Issue 39255 - Problems with Unicode Byte Order Mark (BOM) at start of manifest.xml
Summary: Problems with Unicode Byte Order Mark (BOM) at start of manifest.xml
Status: CLOSED FIXED
Alias: None
Product: General
Classification: Code
Component: code (show other issues)
Version: 680m65
Hardware: All All
: P3 Trivial (vote)
Target Milestone: OOo 2.0
Assignee: mci
QA Contact: issues@framework
URL: http://www.unicode.org/unicode/faq/ut...
Keywords: crash, oooqa
Depends on:
Blocks:
 
Reported: 2004-12-18 10:50 UTC by jlp
Modified: 2005-03-01 10:47 UTC (History)
4 users (show)

See Also:
Issue Type: DEFECT
Latest Confirmation in: ---
Developer Difficulty: ---

Attachments
document form issue 39232 with BOM (U+FEFF) at start of manifest.xml (10.34 KB, application/vnd.sun.xml.math)
2004-12-18 22:53 UTC, lohmaier 		no flags Details
Add an attachment (proposed patch, testcase, etc.)

Note You need to log in before you can comment on or make changes to this issue.
Description jlp 2004-12-18 10:50:34 UTC 
After I upgraded from OOo 1.9.62 to OOo 1.9.65 I had to change some files (for
example META-INF/manifest.xml) inside the documet to change the MIME file type
to correct one. More info here:
http://development.openoffice.org/releases/1.9.m65_snapshot.html
I did this with UltraEdit on Windows XP but it looks like it adds unicode Byte
Order Mark (BOM) characters at the very begining of the file. It looks like this
crashes OOo on Windows. If I understand it correctly on Linux a document with
such a manifest.xml causes OOo not to display the document, but it doesn't crash.

I think that OOo should just ignore the BOM at the begining of the file and load
the document, without crashing.

Here is the report from crash reporter on Windows XP:

<?xml version="1.0" encoding="UTF-8"?>
<!DOCTYPE errormail:errormail PUBLIC "-//OpenOffice.org//DTD ErrorMail 1.0//EN"
"errormail.dtd">
<errormail:errormail xmlns:errormail="http://openoffice.org/2002/errormail"
usertype="">
<reportmail:mail xmlns:reportmail="http://openoffice.org/2002/reportmail"
version="1.1" feedback="true" email="jlp@holodeck1.com">
<reportmail:title></reportmail:title>
<reportmail:attachment name="description.txt"
media-type="text/plain;charset=UTF-8" class="UserComment"/>
<reportmail:attachment name="user.dmp" media-type="application/octet-stream"
class="UserDump"/>
</reportmail:mail>
<officeinfo:officeinfo xmlns:officeinfo="http://openoffice.org/2002/officeinfo"
build="680m65(Build:8851)" platform="wntmsci10.pro" language=""
procpath="D:\Programs\OpenOffice.org\program\" exceptiontype="0xC0000005"
product="OpenOffice.org 1.9.65"/>
<systeminfo:systeminfo xmlns:systeminfo="http://openoffice.org/2002/systeminfo">
<systeminfo:System name="Windows NT" version="5.1" build="2600"
locale="0x00000424"/>
<systeminfo:CPU type="x86"/>
</systeminfo:systeminfo>
<errormail:Stack type="Win32">
<errormail:StackInfo pos="0" ip="0x6777B224" rel="0x0005B224"
ordinal="Ordinal8752+0x00000022" name="bf_svx680mi.dll"
path="D:\Programs\OpenOffice.org\program\"/>
<errormail:StackInfo pos="1" ip="0x6777CEB0" rel="0x0005CEB0"
ordinal="Ordinal8735+0x00000024" name="bf_svx680mi.dll"
path="D:\Programs\OpenOffice.org\program\"/>
<errormail:StackInfo pos="2" ip="0x67830E46" rel="0x00110E46"
ordinal="Ordinal11362+0x0000023D" name="bf_svx680mi.dll"
path="D:\Programs\OpenOffice.org\program\"/>
<errormail:StackInfo pos="3" ip="0x67830C60" rel="0x00110C60"
ordinal="Ordinal11362+0x00000057" name="bf_svx680mi.dll"
path="D:\Programs\OpenOffice.org\program\"/>
<errormail:StackInfo pos="4" ip="0x61D51879" rel="0x00051879"
ordinal="Ordinal196+0x00000011" name="vcl680mi.dll"
path="D:\Programs\OpenOffice.org\program\"/>
<errormail:StackInfo pos="5" ip="0x61D8FF37" rel="0x0008FF37"
ordinal="Ordinal5649+0x00000E7C" name="vcl680mi.dll"
path="D:\Programs\OpenOffice.org\program\"/>
<errormail:StackInfo pos="6" ip="0x61D403B4" rel="0x000403B4"
ordinal="Ordinal12392+0x0000001E" name="vcl680mi.dll"
path="D:\Programs\OpenOffice.org\program\"/>
<errormail:StackInfo pos="7" ip="0x61D466DE" rel="0x000466DE"
ordinal="Ordinal12403+0x00000FFD" name="vcl680mi.dll"
path="D:\Programs\OpenOffice.org\program\"/>
<errormail:StackInfo pos="8" ip="0x61D46959" rel="0x00046959"
ordinal="Ordinal12403+0x00001278" name="vcl680mi.dll"
path="D:\Programs\OpenOffice.org\program\"/>
<errormail:StackInfo pos="9" ip="0x77D48709" rel="0x00008709"
ordinal="GetDC+0x00000072" name="USER32.dll" path="C:\WINDOWS\system32\"/>
<errormail:StackInfo pos="10" ip="0x77D487EB" rel="0x000087EB"
ordinal="GetDC+0x00000154" name="USER32.dll" path="C:\WINDOWS\system32\"/>
<errormail:StackInfo pos="11" ip="0x77D489A5" rel="0x000089A5"
ordinal="GetWindowLongW+0x00000127" name="USER32.dll" path="C:\WINDOWS\system32\"/>
<errormail:StackInfo pos="12" ip="0x77D489E8" rel="0x000089E8"
ordinal="DispatchMessageW+0x0000000F" name="USER32.dll"
path="C:\WINDOWS\system32\"/>
<errormail:StackInfo pos="13" ip="0x61E2D442" rel="0x0012D442"
ordinal="WIN_PolyPolygon+0x0000299F" name="vcl680mi.dll"
path="D:\Programs\OpenOffice.org\program\"/>
<errormail:StackInfo pos="14" ip="0x61E16850" rel="0x00116850"
ordinal="Ordinal3340+0x00000700" name="vcl680mi.dll"
path="D:\Programs\OpenOffice.org\program\"/>
<errormail:StackInfo pos="15" ip="0x61E168AB" rel="0x001168AB"
ordinal="Ordinal3340+0x0000075B" name="vcl680mi.dll"
path="D:\Programs\OpenOffice.org\program\"/>
<errormail:StackInfo pos="16" ip="0x61E1694C" rel="0x0011694C"
ordinal="Ordinal3340+0x000007FC" name="vcl680mi.dll"
path="D:\Programs\OpenOffice.org\program\"/>
<errormail:StackInfo pos="17" ip="0x61D9B9F5" rel="0x0009B9F5"
ordinal="Ordinal602+0x00000037" name="vcl680mi.dll"
path="D:\Programs\OpenOffice.org\program\"/>
<errormail:StackInfo pos="18" ip="0x61E17978" rel="0x00117978"
ordinal="Ordinal729+0x00000050" name="vcl680mi.dll"
path="D:\Programs\OpenOffice.org\program\"/>
<errormail:StackInfo pos="19" ip="0x00401013" rel="0x00001013"
name="soffice.BIN" path="D:\Programs\OpenOffice.org\program\"/>
<errormail:StackInfo pos="20" ip="0x00401055" rel="0x00001055"
name="soffice.BIN" path="D:\Programs\OpenOffice.org\program\"/>
<errormail:StackInfo pos="21" ip="0x7C816D4F" rel="0x00016D4F"
ordinal="RegisterWaitForInputIdle+0x00000049" name="kernel32.dll"
path="C:\WINDOWS\system32\"/>
</errormail:Stack>
<errormail:Checksums type="MD5">
<errormail:Checksum sum="0x4EC64389FBA5D22349989D37F0960FBF" bytes="630784"
file="soffice.BIN"/>
<errormail:Checksum sum="0x53A9A2192A681B10D95BE68345E4F48C" bytes="2854912"
file="bf_svx680mi.dll"/>
<errormail:Checksum sum="0x888190E31455FAD793312F8D087146EB" bytes="983552"
file="kernel32.dll"/>
<errormail:Checksum sum="0x1EBA5C0F2C4D902067BC9BB05B8A7421" bytes="1912832"
file="vcl680mi.dll"/>
<errormail:Checksum sum="0xC72661F8552ACE7C5C85E16A3CF505C4" bytes="577024"
file="USER32.dll"/>
</errormail:Checksums>
</errormail:errormail>
Comment 1 askoorb 2004-12-18 17:44:53 UTC 
For the assigned developer:

this issue was mostly confirmed as being real in issue 39232.  It is up to you
whether to confirm it totaly though.

I can confirm on WinXP SP2.
Comment 2 askoorb 2004-12-18 18:29:26 UTC 
Oh, well, i'd better take qa of this issue:

Confirming.

Also changed the OS to WinXP as Windows 95 does not have much support at all for
unicode.
Comment 3 lohmaier 2004-12-18 22:51:28 UTC 
extended summary, added crash-keyword.
Comment 4 lohmaier 2004-12-18 22:53:30 UTC 
Created attachment 20637 [details]
document form issue 39232 with BOM (U+FEFF) at start of manifest.xml
Comment 5 mikhail.voytenko 2005-01-04 15:20:33 UTC 
The problem seems to take place in the sax parser because of unexpected symbols.
Please see a part of more detaled stack provided by TL below:

sax.uno.dll!_STL::__node_alloc<1,0>::deallocate()  + 0x20cb     C
sax.uno.dll!_XML_GetCurrentLineNumber()  + 0x29 C
sax.uno.dll!sax_expatwrap::LocatorImpl::getLineNumber()  + 0x14 C++
sax.uno.dll!sax_expatwrap::SaxExpatParser_Impl::parse()  + 0x8f C++
sax.uno.dll!sax_expatwrap::SaxExpatParser::parseStream()  + 0x217       C++
package2.dll!ManifestReader::readManifestSequence()  + 0x120    C++
package2.dll!ZipPackage::getZipFileContents()  + 0x482  C++
package2.dll!ZipPackage::initialize()  + 0x548  C++
[...skipped...]
Comment 6 lohmaier 2005-01-16 16:25:54 UTC 
os -> ALL m71 crashes when selecting the file in the File|Open dialog.

It crashes before you can click "open", just selecting it is enough to make it
crash.
Comment 7 mci 2005-02-03 12:02:33 UTC 
reassigned to mav

mci -> mav: 
hi mav, please take take of this issue, thanks... :)

This worked in earlier versions...
Comment 8 mci 2005-02-03 12:31:15 UTC 
sorry, working on the wrong issue... :(

I don't know if this worked in earlier versions!

I sent this to mav since mav sent a comment to this issue and knows about the
prblem...
Comment 9 mikhail.voytenko 2005-02-03 12:37:15 UTC 
MAV->DVO: The crash takes place in sax parser. Please take a look.
Comment 10 mikhail.voytenko 2005-02-04 12:39:10 UTC 
MAV->DVO: The crash takes place in sax parser. Please take a look.
Comment 11 openoffice 2005-02-07 10:50:23 UTC 
dvo: Fixed in CWS swqcore06.

Recognition for UTF-8 BOM (byte order mark) was missing. Strangely, recognition
for some other BOMs was implemented, but erraneous due to implicit C++ type
conversion. That is fixed, too.
Comment 12 openoffice 2005-02-14 10:17:28 UTC 
dvo->mci: Please test.
re-open issue and try to reassign to mci@openoffice.org
Comment 13 openoffice 2005-02-14 10:17:35 UTC 
try to reassign to mci@openoffice.org
Comment 14 openoffice 2005-02-14 10:17:41 UTC 
try to reset resolution to FIXED
Comment 15 mci 2005-02-17 15:38:44 UTC 
bugdoc opens fine in cws swqcore06 on Solaris/Sparc...
Comment 16 mci 2005-02-18 08:35:27 UTC 
bugdoc opens in cws swqcore06 on Linux and Windows
Comment 17 mci 2005-02-18 10:16:52 UTC 
bugdoc loads on all plattform using cws swqcore06

==> verified
Comment 18 mci 2005-03-01 10:47:08 UTC 
bugdoc opens in m81

==> ok

==> closing