fsearch @0.1: patch phase fails

[ryandesign (Ryan Carsten Schmidt)]

​https://build.macports.org/builders/ports-10.14_x86_64-builder/builds/100497/steps/install-port/logs/stdio

patching file configure.ac
Hunk #1 FAILED at 30.
1 out of 1 hunk FAILED -- saving rejects to file configure.ac.rej

[Schamschula (Marius Schamschula)]

For some reason the Portfile needed Epoch 1 to be set to build the new version.

Once I did, I found the above error as well. However, the new release version has some serious Linuxisms malloc_trim, strverscmp and AT_NO_AUTOMOUNT.

I've opened an upstream issue: ​https://github.com/cboxdoerfer/fsearch/issues/297

[ryandesign (Ryan Carsten Schmidt)]

Replying to Schamschula:

For some reason the Portfile needed Epoch 1 to be set to build the new version.

Sure. As far as vercmp is concerned, 0.1beta4 is newer than 0.1. Since that's not true, you have to increase epoch to update to that version.

[ryandesign (Ryan Carsten Schmidt)]

I see you reverted to 0.1beta4 which solved the problem but then updated to 0.1.2 which reintroduced the problem. I see that the upstream issue you opened is still open.

[Schamschula (Marius Schamschula)]

I forgot to update one patch file: [9a89538150f73d7a0889150e007e0c84e991a1c1/macports-ports]

[Schamschula (Marius Schamschula)]

Unfortunately, it seems a changed patch file doesn't seem to trigger a rebuild.

[ryandesign (Ryan Carsten Schmidt)]

It should. Buildbot is currently busy building other things and hasn't yet reached the commit in which you updated the patchfile.

Nevertheless, I usually (though not needed in this case since there was such a short interval between the commits) recommend modifying the Portfile whenever you modify any patchfiles, since MacPorts base only checks the Portfile when deciding whether to clean before trying again and since an interrupted patch phase does not keep track of which patchfiles had been successfully applied (#46927).