13298 Alignment override *.* fails on right side

[chrisjj]

e.g. http://img716.imageshack.us/img716/7...olescreens.png

It works on left side.

[Lutz]

Hi Chris!

As Alignment Override always works from left to right give "*.*" -> "*.mp3" a try or change sides and use "*.*" -> "*.flac".

Greetings Lutz

PS:
I'm thinking it works as follows:
1. BC detects a file named "jyrfgj.jgsd" on left
2. check if it matches an override rule -> yes "*.*" matches
3. get the replacement for the right side (i. e. "*.mp3")
4. look for a file named ""jyrfgj.mp3" on right
5. if existing align the two files

PPS:
Sorry for missing your left side annotation. Viewing your (very short) postings I usually only view your screenshot and don't read any further.

[Lutz]

Trying to check out, why "*.*" on right might not work:

1. BC detects a file named "jyrfgj.flac" on left
2. check if it matches an override rule -> yes "*.flac" matches
3. get the replacement "*.*" for the right side
4. look for a file named ""jyrfgj.flac" on right (same name.extension as left)
5. if existing align the two files (of course they don't)

To get this work the name "*" must match exactly the original name, but the extension "*" must match any extension.

Greetings Lutz

[chrisjj]

> Sorry for missing your left side annotation.

No problem. Yes I too find left side works and that's the workaround I am using until right side is fixed.

[Michael Bulgrien]

By design, Alignment Overrides work from Left to Right. The left side is tweaked and, if the right side matches, the files are aligned. It is like a simple equation: the calculation is defined on the left and the expected answer is provided on the right. Therefore, extra wildcards must be placed on the left-hand side just as regex back-reference captures must be placed on the left side...otherwise BC will not be able to determine what the "expected answer" is.

If you swap sides after creating an alignment override, BC3 keeps the aligned files aligned only because it does not re-apply the override. If you save the session after swapping sides, close BC3, then re-open it, the alignment will be lost if it cannot be processed left to right.

Ideally, the Alignment Overrides dialog should have a checkbox to specify which direction to apply the override (which side is the equation, and which side is the expected answer). Until such an enhancement is implemented, you have no choice but to work left to right.

[chrisjj]

> By design ... extra wildcards must be placed on the left-hand side

If you have any evidence for that being by design, I'd be interested to see it. The UI and documentation show none - they show nothing to indicate inferiority of the right side or of "extra" wildcards.

> Ideally, the Alignment Overrides dialog should have a checkbox to specify
> which direction to apply the override

I don't think that is ideal. I think ideal is *.* simply working on the right side.

[Aaron]

Hello Chris,

We do not currently support right sided regular expression masking. In the Alignment Override dialog, the dropdown arrow to the right of each textbox displays the list of accepted regular expression syntax.

Improving and expanding on the Alignment Override is on our Customer Wishlist.

[chrisjj]

We do not currently support right sided regular expression masking. In the Alignment Override dialog, the dropdown arrow to the right of each textbox displays the list of accepted regular expression syntax.

This is not a regular expression mask. "[ ] Regular expression" is unchecked.

[Michael Bulgrien]

>Ideally, the Alignment Overrides dialog should have a checkbox to specify
> which direction to apply the override

I don't think that is ideal. I think ideal is *.* simply working on the right side.

My use of word "ideally" was in the context of my knowledge that Alignment Overrides are directional. Ideally, it would be nice if Team Scooter could make the direction configurable (short term fix) until they can devote the time necessary for a complete overhaul (long term enhancement).

While there may be a place for directional processing of alignment overrides, it would be nice to have non-directional alignment overrides (i.e. manipulating the value on both sides then comparing the logical result of both sides to determine if there is a match). I've communicated this to Team Scooter before, hence the fact that improving alignment overrides is already on the customer wish list.

[chrisjj]

Ideally, it would be nice if Team Scooter could make the direction configurable (short term fix) until they can devote the time necessary for a complete overhaul (long term enhancement).

Even limited to directional, again that strikes me as far from ideal. It surely is not hard to auto-detect the direction, using a simple string test for *.*.

[Michael Bulgrien]

I wouldn't think Team Scooter would want to check for specific strings. I would think that they would check for patterns (i.e. more asterisks on one side than the other, or the presence of a regex back-reference capture). In this particular case, however, I would care more about having the functionality than I would worry about how it is implemented. If the difference between seeing a short-term fix and not seeing one is a direction checkbox instead of the extra effort of coding and testing "smart logic", then a checkbox is okay by me for a short-term fix.

[chrisjj]

I wouldn't think Team Scooter would want to check for specific strings.

They are already doing so, else there would not be this special case failure.