Add shareable instances of config.toml and openocd.gdb #308

winksaville · 2021-03-05T00:14:19Z

Instead of every chapter having its own copy of the cargo configuration
and openocd.gdb files this change creates a shared set. This will make it
much easier for the user of the discovery book to handle the situation
where their Arm gdb is NOT arm-none-eabi-gdb. As only the shared copy
of .cargo/config.toml will have to be modified.

Also src/05-led-roulette is updated to take advantage of this and I
will go through each of the other chapters changing them to use
the shared set.

I also needed to comment out the rustflags variable in all of the config
files as I'd get an error executing ci/script.sh:

ry.x:5: region 'FLASH' already defined
>>> FLASH : ORIGIN = 0x08000000, LENGTH = 256K
>>>

rust-highfive · 2021-03-05T00:14:21Z

r? @adamgreig

(rust-highfive has picked a reviewer for you, use r? to override)

eldruin

Looks good to me!

src/06-hello-world/.cargo/config

eldruin

Thank you!

bors r+

308: Add shareable instances of config.toml and openocd.gdb r=eldruin a=winksaville Instead of every chapter having its own copy of the cargo configuration and openocd.gdb files this change creates a shared set. This will make it much easier for the user of the discovery book to handle the situation where their Arm gdb is NOT arm-none-eabi-gdb. As only the shared copy of .cargo/config.toml will have to be modified. Also src/05-led-roulette is updated to take advantage of this and I will go through each of the other chapters changing them to use the shared set. I also needed to comment out the rustflags variable in all of the config files as I'd get an error executing ci/script.sh: ry.x:5: region 'FLASH' already defined >>> FLASH : ORIGIN = 0x08000000, LENGTH = 256K >>> Co-authored-by: Wink Saville <[email protected]>

bors · 2021-03-05T09:10:28Z

Build failed:

continuous-integration/travis-ci/push

eldruin · 2021-03-05T09:16:35Z

bors retry

308: Add shareable instances of config.toml and openocd.gdb r=eldruin a=winksaville Instead of every chapter having its own copy of the cargo configuration and openocd.gdb files this change creates a shared set. This will make it much easier for the user of the discovery book to handle the situation where their Arm gdb is NOT arm-none-eabi-gdb. As only the shared copy of .cargo/config.toml will have to be modified. Also src/05-led-roulette is updated to take advantage of this and I will go through each of the other chapters changing them to use the shared set. I also needed to comment out the rustflags variable in all of the config files as I'd get an error executing ci/script.sh: ry.x:5: region 'FLASH' already defined >>> FLASH : ORIGIN = 0x08000000, LENGTH = 256K >>> Co-authored-by: Wink Saville <[email protected]>

bors · 2021-03-05T09:29:31Z

Build failed:

continuous-integration/travis-ci/push

eldruin · 2021-03-05T10:10:14Z

bors retry

308: Add shareable instances of config.toml and openocd.gdb r=eldruin a=winksaville Instead of every chapter having its own copy of the cargo configuration and openocd.gdb files this change creates a shared set. This will make it much easier for the user of the discovery book to handle the situation where their Arm gdb is NOT arm-none-eabi-gdb. As only the shared copy of .cargo/config.toml will have to be modified. Also src/05-led-roulette is updated to take advantage of this and I will go through each of the other chapters changing them to use the shared set. I also needed to comment out the rustflags variable in all of the config files as I'd get an error executing ci/script.sh: ry.x:5: region 'FLASH' already defined >>> FLASH : ORIGIN = 0x08000000, LENGTH = 256K >>> Co-authored-by: Wink Saville <[email protected]>

bors · 2021-03-05T10:23:18Z

Build failed:

continuous-integration/travis-ci/push

eldruin · 2021-03-05T10:53:02Z

It seems building the book just takes too long on Travis :/

bors retry

308: Add shareable instances of config.toml and openocd.gdb r=eldruin a=winksaville Instead of every chapter having its own copy of the cargo configuration and openocd.gdb files this change creates a shared set. This will make it much easier for the user of the discovery book to handle the situation where their Arm gdb is NOT arm-none-eabi-gdb. As only the shared copy of .cargo/config.toml will have to be modified. Also src/05-led-roulette is updated to take advantage of this and I will go through each of the other chapters changing them to use the shared set. I also needed to comment out the rustflags variable in all of the config files as I'd get an error executing ci/script.sh: ry.x:5: region 'FLASH' already defined >>> FLASH : ORIGIN = 0x08000000, LENGTH = 256K >>> Co-authored-by: Wink Saville <[email protected]>

bors · 2021-03-05T11:05:54Z

Build failed:

continuous-integration/travis-ci/push

adamgreig · 2021-03-05T11:07:57Z

I wonder if something is actually wrong - does it work building locally?

adamgreig · 2021-03-05T11:08:56Z

Oh, nevermind, the CI passed on the commit itself, so it should be fine. Just Travis things then?

winksaville · 2021-03-05T15:49:06Z

Do you want me to do anything?

adamgreig · 2021-03-06T13:54:06Z

Let's give it one more go on Travis...

bors retry

308: Add shareable instances of config.toml and openocd.gdb r=eldruin a=winksaville Instead of every chapter having its own copy of the cargo configuration and openocd.gdb files this change creates a shared set. This will make it much easier for the user of the discovery book to handle the situation where their Arm gdb is NOT arm-none-eabi-gdb. As only the shared copy of .cargo/config.toml will have to be modified. Also src/05-led-roulette is updated to take advantage of this and I will go through each of the other chapters changing them to use the shared set. I also needed to comment out the rustflags variable in all of the config files as I'd get an error executing ci/script.sh: ry.x:5: region 'FLASH' already defined >>> FLASH : ORIGIN = 0x08000000, LENGTH = 256K >>> Co-authored-by: Wink Saville <[email protected]>

bors · 2021-03-06T14:06:59Z

Build failed:

continuous-integration/travis-ci/push

winksaville · 2021-03-06T18:02:52Z

Is this failing during ci/script.sh that worked when building on the branch?

winksaville · 2021-03-06T21:09:21Z

Is this failing during ci/script.sh that worked when building on the branch?

So I've done some investigation and there is a difference between how CI works for testing the PR itself and when it's merged. When "testing the PR" mdbook build is run in from ci/after-success while it's run from ci/script.sh when merging because while merging the branch is staging.

The problem appears that mdbook just takes a long time. I modified ci/scripts.sh to adding time RUST_LOG ...:

$ git diff master -- ci/script.sh
diff --git a/ci/script.sh b/ci/script.sh
index acfe2d2..ab6f593 100644
--- a/ci/script.sh
+++ b/ci/script.sh
@@ -2,10 +2,10 @@ set -euxo pipefail
 
 main() {
     # test that building the book works
-    mdbook build
+    time RUST_LOG=error,warn,info mdbook build
 
     # mdbook doesn't handle relative links correctly in print.html so skip it.
-    linkchecker --ignore-url "print.html" book
+    time linkchecker --ignore-url "print.html" book
 
     # now check this as a directory of the bookshelf
     rm -rf shelf

And it took 7 minutes to run mdbook build on Travis when testing the trying branch on my fork.

572.35s$ bash ci/script.sh
+'[' trying '!=' master ']'
+main
+RUST_LOG=error,warn,info
+mdbook build
2021-03-06 20:27:48 [INFO] (mdbook::book): Book building has started
2021-03-06 20:27:48 [INFO] (mdbook::book): Running the html backend
real	7m7.094s
user	0m2.273s
sys	0m10.127s

Although for some reason on my computer it takes 5 seconds, which is weird!
I do have a reasonably quick computer, AMD 3900X with SSD's but it is still
surprising.

$ time TRAVIS_BRANCH=trying bash ci/script.sh
+ '[' trying '!=' master ']'
+ main
+ RUST_LOG=error,warn,info
+ mdbook build
2021-03-06 12:55:00 [INFO] (mdbook::book): Book building has started
2021-03-06 12:55:00 [INFO] (mdbook::book): Running the html backend

real	0m5.397s
user	0m0.312s
sys	0m5.047s
+ linkchecker --ignore-url print.html book
Hello, world!

real	0m0.001s
user	0m0.001s
sys	0m0.000s

Anyway, one solution I thought of was to see I could get mdbook build to output
additional logging information. So I discovered it supported RUST_LOG and as
you can see I have it logging error,warn,info. I chose those because when it was
RUST_LOG=error,warn,info,debug, here, I got an error job exceeded the maximum log length:

2021-03-06 20:24:09 [DEBUG] (mdbook::utils::fs): creating path for file: "/home/travis/build/winksaville/rust-embedded-discovery/book/09-clocks-and-timers/target/thumbv7em-none-eabihf/debug/incremental/aux11-31llwmlgkbkg3/s-fwgnctax4m-mog78m-qtq3c2kgmfbo/query-cache.bin"
2021-03-06 20:24:09 [DEBUG] (mdbook::utils::fs): Copying "/home/travis/build/winksaville/rust-embedded-discovery/src/09-clocks-and-timers/target/thumbv7em-none-eabihf/debug/incremental/aux11-31llwmlgkbkg3/s-fwgnctax4m-mog78m-qtq3c2kgmfbo/query-cache.bin" to "/home/travis/build/winksaville/rust-embedded-discovery/book/09-clocks-and-timers/target/thumbv7em-none-eabihf/debug/incremental/aux11-31llwmlgkbkg3/s-fwgnctax4m-mog78m-qtq3c2kgmfbo/query-cache.bin"
The job exceeded the maximum log length, and has been terminated.

So there appears to be at least three solutions:

Figure out why mdbook build is so slow
Modify mdbook so that RUST_LOG=info will output a little more information,
like displaying the name of each .md file it's processing.
Use a different CI, maybe github actions.

Other idea's?

eldruin · 2021-03-08T08:09:00Z

Thank you for your investigation. We have switched many projects already to GitHub Actions but this one is more complicated. Travis performance has pretty much made an U-turn. See rust-embedded/book#285.

310: Increase Travis no-output waiting time r=therealprof a=eldruin Hopefully this will allow us to merge #308 Co-authored-by: Diego Barrios Romero <[email protected]>

eldruin · 2021-03-08T09:21:38Z

@winksaville could you rebase this onto master? I increased the allowed waiting time in #310

Instead of every chapter having its own copy of the cargo configuration and openocd.gdb files this change creates a shared set. This will make it much easier for the user of the discovery book to handle the situation where their Arm gdb is NOT arm-none-eabi-gdb. As only the shared copy of .cargo/config.toml will have to be modified. Also src/05-led-roulette is updated to take advantage of this and I will go through each of the other chapters changing them to use the shared set. I also needed to comment out the rustflags variable in all of the config files as I'd get an error executing ci/script.sh: ry.x:5: region 'FLASH' already defined >>> FLASH : ORIGIN = 0x08000000, LENGTH = 256K >>>

winksaville · 2021-03-08T17:38:58Z

@winksaville could you rebase this onto master? I increased the allowed waiting time in #310

Done

eldruin

Thank you! Let's see if we get it through this time.

bors r+

308: Add shareable instances of config.toml and openocd.gdb r=eldruin a=winksaville Instead of every chapter having its own copy of the cargo configuration and openocd.gdb files this change creates a shared set. This will make it much easier for the user of the discovery book to handle the situation where their Arm gdb is NOT arm-none-eabi-gdb. As only the shared copy of .cargo/config.toml will have to be modified. Also src/05-led-roulette is updated to take advantage of this and I will go through each of the other chapters changing them to use the shared set. I also needed to comment out the rustflags variable in all of the config files as I'd get an error executing ci/script.sh: ry.x:5: region 'FLASH' already defined >>> FLASH : ORIGIN = 0x08000000, LENGTH = 256K >>> Co-authored-by: Wink Saville <[email protected]>

bors · 2021-03-08T18:53:31Z

Timed out.

winksaville · 2021-03-08T19:00:52Z

Bummer, looks like it got stuck in the job queue and hasn't run, although I could be wrong :(

winksaville · 2021-03-09T04:10:20Z

Travis finally finished, hopefully u can just kick bors :)

eldruin · 2021-03-09T07:21:05Z

Yeah, let's try again...

bors retry

bors · 2021-03-09T07:48:41Z

Build succeeded:

continuous-integration/travis-ci/push

309: Update 06 hello world to use shared cargo config r=eldruin a=winksaville This is dependent on #308 but there is a reduction in number of lines for 06-hello-world, usually a positive :) Co-authored-by: Wink Saville <[email protected]>

winksaville requested a review from a team as a code owner March 5, 2021 00:14

rust-highfive assigned adamgreig Mar 5, 2021

rust-highfive added S-waiting-on-review Status: Awaiting review from the assignee but also interested parties. T-resources labels Mar 5, 2021

eldruin reviewed Mar 5, 2021

View reviewed changes

src/06-hello-world/.cargo/config Show resolved Hide resolved

eldruin previously approved these changes Mar 5, 2021

View reviewed changes

winksaville mentioned this pull request Mar 5, 2021

Update 06 hello world to use shared cargo config #309

Merged

eldruin mentioned this pull request Mar 8, 2021

Increase Travis no-output waiting time #310

Merged

bors bot added a commit that referenced this pull request Mar 8, 2021

Merge #310

f9491d5

310: Increase Travis no-output waiting time r=therealprof a=eldruin Hopefully this will allow us to merge #308 Co-authored-by: Diego Barrios Romero <[email protected]>

winksaville dismissed eldruin’s stale review via 4d57c9a March 8, 2021 17:38

winksaville force-pushed the Common-cargo-config.toml-and-openocd branch from 3f33f3c to 4d57c9a Compare March 8, 2021 17:38

eldruin approved these changes Mar 8, 2021

View reviewed changes

bors bot merged commit 0da0f43 into rust-embedded:master Mar 9, 2021

winksaville deleted the Common-cargo-config.toml-and-openocd branch March 9, 2021 17:58

This was referenced Jul 18, 2021

DO NOT Merge - Test shard cargo config and openocd.gdb #307

Closed

Multiple variations on cargo config #379

Open

Add shareable instances of config.toml and openocd.gdb #308

Add shareable instances of config.toml and openocd.gdb #308

Uh oh!

Conversation

winksaville commented Mar 5, 2021

Uh oh!

rust-highfive commented Mar 5, 2021

Uh oh!

eldruin left a comment

Choose a reason for hiding this comment

Uh oh!

Uh oh!

eldruin left a comment

Choose a reason for hiding this comment

Uh oh!

bors bot commented Mar 5, 2021

Uh oh!

eldruin commented Mar 5, 2021

Uh oh!

bors bot commented Mar 5, 2021

Uh oh!

eldruin commented Mar 5, 2021

Uh oh!

bors bot commented Mar 5, 2021

Uh oh!

eldruin commented Mar 5, 2021

Uh oh!

bors bot commented Mar 5, 2021

Uh oh!

adamgreig commented Mar 5, 2021

Uh oh!

adamgreig commented Mar 5, 2021

Uh oh!

winksaville commented Mar 5, 2021

Uh oh!

adamgreig commented Mar 6, 2021

Uh oh!

bors bot commented Mar 6, 2021

Uh oh!

winksaville commented Mar 6, 2021

Uh oh!

winksaville commented Mar 6, 2021

Uh oh!

eldruin commented Mar 8, 2021

Uh oh!

eldruin commented Mar 8, 2021

Uh oh!

winksaville commented Mar 8, 2021

Uh oh!

eldruin left a comment

Choose a reason for hiding this comment

Uh oh!

bors bot commented Mar 8, 2021

Uh oh!

winksaville commented Mar 8, 2021

Uh oh!

winksaville commented Mar 9, 2021 • edited Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

Uh oh!

eldruin commented Mar 9, 2021

Uh oh!

bors bot commented Mar 9, 2021

Uh oh!

Reviewers

Assignees

Labels

Projects

Milestone

Development

Uh oh!

4 participants

winksaville commented Mar 9, 2021 •

edited

Loading