retrieveRow needs to pad the index, too

better method names

we need the same access methods in both index types

Changed IndexAccessor names based on access method not content

Add support to access multiple rows at once

When saving word indexs (w*.idx) often multiple words of the same length
will need to be accessed. This implements a new method that allows that
in an eff

Add support to access multiple rows at once

When saving word indexs (w*.idx) often multiple words of the same length
will need to be accessed. This implements a new method that allows that
in an efficient way.

Note: this removes the INDEX_MARK_DELETED mechanism to mark deleted
entries. Entries are now deleted using empty lines again. This makes the
batch handling much simpler. If a good reason exists that we should keep
it, it can be readded.

show more ...

rearranging the Index class structure

This is a first step at stuff at restructuring the indexing classes a
bit more.

Some background:

We have basically two different kind of index files:

a) RowI

rearranging the Index class structure

This is a first step at stuff at restructuring the indexing classes a
bit more.

Some background:

We have basically two different kind of index files:

a) RowIndex (like page.idx)

Each line in the index contains a single value. The line number is used
as primary ID. These files can be very large. Thus an index like that
should never be read into memory completely if it can be avoided.

b) TupleIndex (like i12.idx)

Each line contains a list of tuples. The files tend to be smaller so
loading them completely for search and replace is easier.

Since the the access is so completely different, I tried to model that
in the two different classes, basically moving the methods from
\dokuwiki\Search\AbstractIndex to the new classes.

While doing so, I tried to make the doc blocks, variable names and
interface easier to understand. I also added tests for each of the
methods.

The old code has not been touched yet. So these classes do not do
anything outside of tests currently.

show more ...

use Logger::debug() instead of deprecated dbglog()

catch up #3115 Sort with collator

added missing 'notns' related code

fix undefined array in FulltextSearch

this was already fixed by 5afd958 on 2021-02-05

:fire: fix the calculation of file permissons

Our config allows to set the values for `dmode` and `fmode` to allow
users to explicitly define which permissions directories and files
should have.

To

:fire: fix the calculation of file permissons

Our config allows to set the values for `dmode` and `fmode` to allow
users to explicitly define which permissions directories and files
should have.

To avoid unnessary chmod operations, we check the current umask to
compare what permissions files and directories would get witout our
intervention. If the result is already what the user wants, no chmods
will happen later on. Otherwise we set new configs called `dperm` and
`fperm` which will be used in chmod ops. This is done in
`init_creationmodes()`

When we created new directories, we used to pass the original `dmode`
config to `mkdir()`. The system will then apply the umask to that
`dmode`.

This means the resulting directory will *always* have different
permissions than `dmode`, *always* requiring a chmod operation.
That's silly.

**Breaking Change:** This patch removes the passing of `dmode` as
second parameter to all `mkdir` calls, making it default to `0700`
which is also what we test against in `init_creationmodes()`.

Plugins not relying on our `io_*` functions and do create their own
directories and which currenlty pass `dmode` to it need to be
adjusted to remove that second parameter.

Users may want to reapply the proper file permissions to their data
folder.

**Revert:** In 9fdcc8fcd87114ca59a1764a84d213a53c655c8c @movatica
introduced a change to `init_creationmodes()` that compared the umask
against `fmode` instead of `0666`. I merged it because it looked logical
when compared to the code for directories which compared against `dmode`
as described above. However we do not pass `fmode` to any file creation
methods (that's not possible).

The result is that all changes made in the `fmode` setting resulted
in the wrong permissions for newly created files as first reported in
https://forum.dokuwiki.org/d/19463-setting-fmode-not-working-as-intended

I'm unsure about the orginal motivation behind @movatica's change. The
"fix" however, is wrong.

**Tests:** This patch introduces an integration test that will check the
actual results of directory and file creations under various umask,
`dmode` and `fmode` settings.

show more ...

Fix Undefined array key in Search/Indexer.php

Fix Undefined array key "fperm" warning

Also removed a few, now unnecessary `!empty` checks

Fixes #3382

change MetadataSearch and FulltextSearch to non-singleton

singleton is not effective to reduce multiple instantiations, especially for MetadataSearch which is frequently used in ajax call.

change Tokenizer static utility

frequently used in ajax call, singleton is not effective to reduce multiple instantiations.

fix scrutinizer claims

instantiate *Index with numeric page id

will reduce access to static $pidCache

change Index objects to non-singleton

Indexer, FulltextIndex, MetadataIndex uses common directory to store *.idx files, but this does not mean they should be singleton objects to avoid lock confrict

change Index objects to non-singleton

Indexer, FulltextIndex, MetadataIndex uses common directory to store *.idx files, but this does not mean they should be singleton objects to avoid lock confrictions.

show more ...

avoid null in addMetaKeys()

Just ignore $value argument if $key argument is array .
Ignore enpty key of $key argument.
Ensure to treat any null value of $key array as empty string.

remove unnecessary if blocks

getPID(), saveIndex(), saveIndexKey(), getPageWords() return always true, otherwise exceptions.

throw IndexWriteException in saveIndex()/saveIndexkey()

replace user errors with exceptions

Exceptions are better to handle than errors. What I don't like is that
we now have an unfortunate mix of return code and exception signalling
for errors. Some met

replace user errors with exceptions

Exceptions are better to handle than errors. What I don't like is that
we now have an unfortunate mix of return code and exception signalling
for errors. Some methods still return false for errors while others
now throw exceptions (always returning true otherwise).

show more ...

fix method name

make FulltextIndex::getIndexLengths() public

Third-party plugins may use this method. The [cloud plugin](https://github.com/dokufreaks/plugin-cloud) uses idx_indexLength().

Follow up #2985, fperm seetting

Fixed inconsistent handling of falsy values on fperm setting