Issue6009 condition pub priv variants #6270

pnkfelix · 2013-05-06T17:11:31Z

My goal is to resolve the question of whether we want to encourage (by example) consistent use of pub to make identifiers publicly-accessible, even in syntax extensions. (If people don't want that, then we can just let this pull request die.)

This is part one of two. Part two, whose contents should be clear from the FIXME's in this commit, would land after this gets incorporated into a snapshot.

These should work now, I hope

There are some better abstractions for this now

Things like the GC heap and unwinding are desirable everywhere the language might be used, not just in tasks. All Rust code should have access to LocalServices.

I am going to use memory_region and boxed_region as the local heap in the new scheduler, for now at least, and I don't have a rust_env available.

Reusing the existing boxed_region implementation from the runtime

For consistency, for all I/O calls, inspecting the error can be done with the io_error condition.

These will make it easier to write I/O code without worrying about errors

A task without an unwinder will abort the process on failure. I'm using this in the runtime tests to guarantee that a call to `assert!` actually triggers some kind of failure (an abort) instead of silently doing nothing. This is essentially in lieu of a working linked failure implementation.

This also reverts some changes to TLS that were leaking memory. Conflicts: src/libcore/rt/uv/net.rs src/libcore/task/local_data_priv.rs src/libcore/unstable/lang.rs

oops.

@graydon

r? @graydon

Only the first portion has to be owned, as it acts as the buffer for the constructed string. The remaining strings can be static.

…r, r=graydon during task annihilation, since it is easy to tread on freed memory.

…tamorphism

…hism I just had `git apply` fix most of them and then did a quick skim over the diff to fix a few cases where it did the wrong thing (mostly replacing tabs with 4 spaces, when someone's editor had them at 8 spaces).

…sion.

(I wonder if there's a better way to write this in the rule itself; i.e. something like `$$(dirname $$@)`. But for now this will do.)

In commit d7f5e43 "core::rt: Add the local heap to newsched tasks", local_malloc and local_free have become rather big and their forced inlining causes quite a bit of code bloat. Compile times for crates affected by the bloat (e.g. rustc) improve, while others (e.g. libstd) seem to be unaffected, so I guess the inlining doesn't gain us much. Sizes: | librustc | libsytax ---------------|–-----------|------------ with inlining | 18,547,824 | 7,110,848 w/o inlining | 15,092,040 | 5,518,608

In commit d7f5e43 "core::rt: Add the local heap to newsched tasks", local_malloc and local_free have become rather big and their forced inlining causes quite a bit of code bloat. Compile times for crates affected by the bloat (e.g. rustc) improve, while others (e.g. libstd) seem to be unaffected, so I guess the inlining doesn't gain us much. Sizes: ``` | librustc | libsytax ---------------|–-----------|------------ with inlining | 18,547,824 | 7,110,848 w/o inlining | 15,092,040 | 5,518,608

The previous example was erroneously attempting to destroy uninitialized memory, which was often zeroed (masking the bug).

transitional patch to resolve compile/link failure on android after rust-lang#6161 landed, I've encountered below errors since android does not support glob in libc. /opt/ndk_standalone/bin/../lib/gcc/arm-linux-androideabi/4.6/../../../../arm-linux-androideabi/bin/ld: /home/yichoi/rust_work/build/x86_64-unknown-linux-gnu/stage1/lib/rustc/arm-linux-androideabi/lib/libcore-c3ca5d77d81b46c1-0.7-pre.so: error: undefined reference to 'glob' /opt/ndk_standalone/bin/../lib/gcc/arm-linux-androideabi/4.6/../../../../arm-linux-androideabi/bin/ld: /home/yichoi/rust_work/build/x86_64-unknown-linux-gnu/stage1/lib/rustc/arm-linux-androideabi/lib/libcore-c3ca5d77d81b46c1-0.7-pre.so: error: undefined reference to 'globfre Since android does not have `glob.h`, `glob_t` definition comes from https://groups.google.com/forum/?fromgroups=#!topic/android-ndk/vSH6MWPD0Vk rust-lang#6100 should be resolved.

The previous example was erroneously attempting to destroy uninitialized memory, which was often zeroed (masking the bug).

…ectory-orderonly-prereqs, r=z0w0 r? (for rust-lang#3326)

pnkfelix · 2013-05-06T17:12:02Z

(argh the back button is broken on github!)

…e_disable, r=flip1995 Clarify allow/warn/deny documentation. Remove enable/disable. Disable and enable when not specifically explained were not clear to me as an English language speaker, but I was able to figure it out fairly easily due to the examples having A/W, which I assumed meant `allow` and `warn`. I removed both words to be sure it was clear as well as extending the note on what deny means. It now includes a statement on exactly what each word means. Documentation only update. *Please keep the line below* changelog: none

brson added 30 commits April 19, 2013 15:18

wip

e47ee77

core::rt: Just some poking at the I/O docs

6a5c4f6

rt: Add rust_dbg_next_port for generating test port numbers

1f97e6d

core::rt: Add a test mod and put run_in_newsched_task there

744ba62

core::rt: Use generated port numbers in tests

d24a3a4

core::rt: Unignore some networking tests

eac629b

These should work now, I hope

core::rt: Remove redundant copy of run_in_newsched_task

1a7561b

core::rt: Don't directly create scheduler types in I/O tests

e564fc7

There are some better abstractions for this now

core::rt: Fix a broken uvio test

a11c032

core::rt: Add next_test_ip4 for generating test addresses

6e17202

core::rt: Improve docs for run_in_newsched_task testing function

bcb3cfb

core::rt: Listener constructors are called and return a

d818ea8

core: Ignore two long-running tests

00474c1

core: Speed up a test case

fa478f5

core::rt:: Implement Reader/Writer for MemReader/MemWriter

4eff313

core::rt: Add LocalServices for thread-local language services

f9069ba

Things like the GC heap and unwinding are desirable everywhere the language might be used, not just in tasks. All Rust code should have access to LocalServices.

rt: Don't make memory_region depend on rust_env

2fe118b

I am going to use memory_region and boxed_region as the local heap in the new scheduler, for now at least, and I don't have a rust_env available.

core::rt: Add the local heap to newsched tasks

d7f5e43

Reusing the existing boxed_region implementation from the runtime

core::rt Wire up task-local storage to the new scheduler

d0786fd

core::rt: Make I/O constructors return Option instead of Result

e5d21b9

For consistency, for all I/O calls, inspecting the error can be done with the io_error condition.

core::rt: Move the definition of Listener to rt::io

6644a03

Make conditions public. rust-lang#6009

fe13b86

core::rt: Add implementations of Reader, Writer, and Listener for Option

5fbb094

These will make it easier to write I/O code without worrying about errors

core::rt: Add unwinding to newsched tasks

42c0f88

mk: Pass CFLAGS to the uv build

a292d51

core::rt: Fix a use after free in uv 'write'

8a2f9ca

mk: core has another level of directories with rt/io/net

f4af40a

Merge remote-tracking branch 'brson/io'

e944c7d

This also reverts some changes to TLS that were leaking memory. Conflicts: src/libcore/rt/uv/net.rs src/libcore/task/local_data_priv.rs src/libcore/unstable/lang.rs

core: Convert reinterpret_cast to transmute in TLS. rust-lang#6039

6373861

catamorphism and others added 27 commits May 3, 2013 16:56

core: Warning police

5fb5a94

std: Warning police

18bf9bd

syntax: Warning police

1a5f11a

rustc: Warning police

2df8799

add gitattributes and fix whitespace issues

86efd97

rustpkg: Make code actually compile

9f76ca6

oops.

auto merge of rust-lang#6124 : catamorphism/rust/rustpkg, r=catamorphism

7d9c638

r? @graydon

auto merge of rust-lang#6198 : luqmana/rust/linker-arg, r=graydon

5bf7e8b

auto merge of rust-lang#6219 : dotdash/rust/fmt, r=graydon

821979f

Only the first portion has to be owned, as it acts as the buffer for the constructed string. The remaining strings can be static.

auto merge of rust-lang#6227 : graydon/rust/issue-6112-box-annihilato…

c3ab74b

…r, r=graydon during task annihilation, since it is easy to tread on freed memory.

auto merge of rust-lang#6229 : catamorphism/rust/warning-police, r=ca…

a47e4cb

…tamorphism

rt: glob, globfree dummy function for android

7ac6571

Fix syntax: had to use escaped $$ to have an effect after first expan…

495bceb

…sion.

Fix another goof: consistently use parent directory of target for rule.

175a5ee

(I wonder if there's a better way to write this in the rule itself; i.e. something like `$$(dirname $$@)`. But for now this will do.)

Register snapshots

8081e8d

auto merge of rust-lang#6240 : brson/rust/snapshots, r=brson

d74ac9e

small fix to the tutorial-ffi destructor example

8f2d71a

The previous example was erroneously attempting to destroy uninitialized memory, which was often zeroed (masking the bug).

rt: rust_android_dummy.cpp fix for make tidy

987ad9c

auto merge of rust-lang#6244 : thestinger/rust/ffi, r=z0w0

3b299e8

The previous example was erroneously attempting to destroy uninitialized memory, which was often zeroed (masking the bug).

auto merge of rust-lang#6235 : pnkfelix/rust/issue-3326-play-with-dir…

063851f

…ectory-orderonly-prereqs, r=z0w0 r? (for rust-lang#3326)

Fix span tests

aca2a00

auto merge of rust-lang#6208 : sanxiyn/rust/accurate-span, r=sanxiyn

6e6a4be

Step one for 'proper' pub condition: support pub keyword in form.

1cbf0a8

pnkfelix closed this May 6, 2013

Provide feedback

Saved searches

Use saved searches to filter your results more quickly

Uh oh!

Issue6009 condition pub priv variants #6270

Issue6009 condition pub priv variants #6270

Uh oh!

pnkfelix commented May 6, 2013

Uh oh!

pnkfelix commented May 6, 2013

Uh oh!

Uh oh!

Issue6009 condition pub priv variants #6270

Issue6009 condition pub priv variants #6270

Uh oh!

Conversation

pnkfelix commented May 6, 2013

Uh oh!

pnkfelix commented May 6, 2013

Uh oh!

Uh oh!