#58470 closed update (fixed)
Bazel port is permanently outdated
Reported by: | pshabunia-at-epam | Owned by: | cjones051073 (Chris Jones) |
---|---|---|---|
Priority: | Normal | Milestone: | |
Component: | ports | Version: | |
Keywords: | Cc: | RootFunction | |
Port: | bazel |
Description
The most recent port for Bazel build tool is stamped v0.21 which is 5 months and 4 releases beyond "current" as of the day of reporting this issue.
It is important to remind that Bazel Remote Build Executor (RBE) is guaranteed to support only two most recent Bazel clients, hence Bazel developers can find ourselves beyond the horizon of RBE support simply by using MacPorts for managing their Bazel distribution.
Please update the port and keep it up to date, for Rubber Duck's sake! Cheers and thank you.
Change History (13)
comment:1 Changed 5 years ago by pshabunia-at-epam
Owner: | set to RootFunction,cjones051073,openmaintainer |
---|---|
Status: | new → assigned |
comment:2 Changed 5 years ago by pshabunia-at-epam
Owner: | changed from RootFunction,cjones051073,openmaintainer to cjones051073 |
---|
comment:3 Changed 5 years ago by cjones051073 (Chris Jones)
comment:4 Changed 5 years ago by mf2k (Frank Schima)
The Milestone field is for use by Macports team members only, please do not set it.
comment:5 Changed 5 years ago by mf2k (Frank Schima)
Milestone: | MacPorts 2.6.0 |
---|---|
Version: | 2.5.4 |
comment:6 Changed 5 years ago by mf2k (Frank Schima)
Cc: | RootFunction added |
---|
comment:7 Changed 5 years ago by cjones051073 (Chris Jones)
Just to add, currently the only usage of bazel within macports, is the py-tensorflow builds, and this only supports specific old bazel versions.
However.... if there is interest in using bazel further afield, I could be convinced to provide one version, bazel, that is kept more up to date, and an older legacy build that is compatible with whatever tensorflow etc. needs...
comment:8 Changed 5 years ago by pshabunia-at-epam
In the meantime, I had to uninstall the port and install Bazel directly.
I also understand that bazel is both (a) a standalone tool/port and (b) a dependency of py-tensorflow port
Something is certainly wrong if MacPorts prioritizes outdated version of bazel in order to support py-tensorflow port at the cost of the user experience of immediate bazel users. This should be fixed in the future. Thanks for your help.
comment:9 Changed 5 years ago by cjones051073 (Chris Jones)
MacPorts is community maintained, by volunteers. I don't appreciate the tone of your messages here... The best way to get changes made is to submit GitHub pull requests yourself. I look forward to your submission in this regard...
comment:10 Changed 5 years ago by pshabunia-at-epam
Sorry, Chris, I didn't intend to hurt anyone's feeling.
I did my best to draw the attention of maintainers that the port out of date to the extent of uselessness for the majority of users. You invited me to figure it out and make a pull request, but I'm not sure why should I bother? As long as Bazel can be installed directly and via Homebrew, it is indeed an unnecessary burden for the MacPorts community.
As I said, I've solved the problem by uninstalling the port and running the latest Bazel installer.
Thanks for your guidance. Pavel
comment:11 Changed 5 years ago by raimue (Rainer Müller)
Summary: | Bazel port is permamently outdated → Bazel port is permanently outdated |
---|
Yes, you are drawing attention to it, but you do so by making demands. However, by getting free benefits off of the work contributed by others, you are not in the position for that. The right question would have been, how can you help to get the latest version into MacPorts faster? Then you would have learned that there is a good reason why updates to Bazel have been kept back. And as already said, the fastest way to get updated versions into MacPorts is indeed to submit a pull request with a patch.
comment:12 Changed 5 years ago by Chris Jones <jonesc@…>
Resolution: | → fixed |
---|---|
Status: | assigned → closed |
Unfortunately the Bazel port needs to be pinned to the version that tensor flow supports, which is unfortunately currently what the port provides.