GitLab

Since 459093f3 (ERP5Form: implement a basic checkConsistency for
ProxyField, 2021-01-26) we can detect inconsistentencies in ProxyFields
internal data structures. These are the changes after using
checkConsistency(fixit=True) for erp5_web

This relation field is used to select the web pages that will be displayed,
by using the standard ERP5 document management feature of selecting the
best applicable page with the same reference, but this list did not show
the reference of the web page

manage_FTPget is an API to serve the file content to response, it is not
suitable to just get the file/page template content like we need here.
It was causing issues when running through activities, because timerserver
responses does not have the setBase method, but is also wrong because
large files are streamed to response directly.

Zope does not really offer an unified API to get the source of page
templates or files, except the PrincipiaSearchSource method which it uses
for its builtin search engine. It seems suitable to use here as well.

Also add a test for the case where web pages are in a OFS.File, only
the case of PageTemplate was covered.

Use the `ui-list-grid` class on a `ul` element.

```
return gadget.getTranslationDict([Foo])
  .push(function (translation_dict) {
    console.log(translation_dict['Foo'])
  });
```

```
return gadget.getCommandUrlForDict({
  foo_key: {
    command: 'display'
  }
})
  .push(function (url_dict) {
    console.log(url_dict.foo_key);
  });
```

erp5_web_renderjs_ui_test: Add page to be used as example of links to extra menu and cover it with test

The query planner does not seem to notice that we are trying to know if
any row exists matching a set of dependency values, and it keeps scanning
multiple row for each value - which is unproductive.
So split dependency queries from (pseudo-code)
  WHERE <column{,s}> IN <values{, pairs}>
to unions of
  WHERE <column{,s}> = <value{, pair}> LIMIT 1
which produces query plans which do stop immediately when finding a
candidate row.
On a serialization_tag query with 40 values and real-world indexations,
this reduces the number of rows scanned by mariadb from 500 (<10%
efficiency) to 40 (100% efficiency).
The produced SQL is significantly larger (~3x, around 500kB on
real-world sample data, but may vary a lot depending on value length),
but if this has any effect is is more than compensated by the improved
query plan efficiency.

See merge request nexedi/erp5!1355

For test components, we don't use getBusinessTemplateList but define the
business templates to be installed as (test) dependencies of the business
template containing the test.

update test after c3eb0b30 , because non editable editor now render HTML

See merge request nexedi/erp5!957

See merge request nexedi/erp5!1353

currently with a test that tests document_connection

Ensure float/integer field can align right on the listbox
See nexedi/erp5@cfbc621b

This was removed from products in previous commits

Nowadays instead of creating many portal types, we use `use` category to
separate different types of movements, so it makes sense to filter by use,
the same way that we already were able to filter by portal type.

Let's be consistent and use the default size.

When `Form.proxifyField` is used to make a proxy field from a field which has only a TALES and no value, it creates a proxy field with inconsistent internal data structures where `has_value` method does not work as expected. As a consequence, proxy fields to list fields are often broken in the ods/odt rendering.

Here is a scenario to reproduce the problem:

* add `Base_viewDummyFieldLibrary` ERP5 Form
* add a `LinesField`, with id `your_test_base_field`. Change `size` to 1 (not really required but I did this in screenshots here)
* add another `LinesField`, with id `your_test_field` and in this field:
   * set a `default` (`2`) in values, it's only needed to visualise the problem in ods style:
![set default and no items](/uploads/0cb9e229b0ca6761bb15f3d8845ea3f8/image.png)
   * set `items` in TALES (`python: [('one', '1'), ('two', '2')]`) , this will be the problem
![set items in TALES](/uploads/93d90f57e15afdcca60324675b59086a/image.png)
 * So far,...

Try to support displaying web js site from google cache,
which can not access original gadget, due to cross origin ajax query forbidden.

Display original content instead and load the default CSS in such case.

Most ERP5 web sites use a global form wrapping the content.
Modify the html viewer to not drop this content.

See commit 12e4cc8b.

File._setFile does not support that 'data' is a bytes object:
it crashes if there are already data stored on the object.
test_PUT_on_web_page (erp5_dms:testWebDavSupport) is changed to cover
this case and the caller of File._edit is fixed to make that 'file'
is always a file-like object.

File._setFile is also optimized to avoid reading data for nothing
if the document has no data.

also remove dead code

publish_module_standard closes the request, which calls setSite(None),
but the caller still needs the site to be initialized.

Changes dones in tests that have since been moved to bt5 have been reported in them.

Stop using ck_editor in read_only to display HTML.

This allows to get rid of the ck_editor iframe, allowing a faster load and removing the iframe scroll.
The maximise button is no more needed in such case.

It also removes all ck_editor links handling: links can be open in another browser tab now.

This changes impacts both xhtml style and erp5js.

This gadget take an HTML string as parameter.

It first cleans it up (with hardcoded behaviour currently) by dropping unknown tag elements, unknown/unsafe tag attributes.
It is another protection layer on top of asStrippedHTML inside ERP5.

Then, it displays the output HTML and style it with an hardcoded set of rules.