Age | Commit message (Collapse) | Author | |
---|---|---|---|
2010-11-14 | First hack at pruning tags to keep from growing. | Tom Rondeau | |
2010-11-14 | Fixing constructor interface to work better through python (int instead of ↵ | Tom Rondeau | |
uint64_t). | |||
2010-11-14 | Adds argument to annotator test blocks to determine when they add a new tag. | Tom Rondeau | |
2010-11-14 | Fixing some bugs. | Tom Rondeau | |
2010-11-13 | More test methods. Better testing of the all-to-all policy and more tests ↵ | Tom Rondeau | |
for the 1-to-1. | |||
2010-11-13 | Block executor propagation method returns an error indicator to stop the ↵ | Tom Rondeau | |
flowgraph in an error instead of throwing. | |||
2010-11-13 | Fixing how 1-to-1 annotator handles inputs to better test propagataion method. | Tom Rondeau | |
2010-11-13 | Renamed annotator block that goes all-to-all now so that it can take in ↵ | Tom Rondeau | |
arbitrary number of inputs and pass tags on that way. | |||
2010-11-11 | Reworked 1-to-1 annotator block to better testing purposes. | Tom Rondeau | |
2010-11-11 | Switches to using the new annotator block with a known graph structure to ↵ | Tom Rondeau | |
run tests against the all-to-all tag propagation policy to make sure the tags are flowing downstream properly. Adding stub for testing the 1-to-1 tag propagation policy. | |||
2010-11-11 | Modifying the all-to-all stream annotator for better use in testing. It now ↵ | Tom Rondeau | |
add tags at pre-determined points in the stream and also grabs all tags that flow past and stores them locally. These tags are to be used for testing that the correct tags are coming through. | |||
2010-11-11 | Adding an annotator block that moves tags using 1-to-1 policy. | Tom Rondeau | |
2010-11-11 | Renaming random_annotator to annotator_1toall that moves tags from single ↵ | Tom Rondeau | |
input to all outputs. | |||
2010-11-11 | Fixed small bug in how tags are propagated downstream. Also using seq. ↵ | Tom Rondeau | |
numbers in annotator test block to better keep track. | |||
2010-11-08 | Burned by another copy-paste error. Correctly checking and acting if its a sink. | Tom Rondeau | |
2010-11-08 | Adding mutex protection around adding and getting tags from buffers. | Tom Rondeau | |
2010-11-08 | Made propagate_tags a function to be called from block_executor to move tags ↵ | Tom Rondeau | |
downstream. Also made d_start_nitems_read a member of gr_block_executor to better handle allocation. | |||
2010-11-08 | Renaming "handling_method" to "propagation_policy". | Tom Rondeau | |
2010-11-08 | cleaning up; comments, naming, typos, excess code, etc. Moving to make ↵ | Tom Rondeau | |
gr_block_executor more readable. | |||
2010-11-07 | Moving tag handling setup back into gr_block so it can be set in the ↵ | Tom Rondeau | |
constructor of a derived block. | |||
2010-11-07 | Moving tags downstream is moved into gr_block_executor. Predefined three ↵ | Tom Rondeau | |
methods of moving tags that are selectable by a gr_block. | |||
2010-11-07 | Cleaning up. Better use of PMTs; comment mods; returning vectors when ↵ | Tom Rondeau | |
getting tags. | |||
2010-11-06 | Don't directly output tag info to screen from gr_random_annotator; instead, ↵ | Tom Rondeau | |
store the stream and write it to stdout when block is being destroyed. This avoids issues of the muliple threads writing simultaneously to screen. | |||
2010-11-06 | Adding call in scheduler to handle tag movements between blocks and some ↵ | Tom Rondeau | |
helper functions to get access and keep track of tags. | |||
2010-11-06 | Because tags can be appended from upstream, they are not consecutive in ↵ | Tom Rondeau | |
time, so we need to cycle through them all. | |||
2010-11-06 | Better formatting of tag information to make info more readable. | Tom Rondeau | |
2010-11-06 | buffer_reader does not hold its own tags but looks upstream to the ↵ | Tom Rondeau | |
associated buffer to get them. | |||
2010-11-04 | Fixing api call to gr_buffer_reader that no longer takes in which_input. | Tom Rondeau | |
2010-11-04 | gr_buffer_reader doesn't need to know which input it is. | Tom Rondeau | |
2010-11-04 | Moving add_item_tag into gr_buffer and get_tags_in_range to ↵ | Tom Rondeau | |
gr_buffer_reader. gr_block_detail takes care of the high-level adding and retrieving tags via the buffers/readers; also takes care of filtering tags by key. Tags are now added to the gr_buffers on 'which_output' while they are retrieved from the gr_buffer_reader based on 'which_input." | |||
2010-11-04 | Making annotator block build up tags with the noutput_items as the value ↵ | Tom Rondeau | |
held. Simply prints info to stdout when tags are retreived. | |||
2010-11-04 | Makding random_annotator simply copy input to output streams; fixes IO ↵ | Tom Rondeau | |
signatures. | |||
2010-11-04 | Adding shell block for a random annotator. This will be used only for ↵ | Tom Rondeau | |
testing the stream tags, which is why its sitting in runtime. | |||
2010-11-04 | Changing API to match changes to gr_block, including adding "srcid" param to ↵ | Tom Rondeau | |
add_item_tag. Added documentation to header file. Changing to deque from list. Still holding the deque locally in block_detail, but will be moved to gr_buffer. Adding tag just builds the tag tuple and appends it; doesn't worry about duplications. | |||
2010-11-04 | Moves gr_block functions dealing with tags into protected space. Adds ↵ | Tom Rondeau | |
documentation to functions in header. Adds a "srcid" parameter to the add_item_tag function. | |||
2010-11-04 | Fixing buffer to update abs_write_offset counter. Keeping access to counters ↵ | Tom Rondeau | |
exposed through gr_block for now, just remaining to nitem_*. | |||
2010-11-02 | Moved number items read/written from gr_block_detail into gr_buffer ↵ | Tom Rondeau | |
(abs_written_offset) and gr_buffer_reader (abs_read_offset). Keeps the API exposed in gr_blocks for now. | |||
2010-11-02 | Adding QA code to perform some tests on tags. Setting and getting items ↵ | Tom Rondeau | |
read/written; testing adding and retrieving tags. | |||
2010-11-02 | Adding some protection to the nitems read/written accessors. Should this ↵ | Tom Rondeau | |
return 0 or throw? | |||
2010-10-31 | Using pmt_eqv for key testing instead of == to make sure typing is worked ↵ | Tom Rondeau | |
out properly. | |||
2010-10-31 | Checks for duplicate entry when adding a new tag. | Tom Rondeau | |
add_item_tag looks at the last tag entered with the given key and tests if the value is the same. If it is the same value, then don't do add a new item. If the value is different, add a new tag of that key to the list. | |||
2010-10-31 | get_tags_in_range now gets items between start and end INCLUSIVELY. | Tom Rondeau | |
2010-10-31 | Adds functional ability to call get_tags_in_range with a specified key. | Tom Rondeau | |
2010-10-31 | Fix to get_tags_in_range. Returns proper list and handles times when list is ↵ | Tom Rondeau | |
empty. | |||
2010-10-26 | First stab at adding get functions for item tags in a given range. | Tom Rondeau | |
2010-10-26 | Giving gr_block_detail a list of pmt tuples to hold item tagging information. | Tom Rondeau | |
Adds ability to add new tags from a block. | |||
2010-10-19 | Adding vectors to gr_block_detail that keep track of the number of samples ↵ | Tom Rondeau | |
read and written from each block's input. Accessor functions allow query of values through gr_block. Had to add gr_uint64 typedef to SWIG for it to understand how to handle the type. | |||
2010-10-14 | uhd: use start and stop methods for single source/sink blocks | Josh Blum | |
2010-10-06 | Merge branch 'master' into next | Johnathan Corgan | |
* master: Checks to make sure XML path is writable before making the XML runner. Ignores it if directory is now writable. Fixed missing set_relative_rate in these two blocks. The others don't actually do it, even though it's counter-intuitive for the pfb_decimate, which is a sync_block (decimation actually care of in the stream_to_streams). | |||
2010-10-06 | Checks to make sure XML path is writable before making the XML runner. ↵ | Tom Rondeau | |
Ignores it if directory is now writable. |