Closed fennecdjay closed 7 years ago
feea535
fails on OSX valgrind tests.
Showing the log and making valgrind more verbose on failure would ease debugging.
Well this is exactly what we're looking for! To have valgrind to cover us in regard of memory leaks and hidden bugs. So it's all fine!
Now, do you want to try to fix the memory issues in this PR or should I merge this already (haven't done reviewing it yet, still) and we'll catch the memory issues later on?
we'll catch the memory issues later on? I'd prefer that, but we should not merge before we see the valgrind log on errors. I'll work on this ASAP.
Sorry for the delay. I think now (with allowed_failures and loggin on errors), it'll be fine.
I'll try this later tonight. Looking fine and it's a great addition. My only nitpick is having to use $ZPM
in cram tests. Can aliases be used instead?
Regards for v0.1.0
@fennecdjay https://github.com/zpm-project/zpm-zsh/releases/tag/v0.1.0
Can aliases be used instead?
I'd like that, but never managed to do so 😕 Any hint appreciated here 😄
Now all leaks come from url.h. Maybe we should fix it (along with for loop declarations) in a fork and make a pull request. What do you think? (I'm not even sure where to discuss this 😖 )
EDIT: you can have a look a jwerle/url.h#3.
I tend to think we could ship a refined zpm-project version of this.
Can aliases be used instead?
I've been rechecking cram's doc and issue tracker, and did not find a clue on how to do that.
From what I understand, cram
launches a new shell on each test file.
All my previous trials failed, so unless you got a way to achive this that you can show me, I think we'll have to stick to the $ZPM
thing.
@fennecdjay We can contribute to url.h
project, either by raising an issue or creating a PR (it's always better to start with an issue to check if there is an interest from the other side).
In regard of aliases I couldn't do it right away. I'm OK with $ZPM
if there is not another option.
I created an issue, and a PR would be welcome :tada:. see jwerle/url.h#4
@fennecdjay Yay! Awesome!
I'll fix this soon. Should we use something like [astyle](), or whichever you prefer, along with a config file to get rid of those annoyances?
I think there nothing else to add here. we could merge and move to next steps.
Awesome! 👍 Great addition!
Also a few improvements:
Makefile
now properly uses clang.clean
target.