chore(configure): be more defensive about SpiderMonkey location #4380
Conversation
|
Hi, we should always port changes in configure and Makefile to the windows powershell scripts too… |
|
Oh, and it seems I also did not consider the |
|
Hi @big-r81 : I tried to mention that this is for Unix-based systems -- I do not have a Windows instance around to experiment with the changes. But I can take a chance with that. |
|
Sure, I am happy to consider the Windows bits as well. Maybe I can find a Windows system somewhere to poke around with this. |
|
@big-r81 do you have any ideas why this current version fails on the CI? Initially I thought that is due to |
|
Are these the only possibilities for the paths? On my macOs it's |
|
Hrm, that is interesting. I have checked the corresponding |
|
Yeah, I'm wondering too atm... Update: |
ecbe4a4
to
5322b35
Compare
|
I extracted the commit which modified the |
055d6a9
to
6544294
Compare
The `configure` script does not check whether SpiderMonkey actually exists at the presumed location. This may go wrong when the user has a version different from the default one. The mistake is spotted only in build time, indirectly, via missing header files. That is too late and it may not be evident for the user what the problem is. Add a user-friendly safeguard for Unix-like systems to prevent this from happening.
78a671f
to
0ab748a
Compare
The
configurescript does not check whether SpiderMonkey actually exists at the presumed location. This may go wrong when the user has a version different from the default one. The mistake is spotted only in build time, indirectly, via missing header files. That is too late and it may not be evident for the user what the problem is.Add a user-friendly safeguard to prevent this from happening.