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
"-D not hashed" test failure with gcc 6.1.1 #113
Comments
Sure looks like the same issue:
Should have written a test for it. |
Yeah, looks like I botched the refactoring. Now the "# 1" is missing, causing a cache miss... Expected:
Actual:
@jrosdahl : looks we'll be needing a 3.2.7 (problem affecting gcc5/gcc6 only), with a fix and a test |
Actually turns out that it wasn't me, since 4dca108 works but a218001 does not. Have written a test case for it (cf8d693), and fixed the issue (04e36a2). PR upcoming. |
Darn. Thanks for the fix! |
Interestingly, after the change done on master we are now reading the "numbers" that tell whether it's a system header (or not) - but we are not caching those numbers or the trailing quote any longer... |
i.e. we might want to change it back to caching that stuff again |
process_preprocessed_file reads the preprocessed file data and computes a hash. It doesn't cache anything so I don't quite understand what you mean by "caching those numbers or the trailing quote". Could you give me some examples of what you think needs to be fixed after a218001? |
I meant for computing the hash. While troubleshooting the issue (above), I listed all the hash output. The (middle part of the) preprocessor output looks like:
Ccache version 3.2 hashes this as (omitting the header contents):
Whereas ccache 3.3/master hashes this as:
|
OK, so if I understand you correctly you by "after the change done on master we are now reading ... i.e. we might want to change it back to caching that stuff again" meant that master could need fixing but that 3.2-maint works fine. This sounds obvious in retrospect, but I was unsure if it could be interpreted as "the changes made on master regressed and they were then backported to 3.2-maint which now also has regressed". Yes, sounds reasonable to restore hashing of the quotes and numbers again on master. |
Fixed by #115, included in ccache 3.2.7. |
This is with ccache 3.2.6 on Fedora 24 and development, gcc 6.1.1:
Tarball of testdir.8818 contents
The text was updated successfully, but these errors were encountered: