reftest: add test for variables resolutions in filter, for all fields#5643
Conversation
2e16743 to
0f61464
Compare
c191f42 to
d53f6f9
Compare
d53f6f9 to
25837cf
Compare
25837cf to
930e28e
Compare
dra27
left a comment
There was a problem hiding this comment.
I haven't gone through the test properly, but unlike #5642, I think this one is a good candidate for a reftest. I half wondered with the tests which produce messages if it would be worth piping the whole thing through an appropriate grep just to keep the output smaller?
97f6c16 to
d2b04b6
Compare
d2b04b6 to
a9c2b0a
Compare
| ### <pkg:patches.1:some-content> | ||
| blabla | ||
| pioupiou | ||
| bloblob | ||
| ### <pkg:patches.1:to-apply.patch> | ||
| --- a/some-content 2020-12-02 14:22:55.364620832 +0100 | ||
| ++ b/some-content 2020-12-02 14:23:05.668686881 +0100 | ||
| @@ -1,3 +1,3 @@ | ||
| blabla | ||
| -pioupiou | ||
| +ploplop | ||
| bloblob | ||
| ### <pkg:patches.1:to-not-apply.patch> | ||
| --- a/some-content 2020-12-02 14:22:55.364620832 +0100 | ||
| ++ b/some-content 2020-12-02 14:23:05.668686881 +0100 | ||
| @@ -2,3 +2,3 @@ | ||
| ploplop | ||
| -bloblob | ||
| +noooooo |
There was a problem hiding this comment.
these files don't have to be redefined
| ### ::::::::::::::::::::::::::: | ||
| ### :: Self package variable :: | ||
| ### ::::::::::::::::::::::::::: |
There was a problem hiding this comment.
for this section i think it would be also be useful to show the behaviour of opam when the package has already been installed. Like, for each fields, show the behaviour as is done currently, then force-install the package and try to install it again and show if there is a difference of behaviour
There was a problem hiding this comment.
updated with these tests in a separate commit
a9c2b0a to
4972b20
Compare
rjbou
left a comment
There was a problem hiding this comment.
On file definition redundancy, the main idea is to have all information about the given test in its section. But there is 2 different cases:
- patches that never changes, are always redefined, and their content it not so much important to understand the given test
- opam files that are the same for a subset of the tests, they are more important to understand why it fails or not (how is used the variable).
I think that patches can be grouped at the beginning of the test, but opam files add more readability if they are duplicated in their own section.
| ### ::::::::::::::::::::::::::: | ||
| ### :: Self package variable :: | ||
| ### ::::::::::::::::::::::::::: |
There was a problem hiding this comment.
updated with these tests in a separate commit
|
i think this PR is the perfect way to start implementing the new testsuite schema we discussed about. We can use it to see how it works on a small scale, avoid duplication, and start the new testsuite work little by little |
| [ "true" { ?smtg } ] | ||
| [ "false" { !?smtg } ] | ||
| ] | ||
| ### opam install runtest --with-test -vv | sed-cmd true | unordered |
There was a problem hiding this comment.
Are all the unordered necessary?
| ### opam install runtest --with-test -vv | sed-cmd true | unordered | |
| ### opam install runtest --with-test -vv | sed-cmd true |
There was a problem hiding this comment.
Yes, -v displays the Processing [2/3...] and it is unreliable where the output is done (sometimes before the header...)
|
rebased/fixed/updated |
4972b20 to
6c626ef
Compare
6c626ef to
01e0659
Compare
|
Thanks! |
No description provided.