aboutsummaryrefslogtreecommitdiff
path: root/src/http.h (unfollow)
AgeCommit message (Collapse)Author
2006-11-02Added the SBuffer class, a memory buffer that can be used like a stream. Yay!Mike Buland
2006-10-31Update to ConnectionManager, things are much, much better now.Mike Buland
2006-10-31Expanded the scope of the stream system to include positional functions. UpdatedMike Buland
the Connection class so that it won't die horribly if you don't provide the pointers to updatable memory for the amount of time not spent waiting for data during a read. Also fiddled with the http test, as you can see...nothing important.
2006-10-31Changed the file class to be SFile so that it will collide less, all of themMike Buland
will be named like that, make life easier.
2006-10-31Added an sha1 class...Mike Buland
2006-10-31Added the new linkmessenger class that will act as a base-class for anythingMike Buland
that wants to send messages to a containing programlink. Also fiddled with other things...aparently.
2006-10-27Added the beginings of the new stream system.Mike Buland
2006-10-19The plugger now cleans up after you if you're a slob.Mike Buland
2006-10-18Updated it, now it doesn't delete the protocol before telling people about itMike Buland
when shutting down...
2006-10-13Connections are now kept until all pending output has been sent, even if theMike Buland
disconnected flag has been set.
2006-10-11Rearranged the tests, now it's like the old style, which I like more for someMike Buland
reason.
2006-10-11Despite some svn oddness, I'm now moving to a new setup for the tests, that'sMike Buland
very much like the original one, but now using build. You will need the latest build in order to build the tests.
2006-10-11Fixed a wierd bug that caused each element of any part of a copied node tree toMike Buland
be partially cross-linked to the node that was the root of the copy. Also fixed a bug in the xmlfilewriter that should make it both faster, and not mess up % style replacements.
2006-10-11Bugfixes, and pymake removal.Mike Buland
2006-09-26Fixed some whackiness in the connection class, -1 from readInput means nothingMike Buland
there, but there might be later. 0 means death.
2006-09-22Fixed something else in the serializer.Mike Buland
2006-09-22I beleive I fixed the inconsistancies for 64bit machines.Mike Buland
2006-09-21Updated the build.conf to include the tests, not the unit tests yet, but theseMike Buland
I need more anyway.
2006-09-20new build.conf fileMike Buland
2006-09-20Updated the connection class to send all outgoing data when there is some.Mike Buland
Later this should be fixed and another function should be added.
2006-08-30david - creepy forgot to commit hashtable stuff... also fixed something in ↵David
md5... it was using uninitialized data...
2006-08-25Fixed more stuff in the connection class. Yay.Mike Buland
2006-08-25Fixed the connection manager to not suck with the new connection read api.Mike Buland
2006-08-25Added tests and exception codes, so you're program can tell just how bad thingsMike Buland
really are.
2006-08-25Added a new exception for connecitons, and added a new function to theMike Buland
connection class, that reads at least n bytes in x seconds.
2006-08-22Added a new helper to the plugger.Mike Buland
2006-08-22Fixed creepy conflict, two extra spaces in one version...sheeshMike Buland
2006-08-09Fixed a bug in the config manager base that caused config files to never loadMike Buland
or work.
2006-08-09Ok, this one actually builds again.Mike Buland
2006-08-09Renamed it, it also is no longer a singleton, although your child class couldMike Buland
be.
2006-08-09This may work, now the filenames just need to be changed.Mike Buland
2006-08-09The start of the configmanager to libbu++, they are good, but make congo notMike Buland
build now, for a while, and don't actually work yet, so just don't use this revision of libbu++.
2006-08-07Updated the build.conf file, you can build the tests, and more.Mike Buland
2006-08-06It builds with build again, and even the tests now. woo!Mike Buland
2006-08-04Help triggers an exit now in the param processor.Mike Buland