| Summary: | Use artificial recipes for oe-selftest where possible to reduce work | ||
|---|---|---|---|
| Product: | [QA/Testing] Build Testing | Reporter: | Joshua Lock <joshuagloe> |
| Component: | general | Assignee: | Thomas Roos <thomas> |
| Status: | RESOLVED FIXED | QA Contact: | |
| Severity: | enhancement | ||
| Priority: | Low | CC: | benjamin.esquivel, randy.macleod, thomas |
| Version: | 3.0 | ||
| Target Milestone: | 4.99 | ||
| Hardware: | x86 | ||
| OS: | Multiple | ||
| Whiteboard: | NEWCOMER | ||
| OS type for building Yocto: | --- | Type of Regression: | --- |
| Verified: | Documentation change: | No (bug/feature does not impact docs) | |
|
Description
Joshua Lock
2017-03-10 13:25:20 UTC
Moving to 2.5 and reassign this bug to Jair as he will be working on oeqa enhancements So e.g. a simple helloworld.c + recipe should be added to oe-core? found this one here: poky/meta-skeleton/recipes-skeleton/hello-single -> should this then moved/copied to meta-selftest/recipes-test ? Thomas, If you don't get an answer here by early next week, try the email list, perhaps with a patch to focus the discussion. Joshua is mostly on leave for weeks or maybe 2 months so he may not reply quickly if at all. (I'm away next week as well). Hi, just was working on this today again, so what I found:
bbtests.BitbakeTests.test_event_handler
is using this:
bitbake('m4-native')
running just this test takes
orig: Ran 1 test in 75.741s
replaced with hello (copied from oe-skel...)
mod: Ran 1 test in 678.986s
-> so it seems that building a native version will speed up already this a lot.
-> trying hello-native: 29.84s
Wondering if we should always just use -native recipes, then?
Other locations I found, but did not replace or check them, yet:
test_recipe = 'zlib'
test_recipe = 'ed'
bitbake('hicolor-icon-theme')
testrecipe = 'lighttpd'
IMO it makes sense to have a fast hello world (-native?) recipe in meta-selftest /recipes-test to use in some tests where m4-native, zlib, ed... is used!
Or may there is even a faster idea? It really just depends on the specific test, what should be tested.
Please, any thoughts on this?
Also, I will create a patch for this and post it to the mailing list for discussion as a next step.
|