Summary: | app-office/lyx-2.0.5.1-r1 fails in the install stage | ||
---|---|---|---|
Product: | Gentoo Linux | Reporter: | Erik Zeek <zeekec> |
Component: | Current packages | Assignee: | pavel sanda <ps> |
Status: | RESOLVED FIXED | ||
Severity: | normal | CC: | aballier, cjk, proxy-maint, ps, tex |
Priority: | Normal | ||
Version: | unspecified | ||
Hardware: | All | ||
OS: | Linux | ||
Whiteboard: | |||
Package list: | Runtime testing required: | --- | |
Attachments: | Build log |
Description
Erik Zeek
2013-02-22 05:45:08 UTC
The $(scriptdir) is empty in the make files, so the install-data-hook is looking in the wrong place. does lyx-2.0.5.1 compiles for you? i was not able to reproduce the problem here. i also wonder where from comes this 'roundtrip.py'. there is no single string or file like that in the tarball. weren't you playing with something? (In reply to comment #2) > does lyx-2.0.5.1 compiles for you? Not any more, but it is currently installed. I just looked again at the hook. It's wrong. It looks for files in the $(scriptdir)/, but chmods them in $(DESTDIR)$(pkgdatadir)/. On my system $(scriptdir) is empty, so it's looking for files in the root directory, and, if there are no *.py files in your root directory, the body of the for loop is never executed. All that would be fine, except that, strangely, I have four files in the / directory, /roundtrip.m, /roundtrip.py, /test_datatypes.m, and /test_datatypes.py. All of which belong to dev-python/oct2py-0.4.0. I'll bet that if you create a file.py in your / directory, lyx wull fail to install too. Erik After removing the offending files, I was able to update lyx. The for loop is still wrong, but it no-longer breaks the installation. indeed, this part of makefile is wrong, forwarding to upstream typo in makefile causing this and code around was changed in forthcoming 2.1. just backporting the single typo caused different install problem in 2.0.x. due to exotic conditions needed for this bug being triggered i tend to wait for 2.1. (In reply to comment #7) > typo in makefile causing this and code around was changed in forthcoming 2.1. > just backporting the single typo caused different install problem in 2.0.x. > due to exotic conditions needed for this bug being triggered i tend to wait > for 2.1. That sound good to me. The other package, oct2py, fixed their installation, so I'm not tickling this bug anymore. Thanks for all the hard work, Erik 2.1 is out. |