Closed Vladyyy closed 6 months ago
I suspect that you are using the 2.0 release. version This looks a lot like an alignment bug that I fixed a couple of months ago. Is it possible to try again with TOT?
If this is TOT - I'm afraid that I will need a testcase as I don't see the same issue in our reports.
I suspect that you are using the 2.0 release version
yes
Is it possible to try again with TOT?
Yes, I'll give it a shot
I've tried to generate a new RPM from HEAD of master however I think the make dist target is broken:
> make dist
DIST lcov-2.0.tar.gz
make[4]: *** No rule to make target `clean'. Stop.
Error: cannot stat bin/analyzeInfoFiles: No such file or directory
make: *** [lcov-2.0.tar.gz] Error 2
This seems to be a regression introduced by this commit: https://github.com/linux-test-project/lcov/commit/c9b83cf69681b5bca46bde213d2c8e2196395349
Reverting the changes to the makefile fixed the dist target
I've hit some other issues running on TOT and couldn't get a successful run. I opened these two new issues:
The other issues are fixed (I believe) - could you try again with TOT, and report back if there are any other issues or strange behaviour that pop up. Go ahead and close if the issue seems fixed. Thanks Henry
I had to fix a few more issues to get the dist target working.
The rpmbuild picks up local perl modules as dependencies. I applied the fix below
diff --git a/rpm/lcov.spec b/rpm/lcov.spec
index f7af637..cb704ea 100644
--- a/rpm/lcov.spec
+++ b/rpm/lcov.spec
@@ -17,7 +17,7 @@ Prefix: /etc
%define _binary_payload w9.gzdio
# lcov Perl modules are not intended for use by other packages
-%define __requires_exclude ^perl\\(lcovutil\\)$
+%define __requires_exclude ^perl\\(lcovutil\\)$|^perl\\((criteria)\\)$|^perl\\((annotateutil)\\)$|^perl\\((gitblame)\\)$|^perl\\((gitversion)\\)$|^perl\\((p4annotate)\\)
%define __provides_exclude ^perl.*$
Rpm build fails with this failure:
Compiling /tmp/tmp.sz4Cb3LQkG/BUILDROOT/lcov-2.0-111.g79e574b.x86_64/usr/share/lcov/support-scripts/spreadsheet.py ...
File "/usr/share/lcov/support-scripts/spreadsheet.py", line 317
summaryKeys = (*geninfoSpecialKeys, *geninfoChunkKeys, *geninfoKeys)
^
SyntaxError: invalid syntax
error: Bad exit status from /var/tmp/rpm-tmp.CEHhB0 (%install)
I've removed the support-scripts/spreadsheet.py
locally to unblock this
The new tests added to the share
folders are not included in the RPM
error: Installed (but unpackaged) file(s) found:
/usr/share/lcov/example/Makefile
/usr/share/lcov/example/README
/usr/share/lcov/example/descriptions.txt
/usr/share/lcov/example/example.c
/usr/share/lcov/example/example_mod.c
...
Made the change locally
@@ -48,7 +48,7 @@ rm -rf $RPM_BUILD_ROOT
/usr/bin/*
/usr/lib/*
/usr/share/man/man*/*
-/usr/share/lcov/support-scripts/*
+/usr/share/lcov/*
%config /etc/*
It turns out that we use make PREFIX=... install
rather than RPM - and it appears that I can't actually test rpmbuild (for various strange reasons).
As as result: happy to take a PR which fixes whatever RPM issues there are - but I don't think I can fix it.
Wondering if you can test the original issues (this one and #249) with the updated code - so we can make progress on the functionality while \<somebody> figures out the packaging.
Perhaps we can move the RPM issues into a new report. Hopefully, some knowledgeable and motivated person will see it.
My github account is not setup correctly ATM, this is the working diff for the patch
Wondering if you can test the original issues (this one and https://github.com/linux-test-project/lcov/issues/249) with the updated code - so we can make progress on the functionality while
figures out the packaging
This issue has been fixed in TOT
My github account is not setup correctly ATM, this is the working diff for the patch
Thanks for the fix. I think I got it in dcceb6df81a8e9 - but I too have some issues - say, with corporate firewall... Could you test TOT one more time, and make sure that I correctly applied your patch? Then, if this is working - go ahead and close this issue.
Sorry for all the back and forth.
Test locally, works as expeced:
> make dist
DIST lcov-2.0.tar.gz
make[4]: *** No rule to make target `clean'. Stop.
DIST lcov-2.0-112.gdcceb6d.noarch.rpm
make[4]: *** No rule to make target `clean'. Stop.
find: 'debug': No such file or directory
+ test -z ''
+ '[' -z /tmp/tmp.dovd632Lbg/BUILDROOT/lcov-2.0-112.gdcceb6d.x86_64 -o /tmp/tmp.dovd632Lbg/BUILDROOT/lcov-2.0-112.gdcceb6d.x86_64 = / ']'
++ mktemp /tmp/cbr.XXXXXX
+ tmp=/tmp/cbr.wpl1n9
+ trap 'rm -f /tmp/cbr.wpl1n9' EXIT
+ find /tmp/tmp.dovd632Lbg/BUILDROOT/lcov-2.0-112.gdcceb6d.x86_64 '!' '(' -name '*.pyo' -o -name '*.pyc' -o -name '*.elc' -o -name .packlist ')' -type f -print0
+ LANG=C
+ xargs -0r grep -F /tmp/tmp.dovd632Lbg/BUILDROOT/lcov-2.0-112.gdcceb6d.x86_64
+ test -s /tmp/cbr.wpl1n9
+ :
+ rm -f /tmp/cbr.wpl1n9
DIST lcov-2.0-112.gdcceb6d.src.rpm
Excellent - thanks.
In the html report, in the directory view, the differential coverage buckets are missing numbers. See crimson color squares in the picture below
I would expect to see the numbers they way they show up in the top level report