start a script to package data, stick it in dropbox, suck it down for

fast bootstrapping. split off developer instructions. not finished,
because I need a better connection to upload the seed data and test
everything...

.gitignore

@@ -4,6 +4,7 @@
 
 data/input/blockface.bin
 data/input/google_transit_2018_18_08
+data/input/neighborhoods
 data/input/neighborhoods.geojson
 data/input/offstreet_parking.kml
 data/input/osm

abstutil/src/time.rs

@@ -349,6 +349,7 @@ impl<'a> Timer<'a> {
 }
 
 impl<'a> std::ops::Drop for Timer<'a> {
+    // TODO This often hides a panic
     fn drop(&mut self) {
         if self.outermost_name == "throwaway" {
             return;
@@ -408,7 +409,6 @@ impl<'a> std::ops::Drop for Timer<'a> {
 }
 
 // For repeated things
-// TODO Why does the PartialEq derivation in sim require this?
 pub struct Profiler {
     entries: Vec<ProfilerEntry>,
     current_entries: HashMap<String, Instant>,

data/grab_seed_data.sh

@@ -0,0 +1,9 @@
+#!/bin/bash
+# Run from the base repo directory: ./data/grab_seed_data.sh
+
+set -e
+
+curl -L -o seed_data.zip http://dropbox.com/TODO	# TODO Need to upload this
+rm -rf data/input data/system
+unzip seed_data.zip
+rm -f seed_data.zip

data/package_for_devs.sh

@@ -0,0 +1,12 @@
+#!/bin/bash
+# Run from the base repo directory: ./data/package_for_devs.sh
+
+set -e
+
+if [ "$USER" != "dabreegster" ]; then
+	echo "Only Dustin runs this script, to help new developers avoid a long data import process.";
+	exit 1;
+fi
+
+zip -r seed_data data/input data/system
+echo "Fire at will: mv seed_data.zip ~/Dropbox"

docs/INSTRUCTIONS.md

@@ -4,7 +4,9 @@ General disclaimer: This is a very rough demo. The user interface is clunky, and
 gameplay is not cohesively tied together yet. Please email
 <dabreegster@gmail.com> or file a Github issue if you hit problems.
 
-## Installing the game from pre-built binaries
+## Installing the game
+
+Grab a pre-built binary release:
 
 - Windows:
   https://github.com/dabreegster/abstreet/releases/download/v0.1.17/abstreet_windows.zip
@@ -17,6 +19,8 @@ Unzip the folder, then run `play_abstreet.sh` or `play_abstreet.bat`. On
 Windows, you'll probably get a warning about running software from an unknown
 publisher.
 
+Or you can [build from source](/docs/dev.md).
+
 ## Playing the game
 
 General controls:
@@ -43,46 +47,6 @@ Things to try:
 - Go back to the main menu and pick a challenge. Should be self-explanatory from
   there -- leave me feedbck if not.
 
-## For developers: Compiling from source
-
-You will first need:
-
-- Standard dependencies: `bash`, `curl`, `unzip`, `gunzip`
-- `osmconvert`: See https://wiki.openstreetmap.org/wiki/Osmconvert#Download
-- Rust, at least 1.38. https://www.rust-lang.org/tools/install
-
-One-time setup:
-
-1.  Download the repository:
-    `git clone https://github.com/dabreegster/abstreet.git`
-
-2.  Download all input data and build maps. Compilation times will be very slow
-    the first time. `cd abstreet; ./import.sh && ./precompute.sh --release`
-
-3.  Run the game: `cd game; cargo run --release`
-
-### Development tips
-
-- You don't need to rerun `./import.sh` nd `./precompute.sh` most of the time.
-  - If you're just touching code in `game`, `sim`, and `ezgui`, you can just
-    `cargo run` from `game`
-  - If you're modifying the initial OSM data -> RawMap conversion in
-    `convert_osm`, then you do need to rerun `./import.sh` and `precompute.sh`
-    to regenerate the map.
-  - If you're modifying `map_model` but not the OSM -> RawMap conversion, then
-    you can just do `precompute.sh`.
-  - Both of those scripts can just regenerate a single map, which is much
-    faster: `./import.sh caphill; ./precompute.sh caphill`
-- Compile faster by just doing `cargo run`. The executable will have debug stack
-  traces and run more slowly. You can do `cargo run --release` to build in
-  optimized release mode; compilation will be slower, but the executable much
-  faster.
-- To add some extra debug modes to the game, `cargo run -- --dev` or press
-  Control+S to toggle in-game
-- All code is automatically formatted using
-  https://github.com/rust-lang/rustfmt; please run `cargo fmt` before sending a
-  PR.
-
 ## Data source licensing
 
 A/B Street binary releases contain pre-built maps that combine data from:

docs/dev.md

@@ -0,0 +1,54 @@
+# Developer guide
+
+## Getting started
+
+You will first need:
+
+- Standard dependencies: `bash`, `curl`, `unzip`, `gunzip`
+- Rust, at least 1.38. https://www.rust-lang.org/tools/install
+
+One-time setup:
+
+1.  Download the repository:
+    `git clone https://github.com/dabreegster/abstreet.git`
+
+2.  Build all input data. This is very slow, so you should seed from a pre-built
+    copy: `./data/grab_seed_data.sh`. This will download about 1GB and expand to
+    about 5GB.
+
+3.  Run the game: `cd game; cargo run --release`
+
+## Development tips
+
+- Compile faster by just doing `cargo run`. The executable will have debug stack
+  traces and run more slowly. You can do `cargo run --release` to build in
+  optimized release mode; compilation will be slower, but the executable much
+  faster.
+- To add some extra debug modes to the game, `cargo run -- --dev` or press
+  Control+S to toggle in-game
+- All code is automatically formatted using
+  https://github.com/rust-lang/rustfmt; please run `cargo fmt` before sending a
+  PR.
+
+## Building map data
+
+You can skip this section if you're just touching code in `game`, `ezgui`, and
+`sim`.
+
+You'll need some extra dependencies:
+
+- `osmconvert`: See https://wiki.openstreetmap.org/wiki/Osmconvert#Download
+- `cs2cs` from proj4 (See X or `apt-get install proj-bin` on Ubuntu)
+
+The seed data from `data/grab_seed_data.sh` can be built from scratch by doing
+`./import.sh && ./precompute.sh --release`. This takes a while.
+
+Some tips:
+
+- If you're modifying the initial OSM data -> RawMap conversion in
+  `convert_osm`, then you do need to rerun `./import.sh` and `precompute.sh` to
+  regenerate the map.
+- If you're modifying `map_model` but not the OSM -> RawMap conversion, then you
+  can just do `precompute.sh`.
+- Both of those scripts can just regenerate a single map, which is much faster:
+  `./import.sh caphill; ./precompute.sh caphill`