It used the same package name as DocumentLoader and the same .apk name as the
eary sc cppunit test app. Probably having two unrelated apps with the same
package name causes some confusion somewhere.
Change-Id: I11414b9cd59694eb97d39bfaeac4ed1066ae3aab
Only renders on very-first-start after install (oddly).
We initialize vcl in it's own thread to avoid problems.
Thanks to tml for fixing a linking issue.
Change-Id: I960d11c6098681356fea0634970545aa9af9bacb
This reverts commit 3c71fa1e5e. There are cases
(like failure to install a per-user extension during upgrade), where the user
may want to still see the broken extension in the Extension Manager, instead of
it "disappearing completely." A better fix would be to roll back registration
of the broken extension completely, so that "unopkg list" would not list the
first half of its parts as "is registered: yes" and the second half as "is
registered: no."
I doubt any other app would intentionally do such a restart trick, so
it isn't surprising that the app sandbox designers have not taken such
a possibility into account.
So instead of crashing, display a dialog informing the user that LO
must be restarted manually once after installation (or after being
updated, presumably).
Change-Id: I0e6ea8384fe5a9b3a8b3ddb07743374a4c2300cd
Such failure happens e.g. for components that bring along duplicate UNO
implementation names, and if they ever got added to an rdb, LO would afterwards
refuse to start forever.
Change-Id: Ia62ac496547a45f5a72fdc81543f15a417c09687
In a sandboxed LibreOffice, on 10.8.2 at least, creating the Unix
domain socket fails. Ignore that as hopefully people running a
sandboxed LO won't attempt starting it from the command-line or in
other tricky ways, so the normal OS X mechanism that prevents multiple
instances of an app from being started should work fine. I hope.
Change-Id: I4e0aa0c823328d00c8ff1c039859e78ba8ae92c5
...to avoid missing migrations on Windows and Mac OS X, where existing "3" user
profiles got the "MIGRATED" flag file erroneously added when running LO 3
already, presumably due to 6b52267337 "default
config location has changed, look in old config dir when migrating."
Change-Id: I8dd50a36450c167d47411834cc06a754affadf6d
... to look pretty; most of these expect PRODUCTEXTENSION to be a word
like "Beta" but nowadays it is ".micro.rc.alphabeta" numbers so there
should be no space before it.
ReplaceStringHookProc adds insult to injury by refusing to
replace %ABOUTBOX variables unless some %PRODUCT is in the string.
Change-Id: Ib676f3837ab0785529a226fc9359e8defacf3459
daeed90f45 "re-base on ALv2 code" had incorporated
from <http://svn.apache.org/viewvc?view=revision&revision=1172102> "imported
patch extensions_i117681.patch" the parts using BAK_EXTENSIONS in
desktop/source/deployment/manager/, but not the parts setting BAK_EXTENSIONS in
scp2/source/ooo/.
Also, setting the stamp for "bak" in PackageManagerImpl::create needed to be
adapted to bee742eb7a "Fix sense of r/o detection
code, and clean up" and 61eeb689d7 "fix smoketest
- need to check read-only-ness of non-existent paths."
Change-Id: If7cc363b3cf21fd13fb467f5f672eda5312515c8
"The original code is broken to begin with, as it ignores the return values from
the std::unique calls, so excess elements remain at the end of va and vb."
(<http://lists.freedesktop.org/archives/libreoffice/2013-January/043552.html>
"[PATCH] Simplify MigrationImpl::subtract in desktop;" see that mail thread also
for the inspiration to use std::set_difference.)
This fix is not very relevant though, as there are no "ExcludedFiles" lists in
officecfg/registry/data/org/openoffice/Setup.xcu, so the second argument is
always empty, so the return value is always a (sorted) copy of the first
argument, and the "IncludedFiles" lists in Setup.xcu produce no duplicates, so
std::unique does not shrink the first argument anyway.
Change-Id: Ie9fb64ee40fef6e7bdf0f5d0eca5717fec7b0d50
Ignore this option on Windows.
This work-arounds a regression introduced by f4a4ba9ac
Windows appends -Embedding when LibreOffice is called by COM+
Before f4a4ba9ac this option was silently ignored.
Change-Id: I8a727895c9ec8fa4b06aea8ad66b2791ada66dd5
Reviewed-on: https://gerrit.libreoffice.org/1514
Reviewed-by: Michael Stahl <mstahl@redhat.com>
Tested-by: Michael Stahl <mstahl@redhat.com>
