Rename aktualizr-repo to uptane-generator. #1279
Conversation
Codecov Report
@@ Coverage Diff @@
## master #1279 +/- ##
==========================================
- Coverage 79.29% 79.17% -0.13%
==========================================
Files 176 175 -1
Lines 10394 10367 -27
==========================================
- Hits 8242 8208 -34
- Misses 2152 2159 +7
Continue to review full report at Codecov.
|
|
There are some variables and function names with remnants of the former tool name: |
aktualizr-uptane-gen or aktualizr-xxx-gen ? |
pattivacek
force-pushed
the
refactor/rename-aktualizr-repo
branch
from
09b8868
to
892ee34
Compare
|
I think it's clear but a bit wordy. Good for customer to understand what it is but maybe a little bit long to us to say it in discussing. Maybe it could have an abbreviation at same time ? |
aktualizr-gen-uptane? Or is even just aktualizr-generator good enough? I'm open to discussion and will rebase/redo once we reach a final conclusion. |
|
If we want a shorter name, we can also drop the "aktualizr" part: "uptane-generator", "make-uptane"... |
|
Aktualizr-meta-gen？ |
I thought we had decided that all aktualizr tools should be prefixed by |
Ah I was not aware of this specific discussion. But then the question is how "related" and "adjacent" the tool is. It's clear for aktualizr-info which really complements it but in this case it's mostly not aktualizr-specific (in theory at least), we just happen to make use of it for our aktualizr test suite. |
That's a good point. The only things that really need the prefix are probably bundled tools that are useless without aktualizr. Hence, perhaps "uptane-generator" is good enough as it is. Any objections? |
|
what about uptane-repo-gen ? :) |
pattivacek
force-pushed
the
refactor/rename-aktualizr-repo
branch
from
892ee34
to
a5897e9
Compare
|
Updated to "uptane-generator" after today's conversation. Now's your chance to complain. |
This is hopefully somewhat more clear, and since it isn't specific to aktualizr, nor bundled with aktualizr, "aktualizr" doesn't need to be part of the name. Signed-off-by: Patrick Vacek <patrickvacek@gmail.com>
pattivacek
force-pushed
the
refactor/rename-aktualizr-repo
branch
from
a5897e9
to
32c73ef
Compare
|
I cannot get that first Travis test to pass for this. It keeps timing out at 50 minutes. Should we merge anyway, or is there something else to try here? |
|
The problem is that I've been cheating with ccache to make it fast enough recently. Since this PR touches a lot of files, it must have invalidated a lot of cache, so we're back to the old running time + overhead of downloading cache. But whatever, let's not worry about travis for merging. |
Agreed. But I'm confused on one point: does the ccache not get updated if the Travis build doesn't complete? I'm not surprised the first attempt failed, but I was hoping the incomplete build would've made the next one faster, and that doesn't seem to be the case. Also, #1276 is failing for the same reason despite not changing much.
Agreed, at least for this case. |
pattivacek
changed the title
Unfortunately no, if the build timeouts nothing is uploaded. There is a trick we could use that was mentioned in a blog post: set a timeout in the test itself so that it fails but never timeouts from the point of view of Travis. We might resort to that at some point. |
Ah, I see. So that means if a build is timing out, we'll get stuck in a cycle like the current one where we have to wait until we just get lucky and it succeeds. Could we at least update the ccache after compilation but before running the tests? |
Is this too wordy? I'm open to suggestions or abbreviations, but I want to move away from "aktualizr-repo", which is way to ambiguous and confusing.