The form.get function is something like this:
const std::string& get(const std::string& abc) { return abc; }
passing a string literal implicitly gets converted to temporary
std::string whose reference is then returned and used. This causes
crash, atleast for me, on building online with GCC 7
Change-Id: I09d0aeea57a3dbeeefd1bb28ff645723714aa6b4
Reviewed-on: https://gerrit.libreoffice.org/46727
Reviewed-by: Jan Holesovsky <kendy@collabora.com>
Tested-by: Jan Holesovsky <kendy@collabora.com>
(cherry picked from commit 75c2147b7f774afccf55bb93f7fae79efb615361)
Reviewed-on: https://gerrit.libreoffice.org/46730
Reviewed-by: pranavk <pranavk@collabora.co.uk>
Tested-by: pranavk <pranavk@collabora.co.uk>
use a specific message from the client for set the visibility state of
a group instead of hijacking the update row/column header message
Change-Id: I69d66b30db0b4d8a0082cbd2524120491d4f97cb
Reviewed-on: https://gerrit.libreoffice.org/45446
Reviewed-by: Marco Cecchetti <mrcekets@gmail.com>
Tested-by: Marco Cecchetti <mrcekets@gmail.com>
Drop the mention of /opt/poco as the place where to put a self-built
Poco as that seems to make some people think /opt would be a
recommended place to do a clone of the LibreOffice repository, too.
Which is a weird place for it.
Change-Id: I17543a915a7f9ea4716379283b0a51723d39b376
When the name of a document to be edited ends with an extension
matching that mentioned in a prefilter plug-in's configuration file,
the command specified is run and it is the output file that is
actually edited.
Look in the directory LOOLWSD_CONFIGDIR/conf.d (or pathname passed
with --config-dir option) for extra .xml files containing a Poco
configuration tree. These are not merged with the actual configuration
but kept separate. This commit does not yet actually use them for
anything.
To be able to set the support key directly from the command line, and to show
the option, etc.
Change-Id: Iac93bc47a6f4b9d5a5ad0ac8b06bda978e01b760
Reviewed-on: https://gerrit.libreoffice.org/43098
Reviewed-by: Michael Meeks <michael.meeks@collabora.com>
Tested-by: Michael Meeks <michael.meeks@collabora.com>
Assume that when SLEEPFORDEBUGGER or SLEEPKITFORDEBUGGER is mentioned,
loolwsd is being debugged. In that case, don't set any timeout to exit
before first child is forked.
Change-Id: I2527f02187d8452ef6bebb70e1d750b5e796c4b6
The routine for handling the configuration for the max file size
limit, was wrongly using NOFILE. Now we handle both limits correctly.
Change-Id: Ie8b63617286f66af6d4eb1b35b9e4f4b28f3c2a6
Reviewed-on: https://gerrit.libreoffice.org/42803
Reviewed-by: Jan Holesovsky <kendy@collabora.com>
Tested-by: Jan Holesovsky <kendy@collabora.com>
Reviewed-on: https://gerrit.libreoffice.org/42811
Reviewed-by: Marco Cecchetti <mrcekets@gmail.com>
Tested-by: Marco Cecchetti <mrcekets@gmail.com>
... instead of handing it over the Document broker polling thread which
can lead to race conditions, and hence not giving desired behavior when
document is changed externally.
Change-Id: Ib0821d4ae931c357bc4d4c526865eefc090ddc23
Two problem types:
- non-final class has virtual functions but no virtual dtor -> mark the
class final
- abstract class has no virtual dtor -> add a virtual dtor
Change-Id: Iae208b65c774e6da7a3dda5e725fe07d4d589e4f
