| Age | Commit message (Collapse) | Author |
|---|
|
|
|
|
|
|
|
|
|
|
|
- in sequence containers, use operator[] instead of .at() when
accessing indices guaranteed to be in bound;
- in associative containers, prefer .find() and reusing the
returned const iterator to using .contains() and .at(); while
there, make any so obtained iterators const if they are read-only.
|
|
|
|
|
|
...whenever it is given access to a Git repository.
The referenced storage config needs to outlive the repository
config instance.
|
|
When backing up target-cache entries we use parallelism at two
dimensions, the independent cache entries and for each entry we
retrieve the artifacts in parallel. If for each dimension we use
the full amount of parallelism allowed, that gives a number of
threads up to the square of the amount of parallelism specified by
the user. Therefore, use in each dimension only the square root
of the allowed parallelism keeping the total parallelism (up to
rounding) within the specified range.
|
|
... which might be quite ahead of the end time of the invocation if
writing out of the action graph delays the end of the invocation.
|
|
When using a serve end point, the analysis phase might take quite
long if serve has to actually build a delegated target or, at
least, has to synchronize artifacts with the remote end point.
Therefore, also record the time the build phase started (if building
is requested) as an additional time stamp in the profile.
|
|
|
|
Include in the profile also the effective remote-execution endpoint,
properties, and dispatch list. Software projects are often tested
in a variety of environments or hardware configurations; as,
obviously, the performance might differ significantly (especially
depending on the used hardware) a proper analysis therefore requires
the possibility to distinguish the various backends. Adding the
effective configuration adds this posibility.
|
|
|
|
|
|
|
|
|
|
We use a pointer to the actual Profile object to handle profiling,
if requested. For this to work, actual object needs to stay in
scope. However, we handle most of the operations, including parsing
of arguments, in a global try-catch block. In order to be able to
also correctly write a profile file outside this block, move the
scope of the Profile object to top-level in main.
While there, also improve the signature of the Profile class. That
class is only meaningful, if a profile should eventually be writting
to disk. So reflect this in the constructur. Also, once we know the
file name to write the profile to (if any), we have already parsed
the command line; so the making available of the command line to the
profile can be enforced by adding this to the constructor as well.
Co-authored-by: Denisov Maksim <denisov.maksim@huawei.com>
|
|
As parsing the the command-line is non-trivial, we include all the
relevant information about the command line in the profile. This
should also include the subcommand. For sake of completeness, we
also include the non-option arguments of the subcommand.
|
|
|
|
In order to stay backwards compatible, the "tree_overlays" entry
in action-graph descriptions is optional.
|
|
... and only as late as possible wait for that parallel task.
Serializing a large task to JSON can take several seconds and as
building does not depend on it, we shouldn't delay the build for
with extra information.
|
|
... so far, which actions where considered, and which of those
were cached.
|
|
... so far only keeping track basic properties of the invocation.
Still, the code is already organized to support future extensions;
in particular, we do not have to rely on the Profile class
being copyable.
|
|
If --dump-graph or --dump-plain-graph is given several times, the
action graph wil also be written several times. In this way, regular
use of those options will not be affected by adding them implicitly
through invocation-logging options in the rc file.
|
|
|
|
|
|
Do not emit errors when it does not lead to build failures. Callers
handle the log level at which failures of this method should be
logged.
|
|
|
|
|
|
|
|
|
|
To properly use `just serve`, both the client and the serve instance
must talk to the very same execution endpoint. Typically, both the
client and serve can reach out to the execution endpoint via the same
IP address. However, it might be possible that the client and a serve
instance know the same execution endpoint by means of differnet IP
addresses. For example, the client knows the execution endpoint
address through an _external_ IP address, while the serve instance,
deployed within the same network infrastructure, only knows the
_internal_ IP address.
This patch adds the subkey `"client address"` -- of the key
`"execution endpoint"` -- in the serve configuration file, to specify
the alternative pair `address:port` used by the client.
|
|
|
|
And ensure every user obtains HashFunction from corresponding IExecutionApi
|
|
|
|
|
|
... so that tagging in the associated git root can be synchronized
with other services provided by just serve.
|
|
|
|
...instead of adding it preliminarily.
|
|
|
|
...to let it be stored as an independent instance.
|
|
|
|
|
|
... given that a proper reporting of the return code if
repo_config.SetGitCAS is contained in main anyway. Therefore, log
the details trying to set the git cas to a level less than that of
the final reporting.
|
|
... as the base repository of a computed root might be absent.
|
|
|
|
|
|
...and use expected to replace it.
|
