Simplify data embedding #270
Merged
Conversation
This file contains bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
azeey
approved these changes
May 7, 2020
azeey
reviewed
May 7, 2020
This is an internal-only refactoring that changes the mechanics of how the sdf/** file data is embedded into the library. For one, it solves the "static initialization order fiasco" for the embedded data. The data is only loaded into memory upon first use, instead of as the library is being loaded. It also simplifies the ruby embedding code, making it more concise. I hope this is easier to maintain, but it also helps me when sharing Drake with consumers that always build from source, but will not install ruby as a compile-time prerequisite The EmbeddedSdf codegen now emits the cc file with the data; the header is stored in git. This provides an easier way to use a function to retrieve the embedded strings as static locals. The embedded SDF data is now combined into a single map. Embedding files should be boring, ala the WIP std::embed specification. It should not include application-layer logic encoding upgrade paths. This will also make it easier to avoid the "static destruction fiasco" in future commits, due to fewer functions to repair. Signed-off-by: Jeremy Nimmer <jeremy.nimmer@tri.global>
|
FYI I've locally squashed and force-pushed as part of placating the DCO bot
Probably possible, but there's a more complicated use of ruby still in tools/, so we wouldn't be able to drop the ruby dependency yet. That could be future work. For now, this is enough to make my users happier. |
Codecov Report
@@ Coverage Diff @@
## master #270 +/- ##
=======================================
Coverage 86.45% 86.46%
=======================================
Files 59 59
Lines 9051 9065 +14
=======================================
+ Hits 7825 7838 +13
- Misses 1226 1227 +1
Continue to review full report at Codecov.
|
| /// directory such as "1.8/root.sdf", and the values are the contents of | ||
| /// that source file. | ||
| const std::map<std::string, std::string> &GetEmbeddedSdf(); | ||
| } |
There was a problem hiding this comment.
nit Mismatched indentation?
There was a problem hiding this comment.
it looks like we usually close the internal namespace on the same line that it was created (Converter.hh) but not always (SDFImplPrivate.hh), so I'm not worried about it
scpeters
approved these changes
May 7, 2020
scpeters
pushed a commit
that referenced
this pull request
Jun 2, 2020
This is an internal-only refactoring that changes the mechanics of how the sdf/** file data is embedded into the library. For one, it solves the "static initialization order fiasco" for the embedded data. The data is only loaded into memory upon first use, instead of as the library is being loaded. It also simplifies the ruby embedding code, making it more concise. I hope this is easier to maintain, but it also helps me when sharing Drake with consumers that always build from source, but will not install ruby as a compile-time prerequisite The EmbeddedSdf codegen now emits the cc file with the data; the header is stored in git. This provides an easier way to use a function to retrieve the embedded strings as static locals. The embedded SDF data is now combined into a single map. Embedding files should be boring, ala the WIP std::embed specification. It should not include application-layer logic encoding upgrade paths. This will also make it easier to avoid the "static destruction fiasco" in future commits, due to fewer functions to repair. Signed-off-by: Jeremy Nimmer <jeremy.nimmer@tri.global>
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
Add this suggestion to a batch that can be applied as a single commit.
This suggestion is invalid because no changes were made to the code.
Suggestions cannot be applied while the pull request is closed.
Suggestions cannot be applied while viewing a subset of changes.
Only one suggestion per line can be applied in a batch.
Add this suggestion to a batch that can be applied as a single commit.
Applying suggestions on deleted lines is not supported.
You must change the existing code in this line in order to create a valid suggestion.
Outdated suggestions cannot be applied.
This suggestion has been applied or marked resolved.
Suggestions cannot be applied from pending reviews.
Suggestions cannot be applied on multi-line comments.
Suggestions cannot be applied while the pull request is queued to merge.
Suggestion cannot be applied right now. Please check back later.
This is an internal-only refactoring that changes the mechanics of how the
sdf/**file data is embedded into the library.For one, it solves the https://isocpp.org/wiki/faq/ctors#static-init-order fiasco for the embedded data. The data is only loaded into memory upon first use, instead of as the library is being loaded. This is the magic of function-local statics.
It also simplifies the ruby embedding code, making it more concise. I hope this is easier to maintain, but it also helps me when sharing Drake with consumers that always build from source, but will not install ruby as a compile-time prerequisite (RobotLocomotion/drake#13231).
I am happy to take feedback. Is this the correct branch to PR into? Should I be adding a changelog entry? You are welcome to push fixes to my branch -- it looks like by convention I should not be rebasing nor squashing it?