Hi,

I've got a question about the TableDiff. Imagine that i make some versions of a spreadsheet and when comparing them i get something like this:

What i'm trying to know is how would the coordinate of a cell in this view would be in the final merged table. Because i already have the column name on the top i can know the "x" part of the coordinate, but i don't what would be the "y".

I need this because i want to "undo" certain changes shown in the diff view, e.g don't modify certain cell or undo a column removing or similar.

Thanks!

Continue from #5 as that was initially something else.

PSR-0 examples

\Doctrine\Common\IsolatedClassLoader => /path/to/project/lib/vendor/Doctrine/Common/IsolatedClassLoader.php
\Symfony\Core\Request => /path/to/project/lib/vendor/Symfony/Core/Request.php
\Zend\Acl => /path/to/project/lib/vendor/Zend/Acl.php
\Zend\Mail\Message => /path/to/project/lib/vendor/Zend/Mail/Message.php

Or skip to PSR-4

Hi,

I have the following code: https://gist.github.com/gonzaloserrano/07948738bd3abc91545b

I'm trying to patch this table:

$remoteTable = [
    ['col1', 'col2', 'col3', 'col4', 'col5', 'col6'],
    [0, 0, 0, 0, 2, 0]
];

into this one:

$localTable = [
    ['col1', 'col2', 'col3', 'col4', 'col5', 'col6'],
    [0, 0, 0, 0, 1, 0]
];

And i have errors similar to:

PHP Notice:  Undefined offset: 3 in /Users/gonzalo/dev/sp/config-manager-backend/vendor/paulfitz/daff-php/gen/lib/coopy/PhpTableView.class.php on line 131

Notice: Undefined offset: 3 in /Users/gonzalo/dev/sp/config-manager-backend/vendor/paulfitz/daff-php/gen/lib/coopy/PhpTableView.class.php on line 131
PHP Notice:  Undefined offset: 3 in /Users/gonzalo/dev/sp/config-manager-backend/vendor/paulfitz/daff-php/gen/lib/coopy/PhpTableView.class.php on line 132

Notice: Undefined offset: 3 in /Users/gonzalo/dev/sp/config-manager-backend/vendor/paulfitz/daff-php/gen/lib/coopy/PhpTableView.class.php on line 132
PHP Notice:  Undefined offset: 3 in /Users/gonzalo/dev/sp/config-manager-backend/vendor/paulfitz/daff-php/gen/lib/coopy/PhpTableView.class.php on line 133

If you remove ANY column from the tables except the 5th (which is the one which has the change) it works flawlessly, but like its found in the gist it gives those errors.

I'm using last master version which has the 1.2.4 commit message on it.

Thanks!

either hand coded PHPUnit tests or something generated from the main libs?

hand coded will allow people making bugs to include failing tests quite easily. They can also run them to make sure all is well on their platform

Hi,

In https://github.com/paulfitz/daff-php/blob/master/gen/lib/coopy/PhpTableView.class.php#L84 there is this function:

public function &getData() {
    return $data;
}

which i think it should to be:

public function getData() {
    return $this->data;
}

What do you think?

The log message is being generated here:
https://github.com/paulfitz/daff-php/blob/master/gen/lib/coopy/Alignment.class.php#L159

We're comparing the following tables:

Array
(
    [0] => Array
        (
            [0] => lorem
            [1] => ipsum
        )

    [1] => Array
        (
            [0] => 1
            [1] => value
        )

    [2] => Array
        (
            [0] => 2
            [1] => value
        )

)

Array()

Result:

Alignment.hx:153: Ordering took too long, something went wrong
[["@:@",null,"0:-","1:-"],[null,"!","---","---"],[null,"@@","lorem","ipsum"],["1:-","---",1,"value"],["2:-","---",2,"value"]]

The diff generated is completely correct, but the log message being generated is printed to output and the resulting page is wrong.

Hi,

There you are another bug report, this time when diffing files: https://gist.github.com/gonzaloserrano/d74bbbf452f745b91a0f

$localTable = [
    ['col1', 'col2', 'col3', 'col4', 'col5'],
    [0, 0, 0, 0, 0]
];

against

$remoteTable = [
    ['col1', 'col2', 'col3', 'col4', 'col5'],
    ['xxx', 0, 0, 0, 0]
];

With this data it does generate an "empty" diff, but I noticed that changing 'xxx' by 1 generates diff data but says NULL->1 when i think it should be 0->1.

Thanks a lot for your time!

Hi,

I think there is a bug when apply a patch over a table, I created a gist with an example.
https://gist.github.com/sp-ruben-simon/f049beb378851bdfba80

Seems that if you delete a row and there is an identical row then the patcher doesn't have a good behaviour.

In the above example if you have these tables:

$localTable = [
    ['A', 'B'],
    [2, 2],
    [2, 2],
    [3, 8]
];

$remoteTable = [
    ['A', 'B'],
    [2, 2],
    [3, 8]
];

The result should be that:

$resultTable = [
    ['A', 'B'],
    [2, 2],
    [3, 8]
];

But returned a table full of nulls.

What do you think?

Should probably not create a release for every single commit

https://github.com/paulfitz/daff-php/releases

Having "paulfitz/daff-php": "dev-master" in the composer.json should pull the latest code from the master branch without having to make a release.

Gist: https://gist.github.com/gonzaloserrano/f7aaa84f69dbb80d701e

$localTable = [
["fd", "df"],
["fd", "fd"],
[null, "fd"],
["fd", null],
];

against

$remoteTable = [
["A", "new_column_2"],
[null, null],
["fd", "df"],
["fd", "fd"],
];

shoots

Notice: Trying to get property of non-object in /Users/gonzalo/daff-test/vendor/paulfitz/daff-php/gen/lib/coopy/TableDiff.class.php on line 347
<table>
<thead>
<tr class="index"><td class="index">@:@</td><td></td><td>0:0</td><td>1:1</td></tr>
<tr class="spec"><td class="index"></td><td>!</td><td>(fd)</td><td>(df)</td></tr>
<tr class="header"><th class="index"></th><th>@@</th><th>A</th><th>new_column_2</th></tr>
</thead>
<tbody>
<tr class="add"><td class="index">-:1</td><td>+++</td><td></td><td></td></tr>
<tr><td class="index">0:2</td><td></td><td>fd</td><td>df</td></tr>
<tr><td class="index">1:3</td><td></td><td>fd</td><td>fd</td></tr>
<tr class="remove"><td class="index">2:-</td><td>---</td><td></td><td>fd</td></tr>
</tbody>
</table>

Edit: interestingly, if you remove last row from both tables, you get a Alignment.hx:277: Ordering took too long, something went wrong.

Thanks!

Hi,

Testing last version i got an error when aligning a compare table which has floats.

E.g: the first row with a 2.5 value in a cell fails with:

{"success":false,"message":"Warning: array_key_exists(): The first argument should be either a string or an integer in /Users/gonzalo/dev/config-manager/vendor/paulfitz/daff-php/gen/lib/haxe/ds/StringMap.class.php line 21"}

edit: introduced in 1.1.14, 1.1.13 works fine

Thanks

I dont know if there is any value in having a custom error handler in a package that does not specifically deal with handling errors.

EG: I am using Cake which does a decent job of handling errors. Until daff is included. For example I have a custom error handler that emails issues to a bug tracker, spits them out in syslog which is then sent to 'the cloud' as the application is across a number of servers.

Other users that are not using a framework, my be including some package to specifically handle errors.

Daff is not adding value here, just being annoying. Big unformatted error. I would vote that error handling is removed (or at the very lest configurable with default off)

Figured I would make a new ticket here so it does not block up the other repo.

While it can now install and work with composer, there needs to be some auto loading going on so users don't need to manually include files.

I have not found the exact config just yet, will need to make a fork and test it out as the code is not psr-4 compatible. Should be a case of telling composer about the Boot.class.php file and all should be fine.

Hi,

I think there is a bug in the way the diff view is generated.

If i do a diff between two tables where i remove a cell from the remote version i have an html version like this one, which looks great:

But if i don't do the html version an do a $diffTableView->getData() i get this:

Array
(
    [0] => Array
        (
            [0] => @:@
            [1] => 
            [2] => 0:0
        )

    [1] => Array
        (
            [1] => @@
            [2] => A
            [0] => 
        )

    [2] => Array
        (
            [1] => ...
            [2] => ...
            [0] => 
        )

    [3] => Array
        (
            [1] => 
            [2] => zzz
            [0] => 3:3
        )

    [4] => Array
        (
            [1] => ---
            [2] => kkk
            [0] => 4:-
        )

)

Notice 0 row is fine but the 1-4 rows content order should be 0, 1, 2 not 1, 2, 0. They are different data structures, the first one is an integer indexed array in php but the second one is a hashmap, check this out:

php > $good = [0 => '4:-', 1 => '---', 2 => 'kkk'];
php > $bad  = [1 => '---', 2 => 'kkk', 0 => '4:-'];
php > echo json_encode($good);
["4:-","---","kkk"]
php > echo json_encode($bad);
{"1":"---","2":"kkk","0":"4:-"}

Thanks!