pulumi/pkg/diag/sink.go

210 lines
6.4 KiB
Go
Raw Normal View History

2018-05-22 21:43:36 +02:00
// Copyright 2016-2018, Pulumi Corporation.
//
// Licensed under the Apache License, Version 2.0 (the "License");
// you may not use this file except in compliance with the License.
// You may obtain a copy of the License at
//
// http://www.apache.org/licenses/LICENSE-2.0
//
// Unless required by applicable law or agreed to in writing, software
// distributed under the License is distributed on an "AS IS" BASIS,
// WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
// See the License for the specific language governing permissions and
// limitations under the License.
package diag
import (
"bytes"
"fmt"
"io"
"io/ioutil"
"github.com/pulumi/pulumi/pkg/diag/colors"
"github.com/pulumi/pulumi/pkg/util/contract"
"github.com/pulumi/pulumi/pkg/util/logging"
)
// Sink facilitates pluggable diagnostics messages.
type Sink interface {
// Logf issues a log message.
Logf(sev Severity, diag *Diag, args ...interface{})
// Debugf issues a debugging message.
Debugf(diag *Diag, args ...interface{})
Improve output formatting This change improves our output formatting by generally adding fewer prefixes. As shown in pulumi/pulumi#359, we were being excessively verbose in many places, including prefixing every console.out with "langhost[nodejs].stdout: ", displaying full stack traces for simple errors like missing configuration, etc. Overall, this change includes the following: * Don't prefix stdout and stderr output from the program, other than the standard "info:" prefix. I experimented with various schemes here, but they all felt gratuitous. Simply emitting the output seems fine, especially as it's closer to what would happen if you just ran the program under node. * Do NOT make writes to stderr fail the plan/deploy. Previously we assumed that any console.errors, for instance, meant that the overall program should fail. This simply isn't how stderr is treated generally and meant you couldn't use certain logging techniques and libraries, among other things. * Do make sure that stderr writes in the program end up going to stderr in the Pulumi CLI output, however, so that redirection works as it should. This required a new Infoerr log level. * Make a small fix to the planning logic so we don't attempt to print the summary if an error occurs. * Finally, add a new error type, RunError, that when thrown and uncaught does not result in a full stack trace being printed. Anyone can use this, however, we currently use it for config errors so that we can terminate with a pretty error message, rather than the monstrosity shown in pulumi/pulumi#359.
2017-09-23 14:20:11 +02:00
// Infof issues an informational message (to stdout).
Infof(diag *Diag, args ...interface{})
Improve output formatting This change improves our output formatting by generally adding fewer prefixes. As shown in pulumi/pulumi#359, we were being excessively verbose in many places, including prefixing every console.out with "langhost[nodejs].stdout: ", displaying full stack traces for simple errors like missing configuration, etc. Overall, this change includes the following: * Don't prefix stdout and stderr output from the program, other than the standard "info:" prefix. I experimented with various schemes here, but they all felt gratuitous. Simply emitting the output seems fine, especially as it's closer to what would happen if you just ran the program under node. * Do NOT make writes to stderr fail the plan/deploy. Previously we assumed that any console.errors, for instance, meant that the overall program should fail. This simply isn't how stderr is treated generally and meant you couldn't use certain logging techniques and libraries, among other things. * Do make sure that stderr writes in the program end up going to stderr in the Pulumi CLI output, however, so that redirection works as it should. This required a new Infoerr log level. * Make a small fix to the planning logic so we don't attempt to print the summary if an error occurs. * Finally, add a new error type, RunError, that when thrown and uncaught does not result in a full stack trace being printed. Anyone can use this, however, we currently use it for config errors so that we can terminate with a pretty error message, rather than the monstrosity shown in pulumi/pulumi#359.
2017-09-23 14:20:11 +02:00
// Infoerrf issues an informational message (to stderr).
Infoerrf(diag *Diag, args ...interface{})
// Errorf issues a new error diagnostic.
Errorf(diag *Diag, args ...interface{})
// Warningf issues a new warning diagnostic.
Warningf(diag *Diag, args ...interface{})
// Stringify stringifies a diagnostic into a prefix and message that is appropriate for printing.
Stringify(sev Severity, diag *Diag, args ...interface{}) (string, string)
}
// Severity dictates the kind of diagnostic.
type Severity string
const (
Debug Severity = "debug"
Info Severity = "info"
Improve output formatting This change improves our output formatting by generally adding fewer prefixes. As shown in pulumi/pulumi#359, we were being excessively verbose in many places, including prefixing every console.out with "langhost[nodejs].stdout: ", displaying full stack traces for simple errors like missing configuration, etc. Overall, this change includes the following: * Don't prefix stdout and stderr output from the program, other than the standard "info:" prefix. I experimented with various schemes here, but they all felt gratuitous. Simply emitting the output seems fine, especially as it's closer to what would happen if you just ran the program under node. * Do NOT make writes to stderr fail the plan/deploy. Previously we assumed that any console.errors, for instance, meant that the overall program should fail. This simply isn't how stderr is treated generally and meant you couldn't use certain logging techniques and libraries, among other things. * Do make sure that stderr writes in the program end up going to stderr in the Pulumi CLI output, however, so that redirection works as it should. This required a new Infoerr log level. * Make a small fix to the planning logic so we don't attempt to print the summary if an error occurs. * Finally, add a new error type, RunError, that when thrown and uncaught does not result in a full stack trace being printed. Anyone can use this, however, we currently use it for config errors so that we can terminate with a pretty error message, rather than the monstrosity shown in pulumi/pulumi#359.
2017-09-23 14:20:11 +02:00
Infoerr Severity = "info#err"
Warning Severity = "warning"
Error Severity = "error"
)
// FormatOptions controls the output style and content.
type FormatOptions struct {
Pwd string // the working directory.
Color colors.Colorization // how output should be colorized.
Debug bool // if true, debugging will be output to stdout.
}
2017-01-28 00:42:39 +01:00
// DefaultSink returns a default sink that simply logs output to stderr/stdout.
func DefaultSink(stdout io.Writer, stderr io.Writer, opts FormatOptions) Sink {
contract.Require(stdout != nil, "stdout")
contract.Require(stderr != nil, "stderr")
Improve output formatting This change improves our output formatting by generally adding fewer prefixes. As shown in pulumi/pulumi#359, we were being excessively verbose in many places, including prefixing every console.out with "langhost[nodejs].stdout: ", displaying full stack traces for simple errors like missing configuration, etc. Overall, this change includes the following: * Don't prefix stdout and stderr output from the program, other than the standard "info:" prefix. I experimented with various schemes here, but they all felt gratuitous. Simply emitting the output seems fine, especially as it's closer to what would happen if you just ran the program under node. * Do NOT make writes to stderr fail the plan/deploy. Previously we assumed that any console.errors, for instance, meant that the overall program should fail. This simply isn't how stderr is treated generally and meant you couldn't use certain logging techniques and libraries, among other things. * Do make sure that stderr writes in the program end up going to stderr in the Pulumi CLI output, however, so that redirection works as it should. This required a new Infoerr log level. * Make a small fix to the planning logic so we don't attempt to print the summary if an error occurs. * Finally, add a new error type, RunError, that when thrown and uncaught does not result in a full stack trace being printed. Anyone can use this, however, we currently use it for config errors so that we can terminate with a pretty error message, rather than the monstrosity shown in pulumi/pulumi#359.
2017-09-23 14:20:11 +02:00
// Discard debug output by default unless requested.
debug := ioutil.Discard
if opts.Debug {
debug = stdout
}
return newDefaultSink(opts, map[Severity]io.Writer{
Debug: debug,
Info: stdout,
Improve output formatting This change improves our output formatting by generally adding fewer prefixes. As shown in pulumi/pulumi#359, we were being excessively verbose in many places, including prefixing every console.out with "langhost[nodejs].stdout: ", displaying full stack traces for simple errors like missing configuration, etc. Overall, this change includes the following: * Don't prefix stdout and stderr output from the program, other than the standard "info:" prefix. I experimented with various schemes here, but they all felt gratuitous. Simply emitting the output seems fine, especially as it's closer to what would happen if you just ran the program under node. * Do NOT make writes to stderr fail the plan/deploy. Previously we assumed that any console.errors, for instance, meant that the overall program should fail. This simply isn't how stderr is treated generally and meant you couldn't use certain logging techniques and libraries, among other things. * Do make sure that stderr writes in the program end up going to stderr in the Pulumi CLI output, however, so that redirection works as it should. This required a new Infoerr log level. * Make a small fix to the planning logic so we don't attempt to print the summary if an error occurs. * Finally, add a new error type, RunError, that when thrown and uncaught does not result in a full stack trace being printed. Anyone can use this, however, we currently use it for config errors so that we can terminate with a pretty error message, rather than the monstrosity shown in pulumi/pulumi#359.
2017-09-23 14:20:11 +02:00
Infoerr: stderr,
Error: stderr,
Warning: stderr,
})
}
func newDefaultSink(opts FormatOptions, writers map[Severity]io.Writer) *defaultSink {
contract.Assert(writers[Debug] != nil)
contract.Assert(writers[Info] != nil)
Improve output formatting This change improves our output formatting by generally adding fewer prefixes. As shown in pulumi/pulumi#359, we were being excessively verbose in many places, including prefixing every console.out with "langhost[nodejs].stdout: ", displaying full stack traces for simple errors like missing configuration, etc. Overall, this change includes the following: * Don't prefix stdout and stderr output from the program, other than the standard "info:" prefix. I experimented with various schemes here, but they all felt gratuitous. Simply emitting the output seems fine, especially as it's closer to what would happen if you just ran the program under node. * Do NOT make writes to stderr fail the plan/deploy. Previously we assumed that any console.errors, for instance, meant that the overall program should fail. This simply isn't how stderr is treated generally and meant you couldn't use certain logging techniques and libraries, among other things. * Do make sure that stderr writes in the program end up going to stderr in the Pulumi CLI output, however, so that redirection works as it should. This required a new Infoerr log level. * Make a small fix to the planning logic so we don't attempt to print the summary if an error occurs. * Finally, add a new error type, RunError, that when thrown and uncaught does not result in a full stack trace being printed. Anyone can use this, however, we currently use it for config errors so that we can terminate with a pretty error message, rather than the monstrosity shown in pulumi/pulumi#359.
2017-09-23 14:20:11 +02:00
contract.Assert(writers[Infoerr] != nil)
contract.Assert(writers[Error] != nil)
contract.Assert(writers[Warning] != nil)
return &defaultSink{
opts: opts,
writers: writers,
}
}
Improve output formatting This change improves our output formatting by generally adding fewer prefixes. As shown in pulumi/pulumi#359, we were being excessively verbose in many places, including prefixing every console.out with "langhost[nodejs].stdout: ", displaying full stack traces for simple errors like missing configuration, etc. Overall, this change includes the following: * Don't prefix stdout and stderr output from the program, other than the standard "info:" prefix. I experimented with various schemes here, but they all felt gratuitous. Simply emitting the output seems fine, especially as it's closer to what would happen if you just ran the program under node. * Do NOT make writes to stderr fail the plan/deploy. Previously we assumed that any console.errors, for instance, meant that the overall program should fail. This simply isn't how stderr is treated generally and meant you couldn't use certain logging techniques and libraries, among other things. * Do make sure that stderr writes in the program end up going to stderr in the Pulumi CLI output, however, so that redirection works as it should. This required a new Infoerr log level. * Make a small fix to the planning logic so we don't attempt to print the summary if an error occurs. * Finally, add a new error type, RunError, that when thrown and uncaught does not result in a full stack trace being printed. Anyone can use this, however, we currently use it for config errors so that we can terminate with a pretty error message, rather than the monstrosity shown in pulumi/pulumi#359.
2017-09-23 14:20:11 +02:00
const DefaultSinkIDPrefix = "PU"
// defaultSink is the default sink which logs output to stderr/stdout.
type defaultSink struct {
opts FormatOptions // a set of options that control output style and content.
writers map[Severity]io.Writer // the writers to use for each kind of diagnostic severity.
}
func (d *defaultSink) Logf(sev Severity, diag *Diag, args ...interface{}) {
switch sev {
case Debug:
d.Debugf(diag, args...)
case Info:
d.Infof(diag, args...)
Improve output formatting This change improves our output formatting by generally adding fewer prefixes. As shown in pulumi/pulumi#359, we were being excessively verbose in many places, including prefixing every console.out with "langhost[nodejs].stdout: ", displaying full stack traces for simple errors like missing configuration, etc. Overall, this change includes the following: * Don't prefix stdout and stderr output from the program, other than the standard "info:" prefix. I experimented with various schemes here, but they all felt gratuitous. Simply emitting the output seems fine, especially as it's closer to what would happen if you just ran the program under node. * Do NOT make writes to stderr fail the plan/deploy. Previously we assumed that any console.errors, for instance, meant that the overall program should fail. This simply isn't how stderr is treated generally and meant you couldn't use certain logging techniques and libraries, among other things. * Do make sure that stderr writes in the program end up going to stderr in the Pulumi CLI output, however, so that redirection works as it should. This required a new Infoerr log level. * Make a small fix to the planning logic so we don't attempt to print the summary if an error occurs. * Finally, add a new error type, RunError, that when thrown and uncaught does not result in a full stack trace being printed. Anyone can use this, however, we currently use it for config errors so that we can terminate with a pretty error message, rather than the monstrosity shown in pulumi/pulumi#359.
2017-09-23 14:20:11 +02:00
case Infoerr:
d.Infoerrf(diag, args...)
case Warning:
d.Warningf(diag, args...)
case Error:
d.Errorf(diag, args...)
default:
contract.Failf("Unrecognized severity: %v", sev)
}
}
func (d *defaultSink) createMessage(sev Severity, diag *Diag, args ...interface{}) string {
prefix, msg := d.Stringify(sev, diag, args...)
return prefix + msg
}
func (d *defaultSink) Debugf(diag *Diag, args ...interface{}) {
// For debug messages, write both to the glogger and a stream, if there is one.
logging.V(3).Infof(diag.Message, args...)
msg := d.createMessage(Debug, diag, args...)
if logging.V(9) {
logging.V(9).Infof("defaultSink::Debug(%v)", msg[:len(msg)-1])
}
fmt.Fprint(d.writers[Debug], msg)
}
func (d *defaultSink) Infof(diag *Diag, args ...interface{}) {
msg := d.createMessage(Info, diag, args...)
if logging.V(5) {
logging.V(5).Infof("defaultSink::Info(%v)", msg[:len(msg)-1])
}
fmt.Fprint(d.writers[Info], msg)
}
Improve output formatting This change improves our output formatting by generally adding fewer prefixes. As shown in pulumi/pulumi#359, we were being excessively verbose in many places, including prefixing every console.out with "langhost[nodejs].stdout: ", displaying full stack traces for simple errors like missing configuration, etc. Overall, this change includes the following: * Don't prefix stdout and stderr output from the program, other than the standard "info:" prefix. I experimented with various schemes here, but they all felt gratuitous. Simply emitting the output seems fine, especially as it's closer to what would happen if you just ran the program under node. * Do NOT make writes to stderr fail the plan/deploy. Previously we assumed that any console.errors, for instance, meant that the overall program should fail. This simply isn't how stderr is treated generally and meant you couldn't use certain logging techniques and libraries, among other things. * Do make sure that stderr writes in the program end up going to stderr in the Pulumi CLI output, however, so that redirection works as it should. This required a new Infoerr log level. * Make a small fix to the planning logic so we don't attempt to print the summary if an error occurs. * Finally, add a new error type, RunError, that when thrown and uncaught does not result in a full stack trace being printed. Anyone can use this, however, we currently use it for config errors so that we can terminate with a pretty error message, rather than the monstrosity shown in pulumi/pulumi#359.
2017-09-23 14:20:11 +02:00
func (d *defaultSink) Infoerrf(diag *Diag, args ...interface{}) {
msg := d.createMessage(Info /* not Infoerr, just "info: "*/, diag, args...)
if logging.V(5) {
logging.V(5).Infof("defaultSink::Infoerr(%v)", msg[:len(msg)-1])
Improve output formatting This change improves our output formatting by generally adding fewer prefixes. As shown in pulumi/pulumi#359, we were being excessively verbose in many places, including prefixing every console.out with "langhost[nodejs].stdout: ", displaying full stack traces for simple errors like missing configuration, etc. Overall, this change includes the following: * Don't prefix stdout and stderr output from the program, other than the standard "info:" prefix. I experimented with various schemes here, but they all felt gratuitous. Simply emitting the output seems fine, especially as it's closer to what would happen if you just ran the program under node. * Do NOT make writes to stderr fail the plan/deploy. Previously we assumed that any console.errors, for instance, meant that the overall program should fail. This simply isn't how stderr is treated generally and meant you couldn't use certain logging techniques and libraries, among other things. * Do make sure that stderr writes in the program end up going to stderr in the Pulumi CLI output, however, so that redirection works as it should. This required a new Infoerr log level. * Make a small fix to the planning logic so we don't attempt to print the summary if an error occurs. * Finally, add a new error type, RunError, that when thrown and uncaught does not result in a full stack trace being printed. Anyone can use this, however, we currently use it for config errors so that we can terminate with a pretty error message, rather than the monstrosity shown in pulumi/pulumi#359.
2017-09-23 14:20:11 +02:00
}
fmt.Fprint(d.writers[Infoerr], msg)
Improve output formatting This change improves our output formatting by generally adding fewer prefixes. As shown in pulumi/pulumi#359, we were being excessively verbose in many places, including prefixing every console.out with "langhost[nodejs].stdout: ", displaying full stack traces for simple errors like missing configuration, etc. Overall, this change includes the following: * Don't prefix stdout and stderr output from the program, other than the standard "info:" prefix. I experimented with various schemes here, but they all felt gratuitous. Simply emitting the output seems fine, especially as it's closer to what would happen if you just ran the program under node. * Do NOT make writes to stderr fail the plan/deploy. Previously we assumed that any console.errors, for instance, meant that the overall program should fail. This simply isn't how stderr is treated generally and meant you couldn't use certain logging techniques and libraries, among other things. * Do make sure that stderr writes in the program end up going to stderr in the Pulumi CLI output, however, so that redirection works as it should. This required a new Infoerr log level. * Make a small fix to the planning logic so we don't attempt to print the summary if an error occurs. * Finally, add a new error type, RunError, that when thrown and uncaught does not result in a full stack trace being printed. Anyone can use this, however, we currently use it for config errors so that we can terminate with a pretty error message, rather than the monstrosity shown in pulumi/pulumi#359.
2017-09-23 14:20:11 +02:00
}
func (d *defaultSink) Errorf(diag *Diag, args ...interface{}) {
msg := d.createMessage(Error, diag, args...)
if logging.V(5) {
logging.V(5).Infof("defaultSink::Error(%v)", msg[:len(msg)-1])
}
fmt.Fprint(d.writers[Error], msg)
}
func (d *defaultSink) Warningf(diag *Diag, args ...interface{}) {
msg := d.createMessage(Warning, diag, args...)
if logging.V(5) {
logging.V(5).Infof("defaultSink::Warning(%v)", msg[:len(msg)-1])
}
fmt.Fprint(d.writers[Warning], msg)
}
func (d *defaultSink) Stringify(sev Severity, diag *Diag, args ...interface{}) (string, string) {
var prefix bytes.Buffer
// Now print the message category's prefix (error/warning).
switch sev {
case Debug:
prefix.WriteString(colors.SpecDebug)
case Info, Infoerr:
prefix.WriteString(colors.SpecInfo)
case Error:
prefix.WriteString(colors.SpecError)
case Warning:
prefix.WriteString(colors.SpecWarning)
default:
contract.Failf("Unrecognized diagnostic severity: %v", sev)
}
prefix.WriteString(string(sev))
prefix.WriteString(": ")
prefix.WriteString(colors.Reset)
// Finally, actually print the message itself.
var buffer bytes.Buffer
buffer.WriteString(colors.SpecNote)
if diag.Raw {
buffer.WriteString(diag.Message)
} else {
buffer.WriteString(fmt.Sprintf(diag.Message, args...))
}
buffer.WriteString(colors.Reset)
buffer.WriteRune('\n')
// TODO[pulumi/pulumi#15]: support Clang-style expressive diagnostics. This would entail, for example, using
// the buffer within the target document, to demonstrate the offending line/column range of code.
// Ensure that any sensitive data we know about is filtered out preemptively.
filtered := logging.FilterString(buffer.String())
// If colorization was requested, compile and execute the directives now.
return d.opts.Color.Colorize(prefix.String()), d.opts.Color.Colorize(filtered)
}