add --full-version and compiler info#11339
Conversation
|
This is a proposed change, and open to review and suggestions. |
fe384e3 to
75fef23
Compare
|
Can you show how it'll look like in the description? Also, it'd be good to add some motivation. Remember that by default we expect a ticket before a PR. |
|
The motivation was trying to get that information from someone on Discord who was having odd problems with ETA: misremembered; they were using a Linux-built cabal on FreeBSD. As most OSes don't have built-in Linux emulators, this won't be common; the closest you'd get is Apple's Rosetta(2), which at least has the OS right. I don't consider this final by any means; some programs have a ETA: The reason for both variants is that the former makes more semantic sense, but the latter is a faster shell command line edit to get the additional information if you forget to use it the first time. |
916c671 to
aa47f49
Compare
|
I moved that |
12ecf00 to
e9c66f5
Compare
eeda6ef to
475bba6
Compare
|
Updated the title and description to reflect the new implementation. |
--version--version-full and compiler info
475bba6 to
57d796e
Compare
|
Re-requesting reviews because adding a new option makes it a bit more invasive code-wise. |
57d796e to
a617480
Compare
new code needs re-review
|
It just occurred to me: should the |
|
" I think the one you got by running |
Yes that is so for the first one. Not so for the second one. I've added headings to make that clearer. |
|
Hm. How did you build that? It sorta looks like the problem I hit early on with #11372 (which was the flag setting being missing from one of the project files), but reversed. |
|
|
You're getting a different version of |
The project flags look good: |
Could be and now I don't trust |
|
Oh, hm, I think the warning actually comes from the outer run ( Confirmed: Which still leaves why you didn't get full output. I wonder if blowing away |
|
I deleted |
|
Uh, I think this might be #6472. |
Merge Queue Status
This pull request spent 13 minutes 43 seconds in the queue, including 3 minutes 11 seconds running CI. Required conditions to merge
ReasonThe merge conditions cannot be satisfied due to failing checks Failing checks: HintYou may have to fix your CI before adding the pull request to the queue again. |
|
@Mergifyio queue |
Merge Queue Status
This pull request spent 2 hours 13 minutes 3 seconds in the queue, including 2 hours 2 minutes 26 seconds running CI. Required conditions to merge
|
I'm not sure if this counts as a significant change or not, since it's a new out-of-the-way command line option
The information added by this change is the last line. If a development GHC version is used to build
cabal, the version will include a datestamp (e.g.9.14.0.20251128).The original output of
--versionhas been restored; all new information now lives under--full-version.The reason for this is a perplexing cabal bug report on the FP Discord that turned out to be the user running a Linux build of cabal on FreeBSD. Cabal tends to trust
System.Infofor local system information for some things, but it's hardcoded to the system cabal was built on, not the one it's running on. (I think this can only arise on FreeBSD, although presumably a lesser form might happen on older Apple Silicon Macs running an Intel build via Rosetta2.)Template Α: This PR modifies behaviour or interface
Include the following checklist in your PR:
significance: significantin the changelog file.QA notes:
--versionis back to what it was before any of the git revision changes, including spaces--full-versionincludes git information (if supported) and compiler informationcabal exec cabal -- act-as-setup -- --full-versionincludes git (if available) and compiler informationCaballibrary and have aSetup.hssitting around, verify thatrunhaskell Setup.hs -- --full-versionworks and includes themCabalthan comes with the currentghc?