Tdata4 #162

koch-t · 2015-01-06T18:04:10Z

TODO:

skinkie · 2015-01-06T23:17:31Z

tdata.h

@@ -142,42 +160,59 @@ struct tdata {
    vehicle_journey_t *vjs;
    uint32_t *journey_patterns_at_stop;
    spidx_t *transfer_target_stops;
-    uint8_t  *transfer_dist_meters;
+    uint16_t  *transfer_durations;


Que?
On 7 Jan 2015 00:17, "Stefan de Konink" [email protected] wrote:

In tdata.h
#162 (diff):

@@ -142,42 +160,59 @@ struct tdata {
vehicle_journey_t *vjs;
uint32_t *journey_patterns_at_stop;
spidx_t *transfer_target_stops;

uint8_t *transfer_dist_meters;

uint16_t *transfer_durations;

rrrr_t ?

—
Reply to this email directly or view it on GitHub
https://github.com/bliksemlabs/rrrr/pull/162/files#r22560327.

…ur of "usefull" values. Must not be used." This reverts commit 336a174.

…lues.

Additionally move json_dump to RRRR_DEBUG.

…ater.

Since nobody so far had the guts to properly fix the timezone problem when the timetable timezone != machine timezone

…mode. It seems that the __restrict substitute works for all major compilers.

…store).

when searching for one specific itinerary we can do so by smashing two sets of arrival and departure upon eachother and intersect our stop index.

We should have something that generates timetable4.dat based on dummy for correct scaffolding of tests. Just expecting a random timatable4.dat is probably not a good idea.

The compiler uses much faster instructions to zero-out the memory than our home-grown zero-out. Running bitset_clear 100 times on a bitset with 10 million bits takes about 0.05 seconds, with memset 0.015 seconds.

skinkie reviewed Jan 6, 2015
View reviewed changes

koch-t mentioned this pull request Feb 25, 2015

Support for stoparea's #136

Closed

skinkie and others added 26 commits March 6, 2015 20:23

Test btimetext.

0124027

Revert "If rtime >= RTIME_THREE_DAYS it means we have less than an ho…

3f1acac

…ur of "usefull" values. Must not be used." This reverts commit 336a174.

RTIME_THREE_DAYS check.

3d0f0c5

Tests the polyline code. This shows that the current code has a problem.

c873433

The polyline offsets should have been based on the rounded integer va…

d45271e

…lues.

Update the comments in the JSON renderer.

23479d4

Additionally move json_dump to RRRR_DEBUG.

Cleanup JSON code.

894bb5e

Tests for JSON rendering.

15606ac

Try to install check-0.9.14 from sources in travis.

dfdc106

Update .travis.yml

54eff41

This seams to be a typo.

1dc3f35

Add latlon distance prototype

7ae82e7

Initial tests for the geometry.c source.

b22f4a8

Add an street_network_init. Maybe rename it to street_network_reset l…

652e8cb

…ater.

Add some informal comments to street_network.c

7cd6edf

Fix bug in street_network.c, thanks to TTD.

2569c18

Street_network.c unittest

452f35a

Set travis to Europe/Amsterdam

60f2d37

Since nobody so far had the guts to properly fix the timezone problem when the timetable timezone != machine timezone

Add function to retrieve duration for given SP

0a71052

Add test for streetnetwork duration function

c4730c9

Redo router_result.c , only render optimal advices

92823d8

Undo disable check_plan_invariants

d672a28

Don't repeat the same search twice forward

493e97b

REmove virtual storage since we now use the plan as storage

62ac269

Make function intialize plan_t

c04c539

Fix undefined behaviour error

411c703

skinkie and others added 30 commits September 16, 2015 06:54

We may use the restrict keyword for compilers that support it in C89 …

edb19dc

…mode. It seems that the __restrict substitute works for all major compilers.

To initialise tmode_t empty, we introduce m_none.

d5ae8bb

Fix operator query from python interface

ae4e3f4

Python: kwlist should be NULL terminated

9fd743b

Python: introduce mode filter

096b2f5

Python: change the default routing to router_route_first_departure

3fb4da2

Python: change router.stops() output to GTFS stops.txt style (column …

27b05ac

…store).

Python: introduce router.validity()

3f6fb41

Option to select the different routing api's. UNSTABLE API!

d40ae13

Allow to set walk_speed, walk_slack, walk_max_distance from python.

1c0a86c

router_one_itinerary addition

409292b

when searching for one specific itinerary we can do so by smashing two sets of arrival and departure upon eachother and intersect our stop index.

Add const to tdata_t * where appropriate.

28cd12f

Report the validity of the timetable upon startup.

901b485

rtime_t is 16bits, thus must be << 16

d08cb76

Add documentation

3f0d6a1

rrtiimetable Remove unused parameters

a2ebdbc

Include documentation

c531833

Use apt to download check

6fda752

Move to new travis infra

6630f34

Attempting to fix travis

34df456

Stilling fixing travis

1a2ceb0

Add pkg-config depedency

a8e0544

Set travis to trusty

4e74faa

Switch back to apt

e7d860a

Make function call

46b80e4

Reflect hashgrid_streetnetwork.c rename in Makefile

a7bd68b

6 items in bitset are set, so why do we expect 5? Fix test

4dbac4a

Disable inputs that require timetable4.dat

d3187a3

We should have something that generates timetable4.dat based on dummy for correct scaffolding of tests. Just expecting a random timatable4.dat is probably not a good idea.

Use memset to zero out our bitset

470c26f

The compiler uses much faster instructions to zero-out the memory than our home-grown zero-out. Running bitset_clear 100 times on a bitset with 10 million bits takes about 0.05 seconds, with memset 0.015 seconds.

include <string.h> for memset

4fa33c8

Provide feedback

Saved searches

Use saved searches to filter your results more quickly

Tdata4 #162

Tdata4 #162

koch-t commented Jan 6, 2015

skinkie Jan 6, 2015

koch-t Jan 6, 2015

skinkie Jan 6, 2015

Tdata4 #162

Are you sure you want to change the base?

Tdata4 #162

Conversation

koch-t commented Jan 6, 2015

skinkie Jan 6, 2015

Choose a reason for hiding this comment

koch-t Jan 6, 2015

Choose a reason for hiding this comment

skinkie Jan 6, 2015

Choose a reason for hiding this comment