Community
Participate
Working Groups
I am trying to launch a headless application using the launcher executable and the -noSplash option. There is no DISPLAY. Before Mars this worked just fine without errors, but now I get errors printed: Eclipse: Cannot open display: Eclipse: GTK+ Version Check The application is started nonetheless, but the errors are annoying. --launcher.suppressErrors does not help either.
CQ:WIND00-WB4-5749 Cosmetic only, but still an annoying regression in Mars (due to SWT probably).
This is still an issue in Mars Update 1.
Workaround: Instead of ./eclipse -noSplash ... run this: java -jar plugins/org.eclipse.equinox.launcher_*.jar Note that the "java -jar syntax" requires vmargs to be specified *before* application arguments, and without the -vmargs option. Also, in case the eclipse install was the result of "updates", it's possible that there are multiple o.e.equinox.launcher_*.jar files and you'd need to check which version to use (most likely latest). Personally, I'd much prefer a launcher command-line option like ./eclipse -noSplash --launcher.GTK_Check=0 ... to turn off the GTK version check.
(In reply to Martin Oberhuber from comment #3) > > Personally, I'd much prefer a launcher command-line option like > ./eclipse -noSplash --launcher.GTK_Check=0 ... > to turn off the GTK version check. I would much prefer this too. "We" should not have to go to extra programming effort to get "clean output". This has annoyed me for all of Neon (at least), but now I have a use case that it actually causes something "bad" to happen. It messes up my nicely formatted output! Here's roughly the way it is supposed to look, and the way it looked for the several hours I spent getting it to work: Checking repo: file:///home/data/httpd/download.eclipse.org//releases/neon/ Number of IUs in releases_neon: 27724 [INFO] Elapsed seconds for this repo: 9 Checking repo: file:///home/data/httpd/download.eclipse.org//staging/neon/ Number of IUs in staging_neon: 8942 [INFO] Elapsed seconds for this repo: 4 Checking repo: file:///home/data/httpd/download.eclipse.org//releases/mars/ Number of IUs in releases_mars: 27956 [INFO] Elapsed seconds for this repo: 9 Repository Number of IUs releases_neon 27724 staging_neon 8942 releases_mars 27956 = = = = = = = = = And, then, when I moved it to "Hudson", here is the way it looks: Checking repo: file:///home/data/httpd/download.eclipse.org//releases/neon/ Eclipse: Cannot open display: Number of IUs in releases_neon: 27724 [INFO] Elapsed seconds for this repo: 9 Checking repo: file:///home/data/httpd/download.eclipse.org//staging/neon/ Eclipse: Cannot open display: Number of IUs in staging_neon: 8942 [INFO] Elapsed seconds for this repo: 4 Checking repo: file:///home/data/httpd/download.eclipse.org//releases/mars/ Eclipse: Cannot open display: Number of IUs in releases_mars: 27956 [INFO] Elapsed seconds for this repo: 9 Repository Number of IUs releases_neon 27724 staging_neon 8942 releases_mars 27956 = = = = = = = = What a pain to have to spend even more time on what was meant to be a "simple utility" (but with nicely formatted output :)
The bug is in launcher. Is it ok to fix this in 4.6.1?
(In reply to Sravan Kumar Lakkimsetti from comment #5) > The bug is in launcher. Is it ok to fix this in 4.6.1? Yes, this is not a candidate for Neon.0
(In reply to Sravan Kumar Lakkimsetti from comment #5) > The bug is in launcher. Is it ok to fix this in 4.6.1? Any chance of looking at this for Neon.2? Nothing happend for Neon.1
I will look into this in M3
New Gerrit change created: https://git.eclipse.org/r/83036
Gerrit change https://git.eclipse.org/r/83036 was merged to [master]. Commit: http://git.eclipse.org/c/equinox/rt.equinox.framework.git/commit/?id=7259e7994a1de3254e321b735a3f375c5d961d09
(In reply to Eclipse Genie from comment #10) The source change looks good to me (except for the bad indentation). But to test this, we need a rebuild of the natives. Arun?
(In reply to Markus Keller from comment #11) > (In reply to Eclipse Genie from comment #10) > The source change looks good to me (except for the bad indentation). > But to test this, we need a rebuild of the natives. Arun? Thanks for pointing out Markus, we do seem to have missed rebuilding the native binaries. Sravan, are you going to take care of this?
Binaries have been pushed with http://git.eclipse.org/c/equinox/rt.equinox.binaries.git/commit/?id=136a5af29a8c50fca138191b704d86626073796e Verified in 4.7 M3 candidate eclipse-SDK-I20160606-1100-linux-gtk.tar.gz (32bit).
Are there still plans to release this for Neon.2? We are heading into RC builds so it better get in quick.
New Gerrit change created: https://git.eclipse.org/r/85108
Gerrit change https://git.eclipse.org/r/85108 was merged to [R4_6_maintenance]. Commit: http://git.eclipse.org/c/equinox/rt.equinox.framework.git/commit/?id=d91b488d62b57ca9d145bedcfd4aff6c6d327c24
Sravan, I have merged the backport of the source changes, will you be taking care of the binaries rebuilds?
(In reply to Arun Thondapu from comment #17) > Sravan, I have merged the backport of the source changes, will you be taking > care of the binaries rebuilds? Sure I will do that
Binaries have been rebuilt and pushed via https://git.eclipse.org/c/equinox/rt.equinox.binaries.git/commit/?h=R4_6_maintenance&id=114160d56c8af41428d564976801a6d7f63f6cff
Looks like this cause regression via bug 534759.
I opened https://git.eclipse.org/r/#/c/122853/ for bug 534759, would be nice if someone can provide validation steps for bug 472042 (this ticket). So far I've checked with unset DISPLAY and running Java code formatter application. I don't see anything on the console other than the formatter outputs.
(In reply to Simeon Andreev from comment #21) > I opened https://git.eclipse.org/r/#/c/122853/ for bug 534759, would be nice > if someone can provide validation steps for bug 472042 (this ticket). > > So far I've checked with unset DISPLAY and running Java code formatter > application. I don't see anything on the console other than the formatter > outputs. I am not familiar with the "Java code formatter application". The script I mentioned in comment 4 is based on org.eclipse.equinox.p2.director application. For exact steps -- an invocation something like the following should show the issue: ./eclipse4.6.1/eclipse/eclipse -nosplash --launcher.suppressErrors -application org.eclipse.equinox.p2.director -repository "http://download.eclipse.org/releases/oxygen/" -list -vm /shared/common/jdk1.8.0_x64-latest/bin/java 1>repoOutput.txt That is assuming Eclipse 4.6.1 has been extracted to ./eclipse4.6.1/eclipse Such as by: tar -xf eclipse-platform-4.6.1-linux-gtk-x86_64.tar.gz -C ./eclipse4.6.1 I mention "4.6.1" since I could reproduce the issue with that version, but not 4.6.3 or 4.7.3a. If you need it, 4.6.1 is now in 'archives' in /home/data/httpd/archive.eclipse.org/eclipse/downloads/drops4/R-4.6.1-201609071200/eclipse-platform-4.6.1-linux-gtk-x86_64.tar.gz [That's the location to use in rsync, the web URL should be obvious] I did also test if unsetting the DISPLAY variable sufficed -- because , and it did seem to on my system (Ubuntu 16.04). On my latest Ubuntu install, using 4.6.1, the "extraneous" messages after "unset DISPLAY" (and running 4.6.1) are actually worse than comment 4: = = = Failed to connect to Mir: Failed to connect to server socket: No such file or directory Unable to init server: Could not connect: Connection refused Eclipse: Cannot open display: = = = On the Eclipse "build" machine, which literally has no display, the message is only the last line: = = = Eclipse: Cannot open display: = = = probably similar for latest Jenkins HIPP servers, but did not check. I hope all that is the type of information you were looking for. Let me know if not.
(In reply to David Williams from comment #22) Sorry, accidental 'save' while editing. For the following: > I did also test if unsetting the DISPLAY variable sufficed -- because , and > it did seem to on my system (Ubuntu 16.04). I meant to say: I did also test if unsetting the DISPLAY variable sufficed -- because I did not think that it would -- but it did seem to on my system (Ubuntu 16.04).
