Uh oh!
There was an error while loading. Please reload this page.
Add Pass manager for MIR - #31916
Conversation
arielb1
commented
Feb 27, 2016
Is there a way to generate MIR graphs after this PR? |
nagisa
commented
Feb 27, 2016
The graphviz ones? I don’t think so, but an option for |
nikomatsakis
commented
Mar 1, 2016
This basically LGTM. I think the dep-graph node handling here is not right, but it might be worth cleaning that up in a later PR. Sort of sad to lose access to the graphviz code altogether, though I agree that |
bors
commented
Mar 3, 2016
☔ The latest upstream changes (presumably #31979) made this pull request unmergeable. Please resolve the merge conflicts. |
nagisa
commented
Mar 4, 2016
Rebased, comments addressed. |
nikomatsakis
commented
Mar 8, 2016
@bors r+ |
bors
commented
Mar 8, 2016
📌 Commit e30ff06 has been approved by |
nikomatsakis
commented
Mar 8, 2016
@bors r- |
nikomatsakis
commented
Mar 8, 2016
Actually, one question: are we ok with having no more way to get graphviz output (until someone lands a PR)? I am, but I'd love to hear from one or two others on @rust-lang/compiler on this point. |
arielb1
commented
Mar 8, 2016
I like to use graphviz to debug MIR. It is much more readable than the standard debug output. |
arielb1
commented
Mar 8, 2016
I think that having a |
nagisa
commented
Mar 8, 2016
I’ll try implementing that this weekend or something if I don’t get carried away with university things. I’m fine with blocking this on that – it is not like this is important to land this sooner than later. |
Also change output for --unpretty mir to output function names in a prettier way.
nagisa
commented
Mar 12, 2016
EDIT: note that dot does not react well to multigraph dot files with most outputs. e.g. |
arielb1
commented
Mar 13, 2016
@bors r+ |
bors
commented
Mar 13, 2016
📌 Commit bdc176e has been approved by |
bors
commented
Mar 13, 2016
Add Pass manager for MIR A new PR, since rebasing the original one (#31448) properly was a pain. Since then there has been several changes most notable of which: 1. Removed the pretty-printing with `#[rustc_mir(graphviz/pretty)]`, mostly because we now have `--unpretty=mir`, IMHO that’s the direction we should expand this functionality into; 2. Reverted the infercx change done for typeck, because typeck can make an infercx for itself by being a `MirMapPass` r? @nikomatsakis
bors
commented
Mar 13, 2016
driver: restore partially deleted comment Fixesrust-lang#39689. Comment was lost in commit 811b874 (rust-lang#31916@nagisa).
Add missing unpretty option help message There are some missing help messages that is printed `ructc -Zunpretty help` and receiving invalid option. related with rust-lang#16419, rust-lang#45721, rust-lang#21085, rust-lang#31916
Add missing unpretty option help message There are some missing help messages that is printed `ructc -Zunpretty help` and receiving invalid option. related with rust-lang#16419, rust-lang#45721, rust-lang#21085, rust-lang#31916
A new PR, since rebasing the original one (#31448) properly was a pain. Since then there has been several changes most notable of which:
#[rustc_mir(graphviz/pretty)], mostly because we now have--unpretty=mir, IMHO that’s the direction we should expand this functionality into;MirMapPassr? @nikomatsakis