The prldap_tsd_destroy() function (which is called when a
thread exits) was not freeing the information contained within
the PRLDAP_ErrorInfo structure. Added prldap_free_errorinfo()
and a way to determine if that thread-private data looks like
error information (the plei_magic field). At the moment, only
one kind of thread-private data is stored anyway (the error
information).
The prldap_allocate_map() function, which is called when a new
LDAP session is created, was blindly setting the thread-private
error information pointer to NULL. But if a different thread
created or used an LDAP session than called ldap_unbind(), old
error information may have been left in memory. Now the error
info. pointer is reset and reused, which was the original goal.
[@nsAboutProtocolHandler:122]] This is a regression fix from the fix
for bug 163225. The assertions never made sense and now they are
exposed. r=bzbarsky, sr=darin
- make TLS intolerant server detection over proxies work
(this bug 87902)
- on connection failure, only retry without TLS when it is really
likely to help (bug 149910)
- remove obsolete workarounds in SSL i/o layer
(see removed comments in patch)
- avoid to confuse programmers reading code,
by renaming TLSStepUp (which means something else)
to the correct term STARTTLS (what the code is actually doing).
(As suggested by nelsonb)
- If an invalid or expired etc. server certificate is presented,
a warning is shown. If the user decides to cancel,
network activity should stop immediately.
(we currently warn multiple times) (bug 87209)
r=javi/darin/ducarroz/dmose sr=alecf
Added two new libssldap public functions: ldapssl_set_option() and
ldapssl_get_option().
Also fixed a bug in ldapsinit:do_ldapssl_connect() that sometimes
caused PR_Close() to be called twice on an SSL file descriptor
if an error occurred (once in do_ldapssl_connect() itself and
once in the libprldap close function that is called from
do_ldapssl_connect()).
Also updated the NSPR and NSS "error code to string" mapping
tables that are used by ldapssl_err2string().
Also fixed a bug in common.c:ldaptool_print_lderror (LDAP command
line tools) where we did not check for SSL errors when the
error code was "can't connect." We were only checking on
"server down" errors.
Changes:
* Added EXTRA_COMPONENTS variable to build. Use it to automagically install files into the components dir.
* Added SYSINSTALL variable. Use it install files onto the system.
* Move mozilla-config generation from configure to makefiles
* mozilla-config will now generate paths based upon the MRE-based directory layout.
* Generate mozilla script from template
The installation heirarchy currently looks like:
$prefix -> /bin/mozilla
/bin/mozilla-config
/include/mozilla-ver/*
/lib/mozilla-ver (mozappdir)
/components
/res
/defaults
/chrome
There is preliminary build support for MRE dirs as well though the component manager hasn't had MRE support added yet.
$prefix -> /lib/mre/mre-ver (mredir)
/components
Fix bug # 121276 - use sigthreadmask instead of sigprocmask (AIX).
Fix bug # 81584 - strtok use and STRTOK on Linux.
The charray() code now uses strtok_r() if available.
strtok_r() is now used everywhere on Linux if available.
New libssldap public function: ldapssl_set_strength().
New command line tool option: -3 ("no third").
Fix bug # 134035 - LDAP command line tools don't report vendor mismatch.
Improved error checks and reporting for SSL errors in tools.
ldap_url_search() now uses the port from the LDAP * if no hostport
is included in the URL. This is a change in behavior, but the old
behavior was wrong.
Removing support for <outliner> tags
Removing <tree> layout code and moving <tree> tags to outliner layout
Convert all usage of <outliner> to tree tags
Convert all usage of <tree> tags to new <tree> syntax or <listbox>
r=cmanske,varga sr=hyatt,sspitzer a=asa
Also includes fixes by Jan Varga (varga@utcruk.sk) for bugs 132020, 133451, 131393, 115894, and 129327
nsIComponentRegistrar.
2. Converts callers of nsComponentManager::AutoRegister to use
nsIComponentRegistrar's autoRegistrar method.
3. Add nsIComponentRegistrar implmentation to nsComponentManagerImpl.
4. Rearrange nsComponentManager.cpp so that related methods are in the same
place.
5. Added a C-style function NS_GetComponentRegistrar so that getting the
registrar is easier in some places.
6. Added a nsISimpleEnumerator interface on PLDHashTableEnumeratorImpl. in
this way, the same base class can support both old style and new style
enumerations.
7. Fixed a nasty bug where unregistring factories will leave the contract id
hash with a dangling pointer. Now, when unregister is called we search the
contract id hash for entries which have the given doomned cid and remove them.
Bug 115853. r=dp@netscape.com, sr=rpotts@netscape.com
94653 - can't right align text in outliner cell
95730 - White block displays/artifacts in the highlight area after Move/Drag mail to folder.
112238 - [RFE] Outliner widget should support line-through
116855 - Modify outliner builder syntax to match outliner content model
118093 - Make -moz-image-region work in the outliner
118113 - Too little space between pref category tree and pref page
118154 - Category now has a column picker
118647 - Memory leak of 52 bytes from 1 block allocated in nsOutlinerBodyFrame::CheckVerticalOverflow
118660 - left arrow no longer moves to parent category
119751 - Branches are open by default and then closed after launch.
- outermost <outlinerchildren> has "-moz-box-flex: 1" by default
and it actaully replaced <outlinerbody>
- all event handlers moved on <outliner>
- content builder, outliner builder and content view use <outliner> as root
element
- nsIBoxObject has a new attribute "element"
- all CSS rules outlinerbody:foo replaced with outlinerchildren:foo
- drop feedback is now skinable
r=waterson for the changes in mozilla/content
r=sspitzer on the mailnews parts
r=blake on the bookmarks/history
r=bryner on the OutlinerBodyFrame and OutlinerContentView changes
sr=hyatt,hewitt
a function prototype. So we have to decare the function and then
declare it a friend. When Default arguments are used, they need to
be specified in the function prototype and NOT in the friend declaration.
# 43311
r=dbaron@fas.harvard.edu
