Iptc.Application2.DateCreated/TimeCreated formats are different for read/write

[lillolollo]

https://dev.exiv2.org/issues/1380

[clanmills]

Thanks for reporting this. I'm maintain the 0.27-maintenance branch. This looks like quite a substantial change that will ripple through the test suite and I'd prefer not to take this into the 0.27-maintenance branch.

Here is my most recent posting in the "Road Map": #1018 (comment)

[postscript-dev]

I started a GIMP issue #5948 last week to discuss this and mention some issues with the patch.

Since then while looking at something else, I came across another potential problem. The sizes for date and time inside TypeInfoTable are currently incorrect (in types.cpp). The values in the TypeInfoTable definition are correct for the IPTC standard but the output strings of DateValue::write() and TimeValue::write() are of a different length.

Hope that's some help.

[Jehan]

I'm the original patch author. So I apparently posted it at the wrong place (which I was told afterwards but didn't take the time to re-contribute here 😛).

This looks like quite a substantial change that will ripple through the test suite and I'd prefer not to take this into the 0.27-maintenance branch.

So would it be in a new (non-maintenance) version? 🙂

The patch may not be perfect, I didn't spend too much time on it, but in any case, there is definitely something wrong going on as the date metadata set by Exiv2 cannot be read by Exiv2 right now.

[lillolollo]

@clanmills Thanks for your wonderful work on exiv2 but if a program can't read what the same program write maybe bug is more appropriate than enhancement

[clanmills]

This looks like a tough puzzle. I'll set the milestone to v1.00. Setting the milestone means it will be considered for fixing. It doesn't guarantee that a contributor will take on the challenge.

[Jehan]

It doesn't guarantee that a contributor will take on the challenge.

One already has (me), if you check the first comment. I submitted a patch already 5 months ago when I published the report (though apparently not in the right place; by the way I would suggest to close dev.exiv2.org with a redirection here if you don't plan on checking it anymore! 😛): https://dev.exiv2.org/issues/1380

If my patch is not quite right, of course, I would welcome some input so that I can edit it. Of course, if someone else can come to another preferred approach, I don't mind either as long as it ends up fixed. I don't care whose patch gets in in the end. Just noting that if none else is up to the "challenge", then I'm still happy to be the one, as I already started.

[clanmills]

Until recently, Exiv2 was a one-man project and there's a limit to what I can achieve. @nehaljwani is it possible to modify Redmine as suggested?

I am very pleased to say that in the last few months, a few more folks have joined me. On Friday, we released Exiv2 v0.27.4 RC2 and I hope we're now done with v0.27.4. Yesterday I started to go through all the open issues and close some dead matters and assign ones requiring attention to milestone v1.00.

I appreciate that you have provided a patch. However somebody has to review the patch and get it merged. It's quite common that the effort to integrate a code change and update the test suite is more than the effort involved with to create the patch. And, as I don't know the new contributors very well, I can't be sure that if somebody accepts an assignment, they will follow through and deliver.

I want to set your expectation correctly by assigning this to v1.00. I am saying "Will be considered" and not "I promise to deal with this.".

One final point about this. I was 70 in January and had my heart set on retiring. Instead, I am working about 50-60 hours every week to complete v0.27.4 in April and have agreed to be the release engineer for v1.00 which is scheduled for 2021-12-15. My wife is reasonably asking When is this Exiv2 stuff going to stop dominating our life. When will I get some of your time to do various projects around the house and garden?

[clanmills]

I've opened a PR to merge your patch into 'main'. #1547

The damage to the test suite is considerable and every change will have to be investigated. Additionally, we should probably add some tests designed to deal with the matter under discussion.

@Jehan If I grant you write permissions on the Exiv2 repos, can you undertake the work necessary to bring this to a successful conclusion?

[Jehan]

Until recently, Exiv2 was a one-man project and there's a limit to what I can achieve. @nehaljwani is it possible to modify Redmine as suggested?
[…]
I appreciate that you have provided a patch. However somebody has to review the patch and get it merged. It's quite common that the effort to integrate a code change and update the test suite is more than the effort involved with to create the patch. And, as I don't know the new contributors very well, I can't be sure that if somebody accepts an assignment, they will follow through and deliver.

Ahah. Don't worry, I know very well how these kind of things work and how it can be to be alone or too few on a big project. 😉

I was not criticizing the time taken for review, nor anything else for that matter (I myself have many months-old patches I need to review), just reacting to the reaction that it needed a contributor.

@Jehan If I grant you write permissions on the Exiv2 repos, can you undertake the work necessary to bring this to a successful conclusion?

Sure. I'll do it in a branch (to only merge fully working code) and won't promise a specific schedule either (I'm here for the long run anyway), but I'd definitely work on this if you are good with it! 🙂

[Jehan]

P.S.:

by the way I would suggest to close dev.exiv2.org with a redirection here if you don't plan on checking it anymore!

@nehaljwani is it possible to modify Redmine as suggested?

If I can modify my previous suggestion: rather than redirecting dev.exiv2.org, it would be better to make it read-only, and maybe add a banner with a link to the github repo, saying it's the new tracker. This way, Exiv2 won't lose history of bug reports, discussions and old patches, which is often needed when investigating some issues (on GIMP, I regularly make some researches to understand a bug which brings me to look into older reports in our historic bugzilla which is now read-only).

[clanmills]

Thanks very much for such a considerate message. I've triggered an invitation from GitHub and granted you write permission to our repos, so you can can modify PR #1547. To merge, a team member has to approve your changes.

I encourage you to visit our Chat Server and "meet the gang".
(https://matrix.to/#/#exiv2-chat:matrix.org)

I did try to make Redmine read-only. I couldn't find anything in the Redmine Admin UI to control that. I can't believe it's not there. All databases can have readers who cannot write. So, I tried to chmod some files and that totally broke Redmine, so I restored the permissions.

I used to get emails from Redmine and followed up everything reported. A couple of years ago we had to move the hosting and I think after that the emails stopped arriving. In fact, I didn't even notice they had stopped arriving because there was only one per month or so. We moved to GitHub in April 2017. I've been busy and simply forgot about Redmine.

Nehal @nehaljwani worked with Andreas (the founder of Exiv2) to change the hosting in 2019. Let's hear what he has to say.

The move to the new host hasn't cause any particular pain. However, I have noticed that numerous links from Redmine are broken. And while this is a little irritating, as the years go by, those files are fading into history. However, there is a great deal of valuable discussion in Redmine.

In an effort to retire and leave Exiv2 is a maintainable state, I have written a book. Image Metadata and Exiv2 Architecture. https://clanmills.com/exiv2/book/

The book is incredibly detailed and possibly the most boring book ever written. The aim is collect my knowledge of this code and metadata into one place. If I don't discuss something in the book, a search of Redmine and/or GitHub and/or code is necessary. I hope everybody will find answers in the book.

[nehaljwani]

Apologies for the delay in response. I think the r/w access can be controlled using 'roles' at https://dev.exiv2.org/roles. I can revoke all 'edit/create' rights for anyone with no Admin access. Let me know if you wish to proceed, @clanmills :)

[luzpaz]

Any traction on this ?

[lillolollo]

This seems fixed by this #1547 reading #2406 (comment)

[lillolollo]

@neheb ?

[neheb]

no idea.