summaryrefslogtreecommitdiff
path: root/src/include/buf.i
Commit message (Collapse)AuthorAgeFilesLines
* whitespaceKeith Bostic2015-01-051-1/+1
|
* Cache a maximum of 2MB per session for scratch buffers (configurable via an ↵Michael Cahill2015-01-051-1/+8
| | | | | | undocumented "session_scratch_max" config). SERVER-16623
* Copyright notices: add MongoDB, update to 2015.Keith Bostic2015-01-041-0/+1
|
* spell, KNF.Keith Bostic2014-05-231-1/+1
|
* Change WT_METADATA_URI to "metadata:".Michael Cahill2014-05-231-0/+10
|
* Change __wt_buf_grrow_worker to copy non-local data into place, thisKeith Bostic2014-04-291-3/+5
| | | | | | | | simplifies some other code in the system. Change buffer init calls to clear the buffer data size before calling the grow function -- it shouldn't be necessary, but it clarifies what is happening.
* Change the block manager to not touch the WT_ITEM.{mem,memsize} fieldsKeith Bostic2014-04-281-11/+12
| | | | | | | | | | | | when mapping object into memory, remove the WT_ITEM_MAPPED flag, that information is now based on WT_DATA_IN_ITEM (if the WT_ITEM.data field points into the WT_ITEM.mem buffer). This also simplifies some of the buffer handling, there's no longer any reason to "clear" WT_ITEM.mem for mapped objects. Some calls to __wt_buf_steal() used scratch buffers, which is dangerous, scratch buffers can include large allocated chunks; remove/replace all calls to __wt_buf_steal() with __wt_strndup() calls.
* Inline the buffer-size check parts of the standard scratch bufferKeith Bostic2014-04-251-0/+120
functions in a new inline file, include/buf.i. Remove a couple of inlined buffer size tests, they're no longer needed. Fix a few places to use __wt_buf_extend instead of rolling their own versions.