Various issues witn JabRef 5.0

[rvb-033]

Good morning,
I am running JabRef 5.0, release 29/11/2018 and Java 8 update 191 on Windows 10 V.1511. My computer is a I5 3.10Ghz, 8Mo. To my standards it is a big machine, it has been dimensioned to handle .raw image processing and does it pretty well.
My data base is, I believe, large : +/-4Mo, 7000+ records, probably +/-1000 Keywords. It is UTF-8 encoded.
I would like to report three issues :

• every time I open the database, JabRef finds duplicated entry keys, offers to solve them, and does nothing (this worked fine on 3.n versions). I know it does nothing, because if I close the database and open it again, I have the same message.
• If I copy / paste an entry, the pasted one has an encoding problem (accented letters - yes, I am french ! - are replaced by white question marks on black diamonds) in the copy.
• After adding a few new entries, JabRef becomes dead slow when typing in the keywords. I did not toggle off the auto-complete feature, it is probably a mistake. I had an issue like that ages ago and your advice has been to toggle off the feature that counts the number of occurrence of each keyword. I did not find this option on 5.0.

I must admit that I would prefer a feature enabling to copy a keyword from the key word list than a feature that tries to be clever and uses a lot of processing time...

I hope it helps,

Regards,

Hervé

[Siedlerchr]

Thanks for your report. Just a question?
How do you do copy paste?
Are you aware that you can directly paste bibtex code on the maintable? And copy works the same way.
I personally had no problem copying pasting this way. Even with accents or other Unicode chars

[rvb-033]

Copy / paste method is : right click on the entry, select the fist option in the pop-up, right click on an other entry, select paste.

Do you want screenshots ?

[rvb-033]

Good morning,
Copy and paste directly bibtext code is a good workaround, thanks for that.
Addition to the software slow down : when I see it slowing down, I save the library and relaunch JabRef. If I just close the library and reload it without closing and relaunching JabRef, it is still slow.
My understanding is that there is somewhere some memory that is not liberated and progressively saturates.
I hope it helps.
Regards,
Hervé

[wujastyk]

I am experiencing terribly slow data entry in the entry-editor too.

JabRef 5.0-dev--snapshot--2018-12-09--master--b52c97239
Linux 4.15.0-42-generic amd64
Java 1.8.0_191

Both in the field-tabs, and in the biblatex area, it can take two seconds from keypress to the appearance of a single character on the screen. Needless to say, Jabref 5 is unusable until this is improved.

Autocompletion is off.

Jabref 4.3.1 is fine for my daily work.

[rvb-033]

Latest version (december 14th) has a far better behaviour regarding entry-editor performances.
Thanks,
Hervé

[wujastyk]

JabRef 5.0-dev--snapshot--2018-12-17--master--2754c2891 Linux 4.15.0-42-generic amd64 Java 1.8.0_191 No improvement with the latest changes, unfortunately. This screencapture <https://www.dropbox.com/s/s374yj56dfta7ii/jabref-slowkeys.mp4?dl=0> illustrates the lag between typing and the appearance of characters on the screen. I was whacking the keys somewhat to make sure they were audible. :-)

[wujastyk]

JabRef 5.0-dev--snapshot--2018-12-19--master--949d20496
Linux 4.15.0-43-generic amd64
Java 1.8.0_191

Typing lag issue now fixed, at least 90% improvement.

[rvb-033]

Good morning,
I just updated to the latest 5.0 version (2019-07-01) and upgraded java to their latest version (211).
Two of the bug I reported in December have been corrected, thanks for that, but the whole tool is still awfully slow when one try to input something at least keywords and abstract.
Idle, I mean doing nothing, just with the JabRef window open, it consumes above 95% of the processor and 50% of the memory. My machine is an I5/8Mo. It handles 16Mo raw images without any problem !