Confirmed users
574
edits
(→DXR) |
|||
| Line 29: | Line 29: | ||
* Got LLVM 3.2 working on Jenkins. | * Got LLVM 3.2 working on Jenkins. | ||
* Replaced our weird, subprocess-spawning, socket-opening testrunner with nose. (Yay for less code.) | * Replaced our weird, subprocess-spawning, socket-opening testrunner with nose. (Yay for less code.) | ||
* Merged lots of features | * Merged lots of features | ||
** distinguishing between const and non-const functions | |||
** controlling the build order of multiple trees, for when there are inter-tree dependencies | |||
** indexing that makes some queries go 30x faster | |||
** finding callers and callees | |||
** typedefs | |||
** absolute paths for ignores | |||
* Redid the server as a Flask app. CGI is gone in favor of WSGI. Now we have auto-reloading, REPL debugging, better pdb support, etc. | * Redid the server as a Flask app. CGI is gone in favor of WSGI. Now we have auto-reloading, REPL debugging, better pdb support, etc. | ||
* Greatly reduced the number of files spit out by a DXR build. Where before, DXR would emit large "target" folders full of Python source, templates, executables, and the index data, it now spits out just the index and a config file. Since we're no longer generating copies of everything, development is simpler, as we can just edit the originals that are under version control. | * Greatly reduced the number of files spit out by a DXR build. Where before, DXR would emit large "target" folders full of Python source, templates, executables, and the index data, it now spits out just the index and a config file. Since we're no longer generating copies of everything, development is simpler, as we can just edit the originals that are under version control. | ||