test(semgrep): python and javascript translation

- Move rules to .github folder to reduce clutter in root folder
- separate out rules
- add tests for rules
- docs for writing more rules
This commit is contained in:
Ankush Menat 2021-03-11 23:11:14 +05:30
parent 6ed9567b8c
commit d7bc848a52
No known key found for this signature in database
GPG key ID: 8EA82E09BBD13AAF
8 changed files with 208 additions and 30 deletions

38
.github/helper/semgrep_rules/README.md vendored Normal file
View file

@ -0,0 +1,38 @@
# Semgrep linting
## What is semgrep?
Semgrep or "semantic grep" is language agnostic static analysis tool. In simple terms semgrep is syntax-aware `grep`, so unlike regex it doesn't get confused by different ways of writing same thing or whitespaces or code split in multiple lines etc.
Example:
To check if a translate function is using f-string or not the regex would be `r"_\(\s*f[\"']"` while equivalent rule in semgrep would be `_(f"...")`. As semgrep knows grammer of language it takes care of unnecessary whitespace, type of quotation marks etc.
You can read more such examples in `.github/helper/semgrep_rules` directory.
# Why/when to use this?
We want to maintain quality of contributions, at the same time remembering all the good practices can be pain to deal with while evaluating contributions. Using semgrep if you can translate "best practice" into a rule then it can automate the task for us.
## Running locally
Install semgrep using homebrew `brew install semgrep` or pip `pip install semgrep`.
To run locally use following command:
`semgrep --config=.github/helper/semgrep_rules [file/folder names]`
## Testing
semgrep allows testing the tests. Refer to this page: https://semgrep.dev/docs/writing-rules/testing-rules/
When writing new rules you should write few positive and few negative cases as shown in the guide and current tests.
To run current tests: `semgrep --test --test-ignore-todo .github/helper/semgrep_rules`
## Reference
If you are new to Semgrep read following pages to get started on writing/modifying rules:
- https://semgrep.dev/docs/getting-started/
- https://semgrep.dev/docs/writing-rules/rule-syntax
- https://semgrep.dev/docs/writing-rules/pattern-examples/
- https://semgrep.dev/docs/writing-rules/rule-ideas/#common-use-cases

View file

@ -0,0 +1,6 @@
def function_name(input):
# ruleid: frappe-codeinjection-eval
eval(input)
# ok: frappe-codeinjection-eval
eval("1 + 1")

View file

@ -0,0 +1,10 @@
rules:
- id: frappe-codeinjection-eval
patterns:
- pattern-not: eval("...")
- pattern: eval(...)
message: |
Detected the use of eval(). eval() can be dangerous if used to evaluate
dynamic content. Avoid it or use safe_eval().
languages: [python]
severity: ERROR

View file

@ -0,0 +1,37 @@
// ruleid: frappe-translation-empty-string
__("")
// ruleid: frappe-translation-empty-string
__('')
// ok: frappe-translation-js-formatting
__('Welcome {0}, get started with ERPNext in just a few clicks.', [full_name]);
// ruleid: frappe-translation-js-formatting
__(`Welcome ${full_name}, get started with ERPNext in just a few clicks.`);
// ok: frappe-translation-js-formatting
__('This is fine');
// ok: frappe-translation-trailing-spaces
__('This is fine');
// ruleid: frappe-translation-trailing-spaces
__(' this is not ok ');
// ruleid: frappe-translation-trailing-spaces
__('this is not ok ');
// ruleid: frappe-translation-trailing-spaces
__(' this is not ok');
// ok: frappe-translation-js-splitting
__('You have {0} subscribers in your mailing list.', [subscribers.length])
// todoruleid: frappe-translation-js-splitting
__('You have') + subscribers.length + __('subscribers in your mailing list.')
// ruleid: frappe-translation-js-splitting
__('You have' + 'subscribers in your mailing list.')
// ruleid: frappe-translation-js-splitting
__('You have {0} subscribers' +
'in your mailing list', [subscribers.length])

View file

@ -0,0 +1,53 @@
# Examples taken from https://frappeframework.com/docs/user/en/translations
# This file is used for testing the tests.
from frappe import _
full_name = "Jon Doe"
# ok: frappe-translation-python-formatting
_('Welcome {0}, get started with ERPNext in just a few clicks.').format(full_name)
# ruleid: frappe-translation-python-formatting
_('Welcome %s, get started with ERPNext in just a few clicks.' % full_name)
# ruleid: frappe-translation-python-formatting
_('Welcome %(name)s, get started with ERPNext in just a few clicks.' % {'name': full_name})
# ruleid: frappe-translation-python-formatting
_('Welcome {0}, get started with ERPNext in just a few clicks.'.format(full_name))
subscribers = ["Jon", "Doe"]
# ok: frappe-translation-python-formatting
_('You have {0} subscribers in your mailing list.').format(len(subscribers))
# ruleid: frappe-translation-python-splitting
_('You have') + len(subscribers) + _('subscribers in your mailing list.')
# ruleid: frappe-translation-python-splitting
_('You have {0} subscribers \
in your mailing list').format(len(subscribers))
# ok: frappe-translation-python-splitting
_('You have {0} subscribers') \
+ 'in your mailing list'
# ruleid: frappe-translation-trailing-spaces
msg = _(" You have {0} pending invoice ")
# ruleid: frappe-translation-trailing-spaces
msg = _("You have {0} pending invoice ")
# ruleid: frappe-translation-trailing-spaces
msg = _(" You have {0} pending invoice")
# ok: frappe-translation-trailing-spaces
msg = ' ' + _("You have {0} pending invoices") + ' '
# ruleid: frappe-translation-python-formatting
_(f"can not format like this - {subscribers}")
# ruleid: frappe-translation-python-splitting
_(f"what" + f"this is also not cool")
# ruleid: frappe-translation-empty-string
_("")
# ruleid: frappe-translation-empty-string
_('')

View file

@ -0,0 +1,63 @@
rules:
- id: frappe-translation-empty-string
pattern-either:
- pattern: _("")
- pattern: __("")
message: |
Empty string is useless for translation.
Please refer: https://frappeframework.com/docs/user/en/translations
languages: [python, javascript, json]
severity: ERROR
- id: frappe-translation-trailing-spaces
pattern-either:
- pattern: _("=~/(^[ \t]+|[ \t]+$)/")
- pattern: __("=~/(^[ \t]+|[ \t]+$)/")
message: |
Trailing or leading whitespace not allowed in translate strings.
Please refer: https://frappeframework.com/docs/user/en/translations
languages: [python, javascript, json]
severity: ERROR
- id: frappe-translation-python-formatting
pattern-either:
- pattern: _("..." % ...)
- pattern: _("...".format(...))
- pattern: _(f"...")
message: |
Only positional formatters are allowed and formatting should not be done before translating.
Please refer: https://frappeframework.com/docs/user/en/translations
languages: [python]
severity: ERROR
- id: frappe-translation-js-formatting
patterns:
- pattern: __(`...`)
- pattern-not: __("...")
message: |
Template strings are not allowed for text formatting.
Please refer: https://frappeframework.com/docs/user/en/translations
languages: [javascript, json]
severity: ERROR
- id: frappe-translation-python-splitting
pattern-either:
- pattern: _(...) + ... + _(...)
- pattern: _("..." + "...")
- pattern-regex: '_\([^\)]*\\\s*'
message: |
Do not split strings inside translate function. Do not concatenate using translate functions.
Please refer: https://frappeframework.com/docs/user/en/translations
languages: [python]
severity: ERROR
- id: frappe-translation-js-splitting
pattern-either:
- pattern-regex: '__\([^\)]*[\+\\]\s*'
- pattern: __('...' + '...')
- pattern: __('...') + __('...')
message: |
Do not split strings inside translate function. Do not concatenate using translate functions.
Please refer: https://frappeframework.com/docs/user/en/translations
languages: [javascript, json]
severity: ERROR

View file

@ -19,4 +19,4 @@ jobs:
python -m pip install -q semgrep
git fetch origin $GITHUB_BASE_REF:$GITHUB_BASE_REF -q
files=$(git diff --name-only --diff-filter=d $GITHUB_BASE_REF)
if [ -f .semgrep.yml ]; then semgrep --config=.semgrep.yml --quiet --error $files; fi
[[ -d .github/helper/semgrep_rules ]] && semgrep --config=.github/helper/semgrep_rules --quiet --error $files

View file

@ -1,29 +0,0 @@
#Reference: https://semgrep.dev/docs/writing-rules/rule-syntax/
rules:
- id: eval
patterns:
- pattern-not: eval("...")
- pattern: eval(...)
message: |
Detected the use of eval(). eval() can be dangerous if used to evaluate
dynamic content. Avoid it or use safe_eval().
languages:
- python
severity: ERROR
# translations
- id: frappe-translation-syntax-python
pattern-either:
- pattern: _(f"...") # f-strings not allowed
- pattern: _("..." + "...") # concatenation not allowed
- pattern: _("") # empty string is meaningless
- pattern: _("..." % ...) # Only positional formatters are allowed.
- pattern: _("...".format(...)) # format should not be used before translating
- pattern: _("...") + ... + _("...") # don't split strings
message: |
Incorrect use of translation function detected.
Please refer: https://frappeframework.com/docs/user/en/translations
languages:
- python
severity: ERROR