Existing functionality broken in a new release should be fixed and released as point releases.
Its a little insane to lock a fix away and make customers wait potentially months for the next major release, and possibly charge them all over again for something that used to work.
I think it would help Xojos reputation if they got into the habit of releasing a point release or two between major releases.
Right on target. Perfectly put.
There has to be a threshold for that statement. Things are bound to get past QC like this one that only affects listbox if celltextpaint or cellbackgroundpaint are used.
Its also worth noting again that because of how XDC fell in relation to the release of 2019r1, a point release couldnt even be considered before now as we no longer build anything resembling an FC within the month leading up to the show. Theres just too much going on. The week following the show (last week) is equally bad as things dont start to get back to normal for about a week.
Companies do this all the time though. For better or for worse, our bug reporting system is very transparent when it comes to things that have been fixed for an upcoming release. Most other companies dont mark things publicly as fixed until the next alpha cycle begins.
In 2019r1, we broke a vicious cycle which has plagued us for many years in which new features and unscheduled bug fixes could get into a release after we had determined what was going to be in the release without getting proper review. Hopefully this will mean that this current situation happens less often.
The flip side of this is that doing a point release is not an insignificant task and takes resources away from making progress on other things, never mind the additional stress that comes with it. We did this for a few years and it meant that we couldnt work on any multi-cycle features because we were continuously in a release mode, which is very different from feature building mode.
In summary, Im not saying that a point release wont happen. What I am saying is that it also depends on factors other than just it broke.
Answer ?
If a point release targeting a specific bug or two, of something that worked for (say) 14 major releases but was suddenly broken in the newest release is a significant task, then something else is wrong. Branch off from the newest release and fix something that worked previously. I expect every other company/framework/project to do it. I bet you do too.
[snip stuff about XDC etc]
I thought I was clearly making a general point about point releases in between your major releases, not this particular release.
Hi Emile,
I don’t have release dates. I issue builds to clients every few days as they make feature requests and bugs are found. I also use the software I produce on my own shows and build almost daily for this purpose as they evolve.
Delay in answer is because I’m currently on a touring production and have only just reached a day off!
Hi Jean-Yves,
I’m using 2018r4 for end client builds to get around it, and 2019r1 for my own work as I’m able to restart regularly to keep memory down. I’m using the 2019r1 IDE as I’m finding it a lot quicker for UI changes than previous versions. Also some URL/HTTP seems much better in 2019r1 hence my hesitancy to go back to releasing from 2018r4. My main project uses 10s of HTTP requests a second.
Hi Ivan,
I can’t see any mention of how a fix would be released.
Markus, I’m cool and collected thanks.
If you read what I put again, you’ll see that I was merely stating that XDC or that developers being otherwise engaged wasn’t mentioned and that I’d like to know more about a fix release.
Let’s draw a line under that one.
As @Gavin Smith so very well put, issuing a point release for something like this would be very reasonable.
[quote=436104:@Greg O’Lone]There has to be a threshold for that statement. Things are bound to get past QC like this one that only affects listbox if celltextpaint or cellbackgroundpaint are used.
[/quote]
Since listbox doesn’t have alternate row coloring built-in, I doubt there are many pro or semi-pro users that don’t use CellBackgroundPaint. I for one can’t use 2019r1 with any of my projects.
Dito.
+1
this has to hit nearly every xojo project in existence (including the IDE since it uses listboxes in a number of places)
Does this also apply if you use the default listbox without changes, or only when you define your own CellTextPaint / CellBackgroundPaint ?
[quote=436104:@Greg O’Lone]
The flip side of this is that doing a point release is not an insignificant task and takes resources away from making progress on other things, never mind the additional stress that comes with it.[/quote]
have you completely revamped the entire build process in the intervening 8 weeks since I left ???
A point release shouldnt be a huge pile of work otherwise ESP if its to fix ONE specific bug like this
[quote=436104:@Greg O’Lone]There has to be a threshold for that statement. Things are bound to get past QC like this one that only affects listbox if celltextpaint or cellbackgroundpaint are used.
The flip side of this is that doing a point release is not an insignificant task and takes resources away from making progress on other things, never mind the additional stress that comes with it. [/quote]
We are not talking about an unimportant bug but something that makes a release unusable.
That breaks any artificial threshold that Xojo might have set, and certainly warrants the effort to fix and release.
Otherwise every license holder should get his license extended by three months (assuming the bug is fixed in the next release)
I don’t know if I would say nearly every, but I would say many if not most… That Xojo inc may not realize that is disconcerting…
My license will still be good for the next release , but if it was not, I would be VERY upset about there not being a quick point release…
Given that a point release would be available for no extra cost for those who’s license is now expired but was not at the time of the main release, that is the right thing to do for such a significant bug IMO.
-Karen
This bug and the HTTPSocket bugs really deserve a quick point release: They are both regressions with pretty severe consequences, and the fact that many projects are going to run into both of them adds (multiplies?) the urgency.
I’d tend to agree
[quote=436614:@Karen Atkocius]I don’t know if I would say nearly every, but I would say many if not most… That Xojo inc may not realize that is disconcerting…
[/quote]
Of course the listbox bug is only a desktop top issue… maybe they no longer see desktop as the main part of the business?