Miva Merchant 9.08.xx Bug Reports

[kayakbabe]

CAN'T SORT URI's by type in URI Managment

I need to remove all the "normal" types that I just imported and replace them with 301's but it's kinda messy in the admin since I can't sort by type.


It sure would be nice if we could sort by any table column in the admin. take a look at jquery datatables if you need a script that can sort by any column and also by multiple columns.

multiple column sorting would be really nice! as then we could sort by say... active/inactive and then by a custom field if we wanted.

so then I could find all active (miva field) and discontinued (my custom field).

[joshpappas]

I think I may have found a bug with multi-text custom fields. When I submit updates to my product data (not the multi-text field itself), my multi-text values submit as a single serialized value, which breaks the multi-text functionality.

Example color values before change:
Orange
Blue
Yellow
Silver


After update it's just one value all together:
Orange,Blue,Yellow,Silver

This seems to happen because I'm using field groups. It only happens when I "update" from the custom fields tab and not the tab of the field group containing the multi-text field.

[Eric Foresman]

Hi Joshpappas

We have a bug in for this issue, thanks for letting us know about it.

-Eric

[joshpappas]

One more thing... I'm using the "inventory at checkout" feature for the first time and was testing how it works. I had assumed that this checks inventory availability at the final order submission to guarantee that you can't oversell product, but this doesn't seem to be true. It looks like there's no check after the user gets to OPAY. So if multiple users make it to OPAY with the last available item(s) in their cart, they can both check out with it.

[ocpxc02]

I believe I uncovered an issue with the scheduled import functionality. If a file is being processed by the importer and before completing, an updated file is uploaded, the process stops. When the next scheduled time arrives, it seems as though the import attempts to resume where it left off (in my case, the first log entry was at 71%, but immediately fails). This process keeps repeating even after a new file has since been uploaded to the server a number of times. Instead of starting from the beginning, it keeps trying to resume where it first failed.

The only way out of this loop was to:
1) delete the file from the server,
2) allow the scheduled task to run again and not find the file,
3) upload the file again, and then
4) allow the schedule task to run again using the newly uploaded file.

[Eric Foresman]

Hi ocpxc02

I have added a bug for this issue. Thanks for letting us know about it.

-Eric

[kayakbabe]

&mvt:order:payments:descrip; doesn't contain the description, it contains the Module name.

I"m working on my order batch reports and also on some order displaying issues.


I realized that this issue also pertains to why we can't super easily repurpose the COD module for simple thinks like mailed in checks or purchase orders etc. I thought the ability to enter a "new name" for a payment method type in the payment method rules, would just show up everywhere. but it doesn't.

Seems like the PaymentMethodRules.desc isn't fully implemented yet.

It appears the admin_order item or the tbt_order_contents item looks for the OrderPayment.pay_id and uses that to get the module name since
orderpayments.pay_id = module.id

I think it should get the name I enter in the payment method rules instead.

Basically on checkout the PaymentMethodRules.descrip does display instead of the default name . YAY!
But this is the only place it works.
Instead of PaymentMethodRules.descrip being stored in the OrderPayment table the payment module's id is stored instead.


There is no easy association that I can see between the module.id and the PaymentMethodRules.desc
becuase PaymentMethodRules doesn't contain the module.id.
I'd have to do some kind of query on both the mod_code and meth_code to find the description.

Seems like at least the method.id should be stored in PaymentMethodRules as probably PaymentMethodRules.meth_id
Or IMHO an even better solution is to add one field added to the OrderPayment table
I would add OrderPayment.pay_desc
and store the PaymentMethodRules.desc value as it was at the time the order was made.

[kayakbabe]

Page templates

if you change a page template section version to look at something on that older version, but do not save it. (And, do not click reset. ) if you click on clear history, the current template is removed and replaced by the one showing even though it isn't the "saved" version.

So you can lose the version you "think" is saved.

[Eric Foresman]

Page templates

if you change a page template section version to look at something on that older version, but do not save it. (And, do not click reset. ) if you click on clear history, the current template is removed and replaced by the one showing even though it isn't the "saved" version.

So you can lose the version you "think" is saved.

Hi Kayakbabe,

I am not able to reproduce this issue, which template were you editing when you saw this? The specific template might make a difference.

-Eric

[kayakbabe]

eric, it's any page template that has versioning and a clear history. let me know when you see this screen capture.
It may not be a bug per se, but more of a confusing artifact of how browser forms caching works... but maybe refreshing the content view of the text area by using the DOM might make this less confusing.
Also I am using Safari on a Mac. I haven't tested using other browsers to see how they behave.

https://www.dropbox.com/s/dlo6on38qa...avior.mov?dl=0

[Eric Foresman]

eric, it's any page template that has versioning and a clear history. let me know when you see this screen capture.
It may not be a bug per se, but more of a confusing artifact of how browser forms caching works... but maybe refreshing the content view of the text area by using the DOM might make this less confusing.
Also I am using Safari on a Mac. I haven't tested using other browsers to see how they behave.

https://www.dropbox.com/s/dlo6on38qa...avior.mov?dl=0

Ok I see where the confusion is. When you click recall to view an older version of the template the template text area is reloaded with the desired old version. When you click on the Clear history link all of the template versions other than current and original get removed, but that has no affect on the currently displayed template. So while the template that is still being displayed has been deleted and only the original and known good/saved template exist in the database they are not the one displayed.

You can test this by adding 1 step to the procedure in your video. at the end Just click reset on the page or recall again to reload the template text area.


thanks for the video it was very helpful in identifying the issue.

-Eric

[kayakbabe]

Like you describe if you don't click update, click reset.. then you see the old version.

but guess what happens if you click update... it saves the code shown, AND doesn't make a new version. (so you lose the old version that you could have recalled if you had known not to update)

so it really is confusing.

[Eric Foresman]

Like you describe if you don't click update, click reset.. then you see the old version.

but guess what happens if you click update... it saves the code shown, AND doesn't make a new version. (so you lose the old version that you could have recalled if you had known not to update)

so it really is confusing.

I'm sorry, i don't see that behavior, when i click save as the last step, saving the displayed template which we know has been deleted. it does save AND creates a New Version of the template. your known good template is just the oldest one not named original. maybe there is an issue where all three versions are not displayed properly without a reset in safari. im using chrome, Firefox and IE11. i will try to hunt down a native safari machine to see if i can reproduce what you are getting.

-Eric


---edit/additional---

Ok I found someone with a mac who would let me try this out. Im still not seeing a case where the known good template was nuked and unrecoverable. At this point it might be best to open a support ticket to see if support can reproduce this issue while on your store.

Sorry I couldn’t be more help

-Eric

[new_user2018]

I'm so dumbfounded why this isn't a feature that it must almost certainly be a bug -

CSS Resources/JS Resources

Files added via Local File and included using <mvt:item name="head" param="css:resource_code"> are not dynamically loaded onto the page.

This means you can spend the time adding all this information to your JS/CSS resources tab, plug in the <mvt:item> tag with the appropriate param into your template and then if you make a change using the 'Edit Source' button or the file itself...

You'd have to make a new file (with a different name), new CSS/JS resource entry, new <mvt:item> tag just for that resource to avoid the browser cache!

Wouldn't it make more sense for the <mvt:item> tag to handle appending a ?<lastmodifieddatetimeint> to the src/href and changing that value every time the resource is updated? That way you can put in the CSS/JS Resource once and manage/edit effortlessly without need to do any additional versioning.

[rguisewite]

I'm so dumbfounded why this isn't a feature that it must almost certainly be a bug -

CSS Resources/JS Resources

Files added via Local File and included using <mvt:item name="head" param="css:resource_code"> are not dynamically loaded onto the page.

This means you can spend the time adding all this information to your JS/CSS resources tab, plug in the <mvt:item> tag with the appropriate param into your template and then if you make a change using the 'Edit Source' button or the file itself...

You'd have to make a new file (with a different name), new CSS/JS resource entry, new <mvt:item> tag just for that resource to avoid the browser cache!

Wouldn't it make more sense for the <mvt:item> tag to handle appending a ?<lastmodifieddatetimeint> to the src/href and changing that value every time the resource is updated? That way you can put in the CSS/JS Resource once and manage/edit effortlessly without need to do any additional versioning.

I have filed a bug on this. Though not exactly a "bug", it will be a good and useful feature to have. The file is stored on the file system, the modified date is updated when you edit it, so the behavior should be identical to if you had edited the file directly on the server (ie, how it was done prior to the CSS/JS Resources being added). Either way, it's a good thing to have pointed out, so thanks!