-
Notifications
You must be signed in to change notification settings - Fork 17
New issue
Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.
By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.
Already on GitHub? Sign in to your account
Fix #46 and #42 and #50 #51
base: master
Are you sure you want to change the base?
Conversation
Fix segfault on 32 bits architecture (like armv6h, armv7h) by forcing a 64 bits size for off_t by using _FILE_OFFSET_BITS=64
I stumbled upon the same issue on my ARMv7 system and fortunately found your fix :) However, there is another bug if the argument The problem is again related to the What is even more weird is that if I copy the function So, it seems like transferring 64-bit arguments across a library boundary causes the problem on 32-bit ARM. But that doesn't make any sense to me. Any ideas? |
Using my fork, with the patches, I don't get any errors when using In whatever case, what command are you running ? After a quick look at cmp_size, I just noticed that psize is declared as an int64_t, but pu_hr_size expects an off_t. If off_t is not 64bits, that could cause some problem ? Beside that, I don't know what to tell you, because I am not a super expert in C. |
I'm using your fork. Now that you ask I noticed that this segfault only occurs if you have a file in your system where the size doesn't match. Maybe that doesn't happen for you? Because I think the error always occurs if the function
No, that's fine and |
I don't see any error or segfault either when a file size does not match. I did not find by myself that solution. I was asking pacman's dev, about a problem, when simply listing files of a package. This is where the segfault happens for paccheck. So I was given that quick fix (using _FILE_OFFSET_BITS) A demo code showing the problem: https://paste.rs/5Os |
hm, that's strange. No idea why that happens for me then.
Ok. I also checked that again. Your patch is indeed correct. The problem is in The question is only how the problem that it introduces can be fixed... |
Oh, it was my fault. I didn't see that it still used the old Everything works fine with your patch. Sorry for the confusion. |
oh. You got bitten with that problem too. 😁 I was too when trying to run my modified version.... I would suggest you to use pacutils-git PKGBUILD but change the repo to use mine. my 2 cents.. |
Yeah, it's pretty difficult to try it out without installing it apparently 😁 Taking the PKGBUILD from pacutils-git and changing the repo sounds good. |
I was doing |
Add all the code and test to deal with NoProgressBar option in /etc/pacman.conf
and fixes man pages issues
and fixes armv7h issues.