Indeed, even <sep> has worked earlier... I haven't even realized till now. That makes it much more likely that this fix breaks existing templates. I wonder if this fix should only be enabled by incompatibleImprovements.
BTW, this also means that earlier if there was a <sep> outside a listing in a template, that wasn't seen as static text, and the template parsing has failed with "#sep must be inside a #list (or #foreach) block.". Quite a nasty bug that remained hidden for many years. On Fri, Dec 6, 2024 at 7:42 AM Alon Ziv <nola...@google.com.invalid> wrote: > I've tested Google's FreeMarker dependencies. I see a few breakages, but > they are all due to already-broken templates that somehow used to work > (e.g. one template used `<sep>` instead of `<#sep>` and is now broken - > likely due to PR-111). > > +1. > > > On Tue, Dec 3, 2024 at 5:52 PM Christoph Rueger <chrisrue...@gmail.com> > wrote: > >> +1 >> We ran our testsuite successfully with it (OSGi environment). No >> regressions. >> >> Am So., 1. Dez. 2024 um 23:00 Uhr schrieb Daniel Dekany < >> ddek...@apache.org >> >: >> >> > Hi all, >> > >> > Please vote on releasing FreeMarker 2.3.34! >> > >> > Release Notes: >> > https://freemarker.apache.org/builds/2.3.34-voting/versions_2_3_34.html >> > >> > Before proceeding, you should know that FreeMarker 2.3.x, for a long >> > time, always releases a normal and a "gae" variant on the same time, >> > which are technically two independent source trees (Git branches). The >> > "gae" variant contains a few small modification in the Java source >> > code to be Google App Engine compliant, and has freemarker-gae as the >> > Maven artifact name. Otherwise the normal and the "gae" branches are >> > identical. Hence they will be voted on together. >> > >> > The commits to be voted upon are: >> > - Normal (non-gae) variant: >> > >> > >> > >> https://github.com/apache/freemarker/commit/e169fce65f0b513a6e55a220b9733c8e2216e8fd >> > Commit hash: e169fce65f0b513a6e55a220b9733c8e2216e8fd >> > - "gae" variant: >> > >> > >> > >> https://github.com/apache/freemarker/commit/aba94c0d7b4ac2fd8b78a33ce06e8c07df6743e9 >> > Commit hash: aba94c0d7b4ac2fd8b78a33ce06e8c07df6743e9 >> > >> > The artifacts to be voted upon are located here: >> > https://dist.apache.org/repos/dist/dev/freemarker/engine/2.3.34/source/ >> > where the source release artifacts are: >> > - Normal (non-gae) variant: >> > apache-freemarker-src-2.3.34.tar.gz >> > - "gae" variant: >> > apache-freemarker-gae-src-2.3.34.tar.gz >> > >> > See the "Building FreeMarker" section of the README.md inside them for >> > build instructions! >> > As described there: >> > - You need to add the gradle-wrapper.jar manually >> > - You need multiple JDK versions installed, on locations where Gradle >> will >> > find them >> > - You may need -Pfreemarker.allowUnsignedReleaseBuild=true gradle >> command >> > line >> > option to building everything (not just the jar) >> > >> > The release artifacts are signed with the following key: >> > https://people.apache.org/keys/committer/ddekany.asc >> > >> > For convenience, we also provide binaries, which also need to be >> checked: >> > >> https://dist.apache.org/repos/dist/dev/freemarker/engine/2.3.34/binaries/ >> > and Maven artifacts in the ASF Staging Repository: >> > >> > >> https://repository.apache.org/content/repositories/staging/org/freemarker/freemarker/2.3.34/ >> > >> > Please try out the package and vote! >> > >> > The vote is open for a minimum of 72 hours or until the necessary >> number of >> > votes (3 binding +1s) is reached. >> > >> > [ ] +1 Release this package as Apache FreeMarker 2.3.34 >> > [ ] 0 I don't feel strongly about it, but I'm okay with the release >> > [ ] -1 Do not release this package because... >> > >> > Please add "(binding)" if your vote is binding. >> > >> > -- >> > Thanks, >> > Daniel Dekany >> > >> > -- Best regards, Daniel Dekany
