On 07/10/2019 23:27, Vladimir Sitnikov wrote: > Milamber>The first release candidate for JMeter 5.2 (a9349adffb) has been > Milamber>prepared, and your votes are solicited. > > Thank you. > > Here are some comments. > They do not seem to block the release, however, we might want to fix them > in the future: > > 1) Vote mail contains short Git SHA only (a9349adffb). Do we want to have > full one somewhere? No, the short git sha seems sufficient for find the release: ~/GitRepos/JMeter/jmeter [master|✔] 11:36 $ git checkout a9349adffb [...] HEAD is now at a9349adff Current Gradle does not support Java 13, so downgrade GitHub Actions to Java 12 ~/GitRepos/JMeter/jmeter [v5.2-rc1|✔] > > 2) Vote mail misses a link to the signing PGP key (e.g. > https://people.apache.org/keys/committer/$username.asc) See sebb response. The KEYS file is the release way (if you act as the RM, you must add your gpg key into KEYS file) The link to KEYS file is into the Download page too. > 3) The source archive > contains apache-jmeter-5.2/buildSrc/subprojects/batchtest/build/ > and apache-jmeter-5.2/src/build/ folders > We need to exclude the folders via ,gitignore rules. +1, it isn't a blocker but better to fix before a RC2 Milamber > When I build source artifact locally I don't > have apache-jmeter-5.2/src/build/, so the resulting _src.zip is produced > with a different sha > > 4) javadoc task should specify locale explicitly. Currently, it inherits > the system locale, thus it produces for me which is not > right. > That causes my doc/api/...html files to be different from the staged ones > > 5) Javadoc seems to produce different results due to different parsing of > the "header" lines. > Source code: > https://github.com/apache/jmeter/blob/a9349adffb31f0b4d98ca11ce74e04c330661f2d/src/jorphan/src/main/java/org/apache/jorphan/util/JOrphanUtils.java#L168 > > Documentation: > https://apache.github.io/jmeter-site-preview/site/api/org/apache/jorphan/util/JOrphanUtils.html > Method Summary table contains > rightAlign(StringBuilder in, int len) > Right aligns some text in a StringBuilder N.B. <-- note it stops right > after the dot. > > Javadoc on my machine (1.8.0_152-b16) generates > rightAlign(StringBuilder in, int len) > Right aligns some text in a StringBuilder N.B. modifies the input builder > <-- note it includes "modifies ..." into the summary. > > I cannot tell which one follows the spec better, but that is yet another > source of the non-reproducibility of the binary artifacts. > > Vladimir >