Summary: | deskutils/ganttproject: Fails to run with openjdk8, openjdk11 or openjdk16 | ||||||||||||||
---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|
Product: | Ports & Packages | Reporter: | Graham Perrin <grahamperrin> | ||||||||||||
Component: | Individual Port(s) | Assignee: | freebsd-ports-bugs (Nobody) <ports-bugs> | ||||||||||||
Status: | Open --- | ||||||||||||||
Severity: | Affects Many People | CC: | freebsd, rene | ||||||||||||
Priority: | --- | Flags: | freebsd:
maintainer-feedback+
koobs: merge-quarterly? |
||||||||||||
Version: | Latest | ||||||||||||||
Hardware: | Any | ||||||||||||||
OS: | Any | ||||||||||||||
Attachments: |
|
Description
Graham Perrin
2021-07-17 10:45:08 UTC
I'm afraid, you're right. That's a bug. I'll look into it and will do my best to post a patch by tonight. Created attachment 226571 [details]
Updated port diff
I owe an apology to all users of this port... I must have posted an outdated diff, which did not include updates to the ganttproject shell script that starts the application. I've now * updated Makefile with PORTREVISION=1 * included patch for port, which also includes $PREFIX/bin/ganttproject patch, which makes the file way more bourne shell compatible. Java can now be picked up via JAVA_HOME variable or "-j" parameter. Portlint and portclippy both ok on updated Makefile. Test on 12.2p6 and 12.2p9 were both ok. This will fix the breaking issue at hand. I'm planning to improve the binary further to allow for a help output in the shell script and an optional man page, but would rather put that in a separate patch, if that's ok. ^Triage: Please set the maintainer-approval attachment flag (to +) on patches for ports you maintain to signify approval Attachment -> Details -> maintainer-approval [+] Comment on attachment 226571 [details]
Updated port diff
added/updated maintainer approval. apologies for not setting this previously.
(In reply to Chris Moerz from comment #5) No need for apologies, the more we mention it the more people see and learn :) Just to be on the safe side: is there anything else I need to do? Otherwise I'll await any feedback and assume it'll get merged into ports eventually? Thanks (In reply to Chris Moerz from comment #7) I didn't review the patch earlier, but I can see hard coded LOCALBASE (/usr/local) and java versions that will need to be made variable: "/usr/local/openjfx14/lib" Other than that, nothing specific is waiting or needed. Having said that.. You mentioned 'Test on 12.2p6 and 12.2p9 were both ok'. If you can explicitly test with poudriere, that might prove handy (confidence) for committers looking at issues to close, and may indeed pick up additional issues. I would test with a custom PREFIX (to see the failure when its not /usr/local), and probably with multiple java versions If you need help with that: #freebsd-ports or #poudriere on Libera Chat IRC and we'd be happy to assist. Good catch on the /usr/local. Thanks. I'll save ourselves some future trouble by cleaning up my work procedure. I partially fixed that in the Makefile but didn't put the prefix in the patch... I'm now building on poudriere; it says 255 packages. That'll take a while. Will post the result once it's done. Created attachment 226782 [details]
Updated
Will post approval once poudriere is through.
Created attachment 226817 [details]
Updated port diff
Poudriere saved me from another blunder... somehow the PREFIX patch was not in the last file.
Created attachment 226818 [details]
Poudriere Log for FreeBSD 12.2
Created attachment 226819 [details]
Poudriere Log for FreeBSD 13.0
Please be advised, that the poudriere logs say PORTREVISION=2; that's because I had to increment due to another fix.
I still set revision 1 on the patch file, because I assume I'm not supposed to jump revisions when it comes to merging with ports? If leaving out revisions is fine, please let me know and I'm happy to either provide a revision 2 file or keep the jumps in the future.
Further feedback welcome, of course. I'm setting maintainer-feedback to (+) for the post; I assume this should confirm that I'm ok with the result so far. (In reply to Chris Moerz from comment #14) PORTREVISION just needs to be incremented if/when 'the resulting package changes'. An increment per change is not needed, just an increment on the current port In this case given there wasnt a build failure, and the issue is at run time, the PORTREVISION bump is needed (wouldn't have been if all package builds had been failing). PORTREVISION=1 per the patch is fine (if PORTREVISION doesn't currently exist) Note that openjdk16 expired today. Terribly sorry for dropping the ball on this. Just realized I had not completed this. I'm prepping an update to 3.2.3240 right now, that will fix this. It already works on my machine and I am waiting for poudriere to complete its run. If everyone is ok with moving to the next version, I suggest we close this once the updated version is committed. |