| 1 |
Hello all, |
| 2 |
|
| 3 |
Below is the my question about BUILD_DIR variable in cmake.eclass. I |
| 4 |
have previously raised it in IRC channel, but did not get it clear, so I |
| 5 |
tried to ask the author of cmake.eclass. It seems that Andreas may be |
| 6 |
busy recently and I haven't receive a reply, and I'll keep wait. Thus, I |
| 7 |
forward this question here so we can discuss it and keep a public |
| 8 |
record. If the problem is confirmed to be rather important, I think we |
| 9 |
can move the discussion to Gentoo-dev mailing list. |
| 10 |
|
| 11 |
----- Forwarded message from wuyy <xgreenlandforwyy@×××××.com> ----- |
| 12 |
|
| 13 |
Date: Fri, 15 Jul 2022 16:49:45 +0800 |
| 14 |
From: wuyy <xgreenlandforwyy@×××××.com> |
| 15 |
To: Andreas Sturmlechner <asturm@g.o> |
| 16 |
Subject: A question about BUILD_DIR variable in cmake.eclass under EAPI=8 |
| 17 |
|
| 18 |
Hello Andreas, |
| 19 |
|
| 20 |
I'm a student participating in Google Summer of Code this year, |
| 21 |
packaging ROCm. I have met a problem when using cmake.eclass, and I'm |
| 22 |
not sure whether it should be called a bug or it's just a feature, so I |
| 23 |
want to personally ask you, the author of corresponding code. |
| 24 |
|
| 25 |
As I observe the BUILD_DIR variable is no longer globally set in |
| 26 |
cmake.eclass under EAPI=8, with this change: |
| 27 |
|
| 28 |
-# ${WORKDIR}/${P}_build. |
| 29 |
-: ${BUILD_DIR:=${WORKDIR}/${P}_build} |
| 30 |
+# ${CMAKE_USE_DIR}_build (in EAPI-7: ${WORKDIR}/${P}_build). |
| 31 |
+[[ ${EAPI} == 7 ]] && : ${BUILD_DIR:=${WORKDIR}/${P}_build} |
| 32 |
+# EAPI-8: set inside _cmake_check_build_dir |
| 33 |
|
| 34 |
So I have to use this variable in ebuild after _cmake_check_build_dir |
| 35 |
which is inside cmake_src_prepare. That's not an issue in src_prepare. |
| 36 |
Also this variable can be used in following phases. Everything seems |
| 37 |
nice, until I begin debugging and polishing src_test. After a |
| 38 |
successful src_compile, I come to the src_test, which has bugs. I fixed |
| 39 |
the bug and rerun ebuild xxx src_test, but it unexpectedly failed, |
| 40 |
because BUILD_DIR is empty! |
| 41 |
|
| 42 |
To summarize, BUILD_DIR can live to other phases when phases are |
| 43 |
executed in one call, but cannot be kept if the previous steps are |
| 44 |
skipped. |
| 45 |
|
| 46 |
This is not a problem for most scenarios, but it's a bit annoying when |
| 47 |
developing and debugging ebuilds. I have to call _cmake_check_build_dir |
| 48 |
explicitly before using BUILD_DIR, although src_prepare has been run. I |
| 49 |
wonder this may be a bug, but maybe it's designed to work that way for |
| 50 |
some reasons I don't know. If you think this is a bug, I can put it on |
| 51 |
bugzilla and find a fix to it. Thank you very much! |
| 52 |
|
| 53 |
Best regards |
| 54 |
-- |
| 55 |
Yiyang Wu |
| 56 |
|
| 57 |
----- End forwarded message ----- |
| 58 |
|
| 59 |
-- |
Archives