1) Move a bunch of std::map's to std::unordered_map to get constant-time
look-ups
2) Lengthen progress-reporting intervals to spend more time doing work
and less time talking about it
3) Reverse order of SHAPE_LINE_CHAINs in thermal intersection checks to
make (much) better use of bbox caches
4) Don't re-generate bboxes we already have
5) Fix some autos that weren't by reference (and were therefore copying
large datasets)
6) Rename delta progressDelta so it's easier to search for in future
7) Get rid of a few more autos (because I don't like them)
8) Pass large items to lambdas by reference
Fixes https://gitlab.com/kicad/code/kicad/issues/12130
Pins that are explicitly connected in the schematic should not have an
"unconnected pin" ERC error. But stacked pins do not count as
explicitly connected because the schematic designer has not connected
them
(cherry picked from commit 865bb54591)
Thread pools are long-lasting executors that have close to zero overhead
when launching new jobs. This is advantageous over creating new threads
as we can use this for threading smalling jobs and smaller quanta. It
also avoids the heuristics needed to determine the optimal number of
threads to spawn
Now that we are dealing with individual connection elements that do not
update their connected elements as well, we can thread the update, just
being careful to guard any remaining updates (bus_enty/busLine) that
need reciprocal updating
Fixes https://gitlab.com/kicad/code/kicad/issues/10974
(cherry picked from commit 6a53e318e5)
If the SCH_ITEM has already been processed, the extra time needed to
iterated over the memberset and get the SCH_CONNECTION when we won't use
it is not neccesary.
Fixes https://gitlab.com/kicad/code/kicad/issues/10974
(cherry picked from commit 776a28a10e)
Normally, you will gain by resursing a nested loop only over the
pairs that are not already handled. In this case, however, you lose
time because you step outside of the cache by adding the reciprocal test
at each step.
Instead, we process one element at a time, keeping it cached and loop
over all other elements to add to the connection. This saves us about
75% of the time for larger loops (e.g. stacked power pins on a large
BGA)
Fixes https://gitlab.com/kicad/code/kicad/issues/10974
(cherry picked from commit 3a98eacdb9)
We want to shorting the processing time by marking symbols that have
already been processed. But we must avoid marking symbols that will not
be processed due to their other flags
Fixes https://gitlab.com/kicad/code/kicad/issues/11164
(cherry picked from commit 2208e0db16)
Needed to find bottlenecks in fns, so break out individual sections of
the massive function for easier understanding.
buildItemSubgraphs (one section of the previous function) would build
millions of connections that were never used as stacked pins created X!
connections. Also tested using sets instead of lists and keeping unique
lists to avoid flagging but none of these were as performant as using
flags to remember which items had already been processed.
Fixes https://gitlab.com/kicad/code/kicad/issues/10974
(cherry picked from commit 17b1b68ac7)
They don't define a KiCad string class, so the header file name was
somewhat misleading. But the fact that they didn't match definitely
made coding more difficult.
Certain situations can trigger the stale cleanup being run
on a bus that doesn't actually contain the stated member.
In this case we should just keep calm and carry on -- who knows
what netlist will be generated, but we shouldn't crash.
Fixes https://gitlab.com/kicad/code/kicad/-/issues/8530
ADDED a new pin electrical type "free" for internally unconnected pins.
CHANGED the "unconnected" pin electrical type is now represented by
"no_connect" in files and netlists. (The legacy syntax is also accepted
in files.)
This way the net names can be inspected in eeschema and cross-probing works.
Testcases updated for the name changes
CHANGED: all unconnected pins are now included in the netlist with no_connect_ prefixes
This behavior provided shorter net names, but was confusing if
the user gives an explicit strong driver in the subsheet but not
one in the parent sheet.
Testcases updated for net name changes; connectivity is the same
Fixes https://gitlab.com/kicad/code/kicad/-/issues/4201
Depending on what order the subgraphs are processed in, it's
possible to hit a case where the parent sheet renames a bus
member before the bus neighbor renaming logic runs. In this
case, we don't want to apply the bus neighbor logic as it will
break connectivity to the parent sheet if that bus was not
also renamed (for example, if the hier ports only make net
connections and not bus connections)
Fixes https://gitlab.com/kicad/code/kicad/-/issues/6887
conflicts (more than one name for a net) is disabled.
This missing call was creating issues in tests that need a full connection
calculation, especially ercCheckFloatingWires()
Previously disambiguation was looking for exact matches,
but vector buses are permitted to have different ranges and
still participate in merging, so they need to be disambiguated.
Fixes https://gitlab.com/kicad/code/kicad/-/issues/5925
A subsheet may have more than one instance of a hier label,
but it's generally a mistake if there is no matching hier label
for a port on the parent sheet.
This is harder to do today now that we don't have manual tool
for placing hier ports, but it's still possible if you place
ports but then rename the label on the subsheet.
Fixes https://gitlab.com/kicad/code/kicad/-/issues/3468
Don't copy the recursion test cache when copy-constructing
SCH_SHEET_PATHs as it's not needed in connectivity
Fix a few places that were copying unnecessarily
As a visual aid, we can peek into the child to see if a pin
represents a bus in the child (in the case of aliases etc
we might not be able to tell by the name of the pin).
Only do this if there isn't anything else driving a connection
onto the sheet pin in the parent sheet.
Fixes https://gitlab.com/kicad/code/kicad/-/issues/2619
Set up a new lineage for SCH_ITEMS to get back to the SCHEMATIC
they live on: Items will all be parented to the SCH_SCREEN that
they are added to, and each SCH_SCREEN will point back to the
SCHEMATIC that it is part of. Note that this hierarchy is not
the same as the actual schematic hierarchy, which continues to
be managed through SCH_SHEETs and SCH_SHEET_PATHS.
Iterators are invalidated by vector reallocation. We need to keep the
count in the loop in order to have a valid position in the vector
Fixes https://gitlab.com/kicad/code/kicad/issues/4073
CHANGED: Settings are now stored in versioned sub-directories
ADDED: First-run dialog for migrating settings from a previous version
CHANGED: Settings are now stored as JSON files instead of wxConfig-style INI files
CHANGED: Color settings are now all stored in a separate settings file
CHANGED: The symbol editor and footprint editor now have their own settings files
CHANGED: Color settings are no longer exposed through BOARD object
CHANGED: Page layout editor now uses Eeschema's color scheme
Settings are now managed through a central SETTINGS_MANAGER held by PGM_BASE.
Existing settings will be migrated from the wxConfig format on first run of each application.
Per-application settings are now stored in one class for each application.
Scope: NETLIST_ITEM, CONNECTION_TYPE, ELECTRICAL_PINTYPE,
NET_CONNECTION, NETLIST_ITEM, GRAPHIC_PINSHAPE
Note, the pin type enum had PT_ added to the front to prevent
shadowing of the INPUT symbol on msys2 (see discussion at
c17c9960d8)
This also removes vector cover types which do nothing except obfuscate
the underlying implementation.
Mainly changes SCH_SHEET_PINS and CONFIG_PARAM_ARRAY (which will soon
be replaced by Jon's new stuff).
This moves EESchema DLIST structures to rtree. These changes are more
fundamental than the pcbnew changes from 9163ac543888c01d11d1877d7c1
and 961b22d60 as eeschema operations were more dependent on passing
drawing list references around with SCH_ITEM* objects.
"INPUT" declaration in sch_text.h generate *a lot of* compil warnings.
So it is replaced by PS_INPUT.
For consistency, others member of enum class PINSHEETLABEL_SHAPE are also renamed.
For legacy reasons, it stored left and right "spin" as flipped integers in the file format.
But the code handled the flip in multiple locations rather than just doing it on file io.
This change unifies the internal code and does the mapping in the file I/O.
m_connected_bus_item and m_connected_bus_items pointers were not cleared when
rebuilding connection data.
So in some cases these pointers were not updated, and created a crash
once the old data was no longer available (deleted)
Seth Hillbrand
jean-pierre charras
Jeff Young
Alex
Marek Roszko
Jon Evans
Wayne Stambaugh
Tomasz Wlostowski
Ian McInerney
Jon Evans