software servers, but this seems like a possibility to consider. I downloaded both the cdtools zipfile and the Freezone program and unzipped the cdtools file ok. The Freezone program has the right size, though I didn't try to run it. I've successfully downloaded each new Mozilla version with the preceding one for a year and a half. If one is experiencing corrupted downloads, there must be some other explanation than a problem with the integrity of Mozilla's data transport.

Download of large DIVX files in FF seem to be corrupted in FF, same files downloaded in IE, Opera and NS are working fine.

*** Bug 277727 has been marked as a duplicate of this bug. ***

*** Bug 279424 has been marked as a duplicate of this bug. ***

*** Bug 274552 has been marked as a duplicate of this bug. ***

*** Bug 211304 has been marked as a duplicate of this bug. ***

This bug must be a particularly tricky one to squash -- it's been open for two+ years! It (and its relatives, whether of the sequential or multiple download type) seem quite virulent, given the number of reported occurences. It *bugs* me that this is a long-standing problem. I've switched to Firefox, EXCEPT when I need to download files (use IE or something using the IE engine). It's really the only fundamental area where I feel Firefox is lacking... :( I suspect that, for a complex problem of this type, the developers need better information in order to reproduce the behavior. If I could get from them a better idea of what information is desired, perhaps I could construct a better test case? Or... is the work-around to install an external download manager? Is that what the rest of you are doing?

Verified in Mozilla/5.0 (Windows; U; Windows NT 5.1; en-US; rv:1.7.6) Gecko/20050225 Firefox/1.0.1

*** Bug 252719 has been marked as a duplicate of this bug. ***

I'm not so sure this is a duplicate of 252719. Most of the comments here are about a bug where the file size is too small. Yes, Firefox does seem to give up on a file transfer before it is complete and the file ends up being too small. However, I've also seen file sizes be completely correct when downloading multiple files but the md5sums of the files don't match. This would imply (though this hasn't been definitely confirmed) that (1) cross-contamination of data, and if (1) applies then so does (2): a threat that someone who understands the contamination pattern could take advantage of it to insert malicious code.

BAD DOWNLOADS happen a lot recently from Mozilla's own servers. I've been using nightlies most of the time for the past month or two, and in the past two weeks or so have been getting a lot of defective downloads. Today I downloaded the nightly dated 4-10 for Win32, and got a file sized 11,343,904 bytes. The official size is 11,992K according to the server I'm trying again on. It's not the same server, or at least not the same interface. The first time, I used a traditional ftp interface, and now it's a web interface from a German mirror, http://ftp.mozilla.org/pub/mozilla.org/mozilla/nightly/latest-trunk/ gotta go, damn thunderstorm.s here.

To continue, the download completed without any errors reported. I thought that tcp/ip networking included detection of bad packets and requests for resending. This never was a problem doing downloads with Mozilla until very recently, perhaps the past month. Today when I ran the download to install the latest, the installer first reported xpcom.xpi and another xpi module being corrupt. I tried running it again and got just a crash. I then checked the download size and found that the download failed to complete, but the Download Manager indicated it was complete and didn'r report errors. I think this new set of downloading problems got into the 1.8x code while perhaps not into the 1.7x code. I switched back to 1.7x earlier because of some other types of download failures in 1.8a6. These failures are unacceptable. Downloading ought to be a mastered art by now.

To continue, the download completed without any errors reported. I thought that tcp/ip networking included detection of bad packets and requests for resending bad packets. Today when I ran the download to install the latest, the installer first reported xpcom.xpi and another xpi module being corrupt. I tried running it again and got just a crash. I then checked the download size and found that the download failed to complete, but the Download Manager indicated it was complete and didn'r report errors. I think this new set of downloading problems got into the 1.8x code while perhaps not into the 1.7x code. I switched back to 1.7x earlier because of some other types of download failures in 1.8a6. It may be that I didn't notice earlier problems with 1.8x because I stopped using it for a while, but someone should have been noticing and reporting problems downloading in 1.8x. Serving and receiving ftp and http downloads should be a mastered art by now.

Sorry about the duplicate reports. 11,343,904 bytes was the size of the defective download, reported complete with no errors. The second try was good, 12,279,420 bytes.

Today, in trying to download the latest nightly, I was asked if the old mozilla-win32-installer.exe should be overwritten by the new one. I said yes. Then the download started, but after a few seconds the download manager reported that it was finished. I deleted the old nightly file and retried. This time, I noticed that the download was reported complete much too soon, though the download manager reported the right number of bytes for the new file. Here's the real file info---wrong size, obviously not complete. MOZILL~7 EXE 1,770,660 04-16-05 5:19p mozilla-win32-installer.exe The download manager wrongly reported the download as done, wrongly reported the file size, and gave no error messages. The dialup connection stayed up, so that was not the reason for the failed download. This happened using Mozilla/5.0 (Windows; U; Win98; en-US; rv:1.8b2) Gecko/20050411

The download manager seems to stop downloading when another process hogs the I/O device. For example, my email program shows a box to approve a certificate from my real ISP (instead of my mail domain name) and the download from Mozilla stops until the approval request box is dismissed. Then, if the download has quit (either the server or the local program timing out because of inactivity, not sure which), the download manager MARKS THE DOWNLOAD AS COMPLETE AND REPORTS THE CORRECT FILESIZE. The actual file on the disk is of course not complete, but it's renamed from .par to .exe as if it is complete. This happened repeatedly to me today, all due to the security warning box from Mozilla mail (questioning certificate approval for a secure connection) hogged the I/O and stopped the downloading. Ftp/http downloads, page loads, and chat should not be stopped dead in the water by another process's dialog box. The processes should be able to run concurrently even if one is waiting on dialog, etc.

john t: re: comment #28, that is bug 74331. nb: i still see this same behavior with recent firefox releases. i opened bug 301894 to track this for the FF product.

Never mind my remarks here about bad downloads. I found that the disk drive I saved to went bad. Sorry.

While downloading the firefox download to initially install Firefox, I received the corrupt file message when I clicked on the "run" on the message that internet explorer gives at the conclusion of a download. Once I shut IE down and ran the Firefox install program from the "run" command, it opened just fine. You may wish to instruct people who are initially downloading the program to initiate installation by shutting down IE. Further anomalys: when I attempted to download Mozilla-win32-1.7.12-installer.exe (finished @ aprox 9:28pm 12Oct2005), It transfered at only 6k/sec. I was using my new installation of Firefox to initiate it. Typical cable download speeds are usually around 300-500kps. Are you guys serving with 56k modems? (-: I am using wavecable.com (Wave Broadband inc). Their headquarters are in Kirkland WA. There may be interferance eminating from Redmond and affecting broadband transmissions from companies in close proximity to Redmond.

The download mgr messed up when I paused a download, went offline to use my phone, then reconnected and tried to resume. I couldn't resume because the download mgr and properties box indicated that the download was complete, though it certainly was NOT complete. This happens with stalled downloads, I think, but I expected that the pause button might allow for going offline and back online to resume a download, since the program wasn't exited. If the pause button doesn't provide that function, what good is it?

This Oct 26 experience happened with Mozilla/5.0 (Windows; U; Win98; en-US; rv:1.9a1) Gecko/20051021 SeaMonkey/1.1a

related to or dup of Bug 237623, Bug 87151, bug 18004 removed expired blocker reporter?

Comment on attachment 171406 [details] the original image as hosted on another site: http://sis.agr.gc.ca/cansis/taxa/soil/brunisolic/eutric_yt.jpg image is corrupted. Also, reporter is gone.

I recently read a post to www.linuxquestions.org mentioning two programs (uxcook and uxcook 95?) that are made to fix downloads from improperly configured servers that work with MSIE but apparently not with UNIX downloaders. It may be that Mozilla/Firefox is having problems with similarly configured servers, which I assume have nonstandard settings or behavior. Could these misconfigured servers be a source of corrupt Mozilla and Firefox downloads? If so, would adding some uxcook code fix the problem? I assume uxcook is GNU or otherwise freely available.

WFM trunk - tested the reported URLs. I've never had a problem Mozilla/5.0 (Windows; U; Windows NT 5.1; en-US; rv:1.9a1) Gecko/20060206 SeaMonkey/1.5a

no response from reporter, resolving WFM there is an existing bug on file for what's described in comment 32 (resuming downloads)