summaryrefslogtreecommitdiff
diff options
context:
space:
mode:
authorlouiz’ <louiz@louiz.org>2018-04-29 22:18:26 +0200
committerlouiz’ <louiz@louiz.org>2018-04-29 22:18:26 +0200
commitb0168fd45b3683c2d6f61ccae67dcd5b918a363d (patch)
tree6b3e035f6edfef0b9a8aac1813259b9c95ebcf35
parent0887be6eef24929fd594239e9569fc5cee54aba0 (diff)
downloadbiboumi-b0168fd45b3683c2d6f61ccae67dcd5b918a363d.tar.gz
biboumi-b0168fd45b3683c2d6f61ccae67dcd5b918a363d.tar.bz2
biboumi-b0168fd45b3683c2d6f61ccae67dcd5b918a363d.tar.xz
biboumi-b0168fd45b3683c2d6f61ccae67dcd5b918a363d.zip
mam: Send “fin complete” only when appropriate
Also simplify how we did the whole “limit + 1” And fix one bad interpretation of the XEP for the case where the query has no after or before restriction. fix #3349
-rw-r--r--src/bridge/bridge.cpp3
-rw-r--r--src/database/database.cpp26
-rw-r--r--src/database/database.hpp4
-rw-r--r--src/xmpp/biboumi_component.cpp22
-rw-r--r--tests/end_to_end/__main__.py17
5 files changed, 37 insertions, 35 deletions
diff --git a/src/bridge/bridge.cpp b/src/bridge/bridge.cpp
index 1c646fe..7a0157a 100644
--- a/src/bridge/bridge.cpp
+++ b/src/bridge/bridge.cpp
@@ -1005,7 +1005,8 @@ void Bridge::send_room_history(const std::string& hostname, std::string chan_nam
auto limit = coptions.col<Database::MaxHistoryLength>();
if (history_limit.stanzas >= 0 && history_limit.stanzas < limit)
limit = history_limit.stanzas;
- const auto lines = Database::get_muc_logs(this->user_jid, chan_name, hostname, limit, history_limit.since, {}, Id::unset_value, Database::Paging::last);
+ const auto result = Database::get_muc_logs(this->user_jid, chan_name, hostname, limit, history_limit.since, {}, Id::unset_value, Database::Paging::last);
+ const auto& lines = std::get<1>(result);
chan_name.append(utils::empty_if_fixed_server("%" + hostname));
for (const auto& line: lines)
{
diff --git a/src/database/database.cpp b/src/database/database.cpp
index 3b3e890..6e08ee1 100644
--- a/src/database/database.cpp
+++ b/src/database/database.cpp
@@ -190,12 +190,9 @@ std::string Database::store_muc_message(const std::string& owner, const std::str
return uuid;
}
-std::vector<Database::MucLogLine> Database::get_muc_logs(const std::string& owner, const std::string& chan_name, const std::string& server,
- int limit, const std::string& start, const std::string& end, const Id::real_type reference_record_id, Database::Paging paging)
+std::tuple<bool, std::vector<Database::MucLogLine>> Database::get_muc_logs(const std::string& owner, const std::string& chan_name, const std::string& server,
+ std::size_t limit, const std::string& start, const std::string& end, const Id::real_type reference_record_id, Database::Paging paging)
{
- if (limit == 0)
- return {};
-
auto request = select(Database::muc_log_lines);
request.where() << Database::Owner{} << "=" << owner << \
" and " << Database::IrcChanName{} << "=" << chan_name << \
@@ -228,15 +225,26 @@ std::vector<Database::MucLogLine> Database::get_muc_logs(const std::string& owne
else
request.order_by() << Id{} << " DESC ";
- if (limit >= 0)
- request.limit() << limit;
+ // Just a simple trick: to know whether we got the totality of the
+ // possible results matching this query (except for the limit), we just
+ // ask one more element. If we get that additional element, this means
+ // we don’t have everything. And then we just discard it. If we don’t
+ // have more, this means we have everything.
+ request.limit() << limit + 1;
auto result = request.execute(*Database::db);
+ bool complete = true;
+
+ if (result.size() == limit + 1)
+ {
+ complete = false;
+ result.erase(std::prev(result.end()));
+ }
if (paging == Database::Paging::first)
- return result;
+ return {complete, result};
else
- return {result.crbegin(), result.crend()};
+ return {complete, {result.crbegin(), result.crend()}};
}
Database::MucLogLine Database::get_muc_log(const std::string& owner, const std::string& chan_name, const std::string& server,
diff --git a/src/database/database.hpp b/src/database/database.hpp
index 30ffcb4..3e25b30 100644
--- a/src/database/database.hpp
+++ b/src/database/database.hpp
@@ -131,8 +131,8 @@ class Database
* Get all the lines between (optional) start and end dates, with a (optional) limit.
* If after_id is set, only the records after it will be returned.
*/
- static std::vector<MucLogLine> get_muc_logs(const std::string& owner, const std::string& chan_name, const std::string& server,
- int limit=-1, const std::string& start="", const std::string& end="",
+ static std::tuple<bool, std::vector<MucLogLine>> get_muc_logs(const std::string& owner, const std::string& chan_name, const std::string& server,
+ std::size_t limit, const std::string& start="", const std::string& end="",
const Id::real_type reference_record_id=Id::unset_value, Paging=Paging::first);
/**
diff --git a/src/xmpp/biboumi_component.cpp b/src/xmpp/biboumi_component.cpp
index 6dc5fc5..dbaf8a4 100644
--- a/src/xmpp/biboumi_component.cpp
+++ b/src/xmpp/biboumi_component.cpp
@@ -743,19 +743,15 @@ bool BiboumiComponent::handle_mam_request(const Stanza& stanza)
}
// Do not send more than 100 messages, even if the client asked for more,
// or if it didn’t specify any limit.
- // 101 is just a trick to know if there are more available messages.
- // If our query returns 101 message, we know it’s incomplete, but we
- // still send only 100
- if ((limit == -1 && start.empty() && end.empty())
- || limit > 100)
- limit = 101;
- auto lines = Database::get_muc_logs(from.bare(), iid.get_local(), iid.get_server(), limit, start, end, reference_record_id, paging_order);
- bool complete = true;
- if (lines.size() > 100)
- {
- complete = false;
- lines.erase(lines.begin(), std::prev(lines.end(), 100));
- }
+ if (limit < 0 || limit > 100)
+ limit = 100;
+ auto result = Database::get_muc_logs(from.bare(), iid.get_local(), iid.get_server(),
+ limit,
+ start, end,
+ reference_record_id, paging_order);
+ bool complete = std::get<bool>(result);
+ auto& lines = std::get<1>(result);
+
for (const Database::MucLogLine& line: lines)
{
if (!line.col<Database::Nick>().empty())
diff --git a/tests/end_to_end/__main__.py b/tests/end_to_end/__main__.py
index 2f6fc28..1ec5a74 100644
--- a/tests/end_to_end/__main__.py
+++ b/tests/end_to_end/__main__.py
@@ -1961,7 +1961,7 @@ if __name__ == '__main__':
("/iq[@type='result'][@id='id3'][@from='#foo%{irc_server_one}'][@to='{jid_one}/{resource_one}']",
"/iq/mam:fin[@complete='true']/rsm:set")),
- # Retrieve a limited archive
+ # Retrieve the whole archive, but limit the response to one elemet
partial(send_stanza, "<iq to='#foo%{irc_server_one}' from='{jid_one}/{resource_one}' type='set' id='id4'><query xmlns='urn:xmpp:mam:2' queryid='qid4'><set xmlns='http://jabber.org/protocol/rsm'><max>1</max></set></query></iq>"),
partial(expect_stanza,
@@ -1971,7 +1971,7 @@ if __name__ == '__main__':
partial(expect_stanza,
("/iq[@type='result'][@id='id4'][@from='#foo%{irc_server_one}'][@to='{jid_one}/{resource_one}']",
- "/iq/mam:fin[@complete='true']/rsm:set")),
+ "!/iq/mam:fin[@complete='true']/rsm:set")),
]),
Scenario("mam_with_timestamps",
@@ -2205,11 +2205,9 @@ if __name__ == '__main__':
] * 150 + [
# Retrieve the archive, without any restriction
partial(send_stanza, "<iq to='#foo%{irc_server_one}' from='{jid_one}/{resource_one}' type='set' id='id1'><query xmlns='urn:xmpp:mam:2' queryid='qid1' /></iq>"),
- # Since we should only receive the last 100 messages from the archive,
- # it should start with message "1"
partial(expect_stanza,
("/message/mam:result[@queryid='qid1']/forward:forwarded/delay:delay",
- "/message/mam:result[@queryid='qid1']/forward:forwarded/client:message[@from='#foo%{irc_server_one}/{nick_one}'][@type='groupchat']/client:body[text()='1']")
+ "/message/mam:result[@queryid='qid1']/forward:forwarded/client:message[@from='#foo%{irc_server_one}/{nick_one}'][@type='groupchat']/client:body[text()='0']")
),
] + [
# followed by 98 more messages
@@ -2221,7 +2219,7 @@ if __name__ == '__main__':
# and finally the message "99"
partial(expect_stanza,
("/message/mam:result[@queryid='qid1']/forward:forwarded/delay:delay",
- "/message/mam:result[@queryid='qid1']/forward:forwarded/client:message[@from='#foo%{irc_server_one}/{nick_one}'][@type='groupchat']/client:body[text()='100']"),
+ "/message/mam:result[@queryid='qid1']/forward:forwarded/client:message[@from='#foo%{irc_server_one}/{nick_one}'][@type='groupchat']/client:body[text()='99']"),
after = partial(save_value, "last_uuid", partial(extract_attribute, "/message/mam:result", "id"))
),
# And it should not be marked as complete
@@ -2236,9 +2234,9 @@ if __name__ == '__main__':
partial(expect_stanza,
("/message/mam:result[@queryid='qid2']/forward:forwarded/delay:delay",
- "/message/mam:result[@queryid='qid2']/forward:forwarded/client:message[@from='#foo%{irc_server_one}/{nick_one}'][@type='groupchat']/client:body[text()='101']")
+ "/message/mam:result[@queryid='qid2']/forward:forwarded/client:message[@from='#foo%{irc_server_one}/{nick_one}'][@type='groupchat']/client:body[text()='100']")
),
- ] + 47 * [
+ ] + 48 * [
partial(expect_stanza,
("/message/mam:result[@queryid='qid2']/forward:forwarded/delay:delay",
"/message/mam:result[@queryid='qid2']/forward:forwarded/client:message[@from='#foo%{irc_server_one}/{nick_one}'][@type='groupchat']/client:body")
@@ -2297,8 +2295,7 @@ if __name__ == '__main__':
partial(expect_stanza,
("/iq[@type='result'][@id='id4'][@from='#foo%{irc_server_one}'][@to='{jid_one}/{resource_one}']",
"/iq/mam:fin/rsm:set/rsm:last[text()='{last_uuid}']",
- "/iq/mam:fin[@complete='true']",
- "/iq/mam:fin")),
+ "!/iq/mam:fin[@complete='true']",)),
]),
Scenario("channel_history_on_fixed_server",
[