I'm stealing Mat Booth's thunder, but EPIC will shortly be available in Fedora (definitely in F9, due out in a couple of weeks, and possibly backported for F8 as well).
During the reviews I noticed that there was a dependency on gnu-regexp. Why is that? Is java.util.regex.Pattern insufficient?
If you would like to refer to this comment somewhere else in this project, copy and paste the following link:
The dependency on gnu.regexp is historical only. I've replaced it by java.util.regex in some source files that I modified, but not everywhere. I vaguely remember that java.util.regex was not just functionally equivalent, but more efficient, too.
If you would like to refer to this comment somewhere else in this project, copy and paste the following link:
Regarding the review, I don't have much to add. The ANTLR warnings are because the grammar specifications were hacked together in hurry (a trial-and-error process more than relying on any specifications). As far as the test suites are concerned, there are two of them: the first one - classes in "org.epic.perleditor-test/src" - contains low-level standalone JUnit tests that don't generate false alarms. Perhaps the most valuable among them are those that exercise the ANTLR grammars. The second test suite - classes in "org.epic.perleditor-test/src-pde" - contains JUnit tests that can only be run within a hosted Eclipse workbench. Some of them do generate false alarms depending on the environment - they are basically only guaranteed to run on my machine.
I will take patches for replacing gnu.regexp with java.util.regex. Just take care that you test that the replacements are still working.
A link to CVS is available from the menu on the home page. The downloadable files are outdated because I am too lazy to go through the sf.net file management workflow on every minor release. One day I might automate it. Meanwhile, I just added a note to the downloads section.
If you would like to refer to this comment somewhere else in this project, copy and paste the following link:
Hi Jan, I'm the guy who packaged Epic up for Fedora.
Thanks for posting on here, Jack. It was a bit remiss my part not to really.
One quick question for you, Jan: Would you be willing to update to the latest version of JDOM? In order to get Epic into Fedora, I wrote a small patch that fixes its use of APIs that have changed in the last major version bump. It's not a big patch and I have been using it since Christmas when I first submitted the package for review without trouble.
I'm stealing Mat Booth's thunder, but EPIC will shortly be available in Fedora (definitely in F9, due out in a couple of weeks, and possibly backported for F8 as well).
During the reviews I noticed that there was a dependency on gnu-regexp. Why is that? Is java.util.regex.Pattern insufficient?
Which version of EPIC is going to be included?
The dependency on gnu.regexp is historical only. I've replaced it by java.util.regex in some source files that I modified, but not everywhere. I vaguely remember that java.util.regex was not just functionally equivalent, but more efficient, too.
Fedora is apparently tracking the unstable branch.
Once it's in the build system, it should show up here:
http://koji.fedoraproject.org/koji/packageinfo?packageID=6115
You may also be interested in the discussion regarding its review for inclusion:
https://bugzilla.redhat.com/show_bug.cgi?id=426884
Will you take patches for replacing gnu.regexp with java.util.regex?
By the way, a suggestion for the homepage: the Downloads link points to outdated version info, and there's no pointer to EPIC's CVS.
Tracking the unstable branch is good.
Regarding the review, I don't have much to add. The ANTLR warnings are because the grammar specifications were hacked together in hurry (a trial-and-error process more than relying on any specifications). As far as the test suites are concerned, there are two of them: the first one - classes in "org.epic.perleditor-test/src" - contains low-level standalone JUnit tests that don't generate false alarms. Perhaps the most valuable among them are those that exercise the ANTLR grammars. The second test suite - classes in "org.epic.perleditor-test/src-pde" - contains JUnit tests that can only be run within a hosted Eclipse workbench. Some of them do generate false alarms depending on the environment - they are basically only guaranteed to run on my machine.
I will take patches for replacing gnu.regexp with java.util.regex. Just take care that you test that the replacements are still working.
A link to CVS is available from the menu on the home page. The downloadable files are outdated because I am too lazy to go through the sf.net file management workflow on every minor release. One day I might automate it. Meanwhile, I just added a note to the downloads section.
Hi Jan, I'm the guy who packaged Epic up for Fedora.
Thanks for posting on here, Jack. It was a bit remiss my part not to really.
One quick question for you, Jan: Would you be willing to update to the latest version of JDOM? In order to get Epic into Fedora, I wrote a small patch that fixes its use of APIs that have changed in the last major version bump. It's not a big patch and I have been using it since Christmas when I first submitted the package for review without trouble.
Here's a link to it: http://cvs.fedoraproject.org/viewcvs/rpms/eclipse-epic/devel/epic-jdom-api-changes.patch?rev=1.1
Regards,
Mat
I upgraded to jdom-1.1 and applied your patch in the just released 0.6.23.
Excellent, thanks!