Issue parsing linebreaks in segments

[johanib]

Issue parsing linebreaks in segments

[johanib]

@sabas
I created this pr:
#135

I did some small refactoring to reduce the amount of nested if/else constructions.
Please let me know if you have any comments.

[sabas]

The processs doesn't exist actually, in my projects I use directly the master branch 🤣
I usually do it after someone reminds me to do it... I'll tag 1.1 in a few hours!

[iGrog]

unwrap function in Reader removes CR+LF+0x00, but unwrap in Parser doesn't:

$string = \preg_replace('#[\x00\r\n]#', '', $string);

Is that how it should be?

https://github.com/php-edifact/edifact/blob/master/src/EDI/Reader.php#L557

https://github.com/php-edifact/edifact/blob/master/src/EDI/Parser.php#L467

[sabas]

Hi, your input EDI has a line-break inside the segment, it shouldn't be valid as the linebreak character isn't in the UNOB charset...
I modified the Parser to add a feature to detect an error where there's no segment terminator in files that are formatted with one segment per line.
If you pass ->setStrict(true) to the Parser instance before running it on the edi it reverts to the previous behaviour... @johanib
@iGrog I believe the unwrap function in Reader is a version of more than 5 years ago of the same function in Parser :-D
Reader and Analyser need some refactoring probably....

[johanib]

@sabas Thanks! I'm not sure how I missed ->setStrict(true), but it solves my linebreak issue 😅

One more thing I don't understand, given these EDI segments:

FTX+AAI+++PLS ENSURE TO TAKE OUR APPROVAL PRIOR STUFFING ANY NON HAZ CHEMICA:LS'
FTX+ABV+++THIS BOOKING CONFIRMATION IS SUBJECT TO SEALING'
FTX+AAI+++THE SHIPPER SHALL NOT BE RESPONSIBLE FOR ANY COSTS/DELAYS OCCUR

I really can't figure out how to extract the FTX+AAI segments...
Do you know if this library is capable of reading these segments?
It seems it always returns with the error 'Segment is ambiguous'.

I wrote some hacky solution to accomplish this and a test to illustrate what I expect as output:
https://github.com/johanib/edifact/pull/1/files

[sabas]

What about returning you an array of results instead of null? Would it make sense? Are you using Reader to extract specific information?

[johanib]

For my purposes I do not need to extract the individual FTX+AAI segments.
But afaik, the readEdiDataValue function is build on the idea that the caller specifies what it wants to read, so I added the $offset parameter.

Returning an array would work for my case.

There are also some checks at the bottom of the readEdiDataValue function. They would need to be looped if the function would start to return arrays. So I'm not sure if that will make it better.

I can probably improve my initial suggestion to make it more nice if you think it's a good addition.

[sabas]

Of course, feel free to send a PR :)

[johanib]

🎊
Just a question: What is the release process like?
I see the latest tag / composer package is from October last year.
Can we release this version? :)