Age | Commit message (Collapse) | Author |
|
An unofficial HTTP response status code is defined and used
internally in LibreOffice WebDAV ucp provider in order to
catch either the DAVException::DAV_HTTP_TIMEOUT state or the
DAVException::DAV_HTTP_CONNECT state.
Change-Id: Iec4927d18ace5384fed16a7ec2a2620dc8305c5b
Reviewed-on: https://gerrit.libreoffice.org/29757
Tested-by: Jenkins <ci@libreoffice.org>
Reviewed-by: Giuseppe Castagno <giuseppe.castagno@acca-esse.eu>
|
|
Change-Id: I08c55a3023ec2e8990098eeb60e91cd18556e7ae
Reviewed-on: https://gerrit.libreoffice.org/29656
Reviewed-by: Noel Grandin <noel.grandin@collabora.co.uk>
Tested-by: Noel Grandin <noel.grandin@collabora.co.uk>
|
|
Add the OPTIONS information and response status code into the cache:
- if the OPTIONS information is already cached, update the cache only
if the lifetime is different;
- if the OPTIONS information is not cached, cache it.
Add some new functions in DAVOptions to support the change and remove
a function no longer used.
Change-Id: I2f28f0ee793ec7d898caa61cc0a4962334e6e068
Reviewed-on: https://gerrit.libreoffice.org/29733
Tested-by: Jenkins <ci@libreoffice.org>
Reviewed-by: Giuseppe Castagno <giuseppe.castagno@acca-esse.eu>
|
|
A reference can be found here:
<https://en.wikipedia.org/wiki/List_of_HTTP_status_codes>
(retrieved 2016-09-13).
Changes done:
Add set of 'HEAD method not available' before using fall back GET method.
Add new method in OPTIONS cache.
Add response status code if fall-back GET didn't succeeded.
Add copy-assignement operator to DAVOptions.
Fix behaviour of GET fall back when HEAD missing.
Change-Id: I6aad0e22bb444abf37b90186588f64f79b03cc3d
Reviewed-on: https://gerrit.libreoffice.org/29680
Tested-by: Jenkins <ci@libreoffice.org>
Reviewed-by: Giuseppe Castagno <giuseppe.castagno@acca-esse.eu>
|
|
Instead of caching only a single status flag, now both the HTTP
response status code and the message accompanying it are cached.
Change-Id: If7352f6b0cb7c7dab6af3cede96647308baa5ce2
Reviewed-on: https://gerrit.libreoffice.org/29538
Tested-by: Jenkins <ci@libreoffice.org>
Reviewed-by: Giuseppe Castagno <giuseppe.castagno@acca-esse.eu>
|
|
Implement a better way to GET only headers, as HEAD method
substitute.
Wrote a new version of GET, specifically for 'no payload' service
because sometimes neon was returning an internal error about:
'Compressed response was truncated' but without a specific error
state.
Change-Id: I84a3e5635193c3f68d335eba7af41b05980f4192
Reviewed-on: https://gerrit.libreoffice.org/29502
Tested-by: Jenkins <ci@libreoffice.org>
Reviewed-by: Giuseppe Castagno <giuseppe.castagno@acca-esse.eu>
|
|
Change-Id: I152df55f5a30e073c66cfd3a64d5090a20cb232d
Reviewed-on: https://gerrit.libreoffice.org/29471
Tested-by: Jenkins <ci@libreoffice.org>
Reviewed-by: Noel Grandin <noel.grandin@collabora.co.uk>
|
|
The function call will act as a precheck on Web resource access
error when accessing the open directly without the normal file
open procedure, e.g. accessing Web resource property values first.
An example of this direct open call are the extension fetching and
the extension update check.
Change-Id: I765a2fe29f432a07a649e5a35440ae351c681af9
Reviewed-on: https://gerrit.libreoffice.org/29454
Tested-by: Jenkins <ci@libreoffice.org>
Reviewed-by: Giuseppe Castagno <giuseppe.castagno@acca-esse.eu>
|
|
Change-Id: I6aeda978911b25caa45d4e459e581fc743e93d2c
Reviewed-on: https://gerrit.libreoffice.org/29442
Tested-by: Jenkins <ci@libreoffice.org>
Reviewed-by: Noel Grandin <noel.grandin@collabora.co.uk>
|
|
Change-Id: I6be4a7861f2a978c260defd54dcbc8d124017439
Reviewed-on: https://gerrit.libreoffice.org/29431
Tested-by: Jenkins <ci@libreoffice.org>
Reviewed-by: Giuseppe Castagno <giuseppe.castagno@acca-esse.eu>
|
|
Change-Id: Ia29560d54a61f1238f3b4e945d78308a3a68c483
Reviewed-on: https://gerrit.libreoffice.org/29269
Tested-by: Jenkins <ci@libreoffice.org>
Reviewed-by: Giuseppe Castagno <giuseppe.castagno@acca-esse.eu>
|
|
Change-Id: I41e7cc0a1e176e44cb1b3fb87269eb6e052c6478
Reviewed-on: https://gerrit.libreoffice.org/29273
Tested-by: Jenkins <ci@libreoffice.org>
Reviewed-by: Giuseppe Castagno <giuseppe.castagno@acca-esse.eu>
|
|
Change-Id: I38333e5d229aa520fbe0a8ad72007c503853956e
|
|
Change-Id: I4ff36df3ac1cc0788322768378fb74f70892f922
Reviewed-on: https://gerrit.libreoffice.org/29245
Tested-by: Giuseppe Castagno <giuseppe.castagno@acca-esse.eu>
Reviewed-by: Giuseppe Castagno <giuseppe.castagno@acca-esse.eu>
|
|
Change-Id: I70fcf95dfd3db05b4fd6e5cee37866f673d3afa8
Reviewed-on: https://gerrit.libreoffice.org/29183
Reviewed-by: Noel Grandin <noel.grandin@collabora.co.uk>
Tested-by: Noel Grandin <noel.grandin@collabora.co.uk>
|
|
Change-Id: I852e98b16fdcb88b04e39d11e3101d502c918c24
Reviewed-on: https://gerrit.libreoffice.org/29078
Reviewed-by: Noel Grandin <noel.grandin@collabora.co.uk>
Tested-by: Noel Grandin <noel.grandin@collabora.co.uk>
|
|
... partial GET means a GET with no data returned, to test for
resource 'real' availability in case HEAD fails or is disabled.
At the same time disable use of HEAD method on the resource
for the whole access procedure.
This change is needed to manage web servers that don't permit the
use of either HEAD or OPTIONS methods.
To accomodate pure web sites that enable only GET to fetch contents.
Change-Id: If52316ddd6ce637fb7d2ef518423de246ee2dcfb
Reviewed-on: https://gerrit.libreoffice.org/28992
Tested-by: Giuseppe Castagno <giuseppe.castagno@acca-esse.eu>
Reviewed-by: Giuseppe Castagno <giuseppe.castagno@acca-esse.eu>
|
|
Some server have HEAD method disabled or not implemented.
So we can use GET with no data, retrieving only the header info.
The implementing code comes from LO WebDAV serf implementation.
Change-Id: I3b3d81f5a07232c4f1647888685e820939bc978c
Reviewed-on: https://gerrit.libreoffice.org/28982
Tested-by: Jenkins <ci@libreoffice.org>
Reviewed-by: Giuseppe Castagno <giuseppe.castagno@acca-esse.eu>
|
|
The issue of 362d4f0cd4e50111edfae9d30c90602c37ed65a2 "Explicitly mark
overriding destructors as 'virtual'" appears to no longer be a problem with
MSVC 2013.
(The little change in the rewriting code of compilerplugins/clang/override.cxx
was necessary to prevent an endless loop when adding "override" to
OOO_DLLPUBLIC_CHARTTOOLS virtual ~CloseableLifeTimeManager();
in chart2/source/inc/LifeTime.hxx, getting stuck in the leading
OOO_DLLPUBLIC_CHARTTOOLS macro. Can't remember what that
isAtEndOfImmediateMacroExpansion thing was originally necessary for, anyway.)
Change-Id: I534c634504d7216b9bb632c2775c04eaf27e927e
|
|
Surely the actual bitfield syntax is enough to tell the code reader
that it is a bitfield.
Change-Id: Ic9552e01b19c8b34b2a17db56b9ff63e7c7de926
|
|
Change-Id: Iea0eef5d3cd67c73325eb7af883ef8ce286187cd
Reviewed-on: https://gerrit.libreoffice.org/28850
Tested-by: Jenkins <ci@libreoffice.org>
Reviewed-by: Giuseppe Castagno <giuseppe.castagno@acca-esse.eu>
|
|
Change-Id: I1a7ed38a95b06029d821b000324b2cc880df1ebb
|
|
Change-Id: I592f73012ae8a4df8488ef04dc0956d9400625bc
|
|
Change-Id: Ib854d1ac74e952c786136176899e533804eb1e61
Reviewed-on: https://gerrit.libreoffice.org/28459
Tested-by: Jenkins <ci@libreoffice.org>
Reviewed-by: Giuseppe Castagno <giuseppe.castagno@acca-esse.eu>
|
|
Change-Id: Ifa588c37added377d753cb7b32a6f1e6a3672c94
Reviewed-on: https://gerrit.libreoffice.org/28453
Tested-by: Jenkins <ci@libreoffice.org>
Reviewed-by: Giuseppe Castagno <giuseppe.castagno@acca-esse.eu>
|
|
When OPTIONS methods (or verb) is disabled (or denied) on a IIS 10.0
web server, error 404 (e.g. 'Not Found') is emitted, so we need to deal
with it.
Change-Id: Ifd6df0ea4b89824133e212f73950d6c3acd00dd7
Reviewed-on: https://gerrit.libreoffice.org/28430
Tested-by: Jenkins <ci@libreoffice.org>
Reviewed-by: Giuseppe Castagno <giuseppe.castagno@acca-esse.eu>
|
|
...the one I wrongly removed in commit
72589082945e5a197f3fb4b884f48dd4ae275e5f
Change-Id: I9f8e2f34a91b0fe481b1fe0f4655e3899a0589a6
Reviewed-on: https://gerrit.libreoffice.org/28412
Tested-by: Jenkins <ci@libreoffice.org>
Reviewed-by: Giuseppe Castagno <giuseppe.castagno@acca-esse.eu>
|
|
This is a sort of follow up to commit:
510da29131e56ba0e1783e505e7131e6f7e0d6b4:
"forced to make certain neon api calls thread-safe due to cups
rhbz#637738 libgcrypt that's been initialized already by cups to be
non-threadsafe is therefore then also non-threadsafe. neon uses
libgcrypt. So in the (as this is the usual circumstance) case that
cups is initialized before neon, then libgcrypt is not threadsafe
even though neon attempt to initialize it as such."
Change-Id: I1f868642ac15bf9fa25d3edd46390467d8fce1f5
Reviewed-on: https://gerrit.libreoffice.org/28401
Tested-by: Jenkins <ci@libreoffice.org>
Reviewed-by: Giuseppe Castagno <giuseppe.castagno@acca-esse.eu>
|
|
Change-Id: I3e585fb44612cb7ab2e6f04570b233310b01d8c8
|
|
Change-Id: Ib7ba126d53e81dbd65451aa4bb33033f468ca295
|
|
Change-Id: I6f9e1218c2767ec69ca371c4f5a12c2f385c7363
Reviewed-on: https://gerrit.libreoffice.org/28275
Tested-by: Jenkins <ci@libreoffice.org>
Reviewed-by: Giuseppe Castagno <giuseppe.castagno@acca-esse.eu>
|
|
Change-Id: I47f583c808498aaf0f7cca1749fb5ef841ad10f3
Reviewed-on: https://gerrit.libreoffice.org/28274
Tested-by: Jenkins <ci@libreoffice.org>
Reviewed-by: Giuseppe Castagno <giuseppe.castagno@acca-esse.eu>
|
|
PROPFIND 'propname' is the special usage to retrieve all the
properties available on the URI resource, their names only.
See <https://tools.ietf.org/html/rfc4918#section-9.1> for
PROPFIND 'propname' definition.
Add cache usage in Content::getProperties as well.
The caching model is simple: a simple lifetime limit of 10 seconds
to declare the property name list stale and request another list,
accessing the Net.
This should reduce the number of PROPFIND calls on the Net.
Change-Id: Ie4ebd946dd81583dc964a62c7744f3e2c716c737
Reviewed-on: https://gerrit.libreoffice.org/28273
Tested-by: Jenkins <ci@libreoffice.org>
Reviewed-by: Giuseppe Castagno <giuseppe.castagno@acca-esse.eu>
|
|
...where their use of boost::shared_ptr (instead of std::shared_ptr) matches the
use in libcmis, but makes them unlikely to be useful anywhere else. So move
them into ucb/source/ucb/cmis/.
Change-Id: I68359be6b43d6889af4f241dcdcbdc0d9d70d717
|
|
Change-Id: If42393b2738ab107d380e55a0cbf2d6fa9940130
Reviewed-on: https://gerrit.libreoffice.org/28265
Reviewed-by: Giuseppe Castagno <giuseppe.castagno@acca-esse.eu>
Tested-by: Giuseppe Castagno <giuseppe.castagno@acca-esse.eu>
|
|
...by turning the relevant ctor parameters into "const &".
Change-Id: Ia8d0aba5da10ad6b25f8689e2281e45b3d71c1fc
|
|
Change-Id: I83818be02fd7479bf6990dde076aad582a7f8f44
|
|
Change-Id: I642be9296babf077588eaf4c9da0aa11dc184fa0
|
|
Change-Id: I5cd6e9f4f5b3d9d7086229e4753d443d0f82bed7
|
|
Change-Id: Id5323cb6f52666f85965e11b07e4f2bca8af4e78
|
|
... and a bit of function member signature change.
Change-Id: I0d41277bca3991c7c105400e932df96f7bcf0d79
Reviewed-on: https://gerrit.libreoffice.org/28172
Tested-by: Jenkins <ci@libreoffice.org>
Reviewed-by: Giuseppe Castagno <giuseppe.castagno@acca-esse.eu>
|
|
Plus another PUT forgotten on a previuos commit.
Change-Id: I9e15cae2e858884cfd4c827ae767a2c5fcef1782
Reviewed-on: https://gerrit.libreoffice.org/28158
Tested-by: Jenkins <ci@libreoffice.org>
Reviewed-by: Giuseppe Castagno <giuseppe.castagno@acca-esse.eu>
|
|
Change-Id: Id7ddc0fc1f57c5e8e7fb002e31d54fb8e9f8ffab
Reviewed-on: https://gerrit.libreoffice.org/28050
Tested-by: Jenkins <ci@libreoffice.org>
Reviewed-by: Noel Grandin <noelgrandin@gmail.com>
|
|
Change-Id: Ia9ef644359c34ac30b78d48c79e8bc9d5c639f92
Reviewed-on: https://gerrit.libreoffice.org/28111
Tested-by: Jenkins <ci@libreoffice.org>
Reviewed-by: Giuseppe Castagno <giuseppe.castagno@acca-esse.eu>
|
|
Change-Id: Id49254f3f3b230de4e7d3f116b52a38f6042f6c1
Reviewed-on: https://gerrit.libreoffice.org/28070
Tested-by: Jenkins <ci@libreoffice.org>
Reviewed-by: Giuseppe Castagno <giuseppe.castagno@acca-esse.eu>
|
|
Check for maximum number of redirections according to
<https://tools.ietf.org/html/rfc7231#section-6.4>.
A practical limit can be 5, due to old RFC:
<https://tools.ietf.org/html/rfc2068#section-10.3>, this limit is
reported also in more recent RFCs, see final paragraph of RFC7231, 6.4.
Change-Id: I2b394ef8d1ef391a527df349aa749819c496657b
Reviewed-on: https://gerrit.libreoffice.org/28066
Tested-by: Jenkins <ci@libreoffice.org>
Reviewed-by: Giuseppe Castagno <giuseppe.castagno@acca-esse.eu>
|
|
Change-Id: Id2c9f87e408b6754e17ef3792691d848deabfcd5
Reviewed-on: https://gerrit.libreoffice.org/28056
Tested-by: Jenkins <ci@libreoffice.org>
Reviewed-by: Giuseppe Castagno <giuseppe.castagno@acca-esse.eu>
|
|
Change-Id: Ib963fc6e0834606dcb6f42b3c0df5a408ebc5351
Reviewed-on: https://gerrit.libreoffice.org/28030
Tested-by: Jenkins <ci@libreoffice.org>
Reviewed-by: Giuseppe Castagno <giuseppe.castagno@acca-esse.eu>
|
|
Change-Id: Ie3d2cf29e99b8a51b80246aafa23a92e6c3404b2
Reviewed-on: https://gerrit.libreoffice.org/28017
Tested-by: Jenkins <ci@libreoffice.org>
Reviewed-by: Noel Grandin <noelgrandin@gmail.com>
|
|
for a given site.
This patch uses the fact that getFolderByServerRelativeUrl API returns
site's root folder if used with empty path ('').
Hopefully this isn't otherwise context-dependent, so that it doesn't
return other folders in some unknown circumstances.
More correct would be instead to use the site's root path directly,
like '/sites/mysite'. But I don't know a reliable way to filter the
API part from the base URI.
To make use of the corrected SharePoint repository root id, the patch
also redirects all cases where "/" object is queried, to session's
getRootFolder().
libcmis part of the patch will hopefully soon be merged upstream.
Change-Id: I0f3b9930748de3fe4763eb27721739f097b9892d
Reviewed-on: https://gerrit.libreoffice.org/27987
Tested-by: Jenkins <ci@libreoffice.org>
Reviewed-by: David Tardon <dtardon@redhat.com>
|